                          M A N U A L . T X T
                         == How to Use MCPP ==
                Kiyoshi Matsui      kmatsui@t3.rim.or.jp

V.2.0   1998/08     First released.                             kmatsui
V.2.1   1998/09     Updated according to C99 1998/08 draft.     kmatsui
V.2.2   1998/11     Updated according to C++98 Standard.        kmatsui
V.2.3 pre-release 1     2002/08     Updated according to C99 Standard.
                    Ported to Linux/ GNU C, CYGWIN and LCC-WIN32.
                    GNU C-compatible features augmented.        kmatsui
V.2.3 pre-release 2     2002/12     Ported to GNU C V.3.2.
                    Revised some wording.                       kmatsui
V.2.3 release   2003/02     Finally released.                   kmatsui
V.2.3 patch 1   2003/03     Slightly modified.                  kmatsui
V.2.4 prerelease    2003/11     Added porting to Visual C++.
                Added #pragma __preprocess, #pragma __preprocessed
                                                                kmatsui
V.2.4 release       2004/02     Extended multi-byte character handling.
                    Added porting to Plan 9/pcc.
                                                                kmatsui
V.2.4.1         2004/03     Revised recursive macro expansion, and added
                    -c option.
                                                                kmatsui


                                Contents

1.  Overview
    1.1     High Portability
    1.2     Standard C Mode with Highest Conformance and Other Modes

2.  Invocation Options and Environment Settings
    2.1     How to Pass Options to MCPP
    2.2     How to Specify Invocation Options
    2.3     General Options
    2.4     Options by MCPP Mode
    2.5     General Options Except for Some Compiler Systems
    2.6     Options by Compiler System
    2.7     Environment Variables
    2.8     Multi-Byte Character Encodings
    2.9     How to Use MCPP in One-Path Compilers
    2.10    How to Make MCPP Available in IDE

3.  Enhancement and Compatibility
    3.1     #pragma __put_defines,
                    #pragma __preprocess, #pragma __preprocessed,
                    #put_defines, #preprocess, #preprocessed
        3.1.1   Pre-preprocessing of Header File
    3.2     #pragma __once, #pragma once
        3.2.1   Tool to Write #pragma __once to Header Files
    3.3     #pragma __include_next, #pragma __warning_cpp,
                    #include_next, #warning
    3.4     #pragma __push_macro, #pragma __pop_macro,
                    #pragma push_macro, #pragma pop_macro
                    #pragma __setlocale, #pragma setlocale
    3.5     #pragma __debug_cpp, #pragma __end_debug_cpp,
                    #debug, #end_debug
    3.6     #assert, #asm, #endasm
    3.7     New C99 Features (_Pragma() Operator, Variable Argument
                    Macro and others)
    3.8     Asm Statement in Borland C and Other Special Syntaxes
    3.9     Compatibility with GNU C/Cpp
        3.9.1   Preprocessing FreeBSD 2/Kernel
        3.9.2   Preprocessing FreeBSD 2/Libc
        3.9.3   Problems Concerning GNU C 2/cpp
        3.9.4   Preprocessing Linux/glibc 2.1
        3.9.5   To Use MCPP with GNU C 2
        3.9.6   Preprocessing GNU C 3.2R
        3.9.7   To Use MCPP with GNU C 3
    3.10    Visual C++ .net System Header Problems
        3.10.1  Comment Generating Macro?

4.  Implementation-defined Behaviors
    4.1     Status Value on Exit
    4.2     Include Directory Search Path
    4.3     How to Construct Header Name
    4.4     Evaluation of #if Expression
    4.5     Character Constant Evaluation in #if Expression
    4.6     #if sizeof (type)
    4.7     How to Handle White-Space Sequence
    4.8     Default Specifications for MCPP Executables

5.  Diagnostic Messages
    5.1     Diagnostic Messages Format
    5.2     Translation Limits
    5.3     Fatal Errors
        5.3.1   MCPP's Own Bugs
        5.3.2   Physical Errors
        5.3.3   Translation Limits and Internal Buffer Errors
        5.3.4   #pragma __preprocessed Related Errors
    5.4     Errors
        5.4.1   Character and Token Related Errors
        5.4.2   Unterminated Source File Related Errors
        5.4.3   Ill-Balanced Preprocessing Group Related Errors
        5.4.4   Simple Syntax Errors on Directive Lines
        5.4.5   Syntax Errors in #if Expressions
        5.4.6   #if Expression Evaluation Errors
        5.4.7   #define Related Errors
        5.4.8   #undef Related Errors
        5.4.9   Macro Expansion Errors
        5.4.10  #error and #assert
        5.4.11  Failure of #include
        5.4.12  Other Errors
    5.5     Warnings (Class 1)
        5.5.1   Character, Token and Comment Related Warnings
        5.5.2   Unterminated Source File Related Warnings
        5.5.3   Directive Line Related Warnings
        5.5.4   #if Expression Related Warnings
        5.5.5   Macro Expansion Related Warnings
        5.5.6   Line Number Related Warnings
        5.5.7   #pragma __warning_cpp, #warning
    5.6     Warnings (Class 2)
    5.7     Warnings (Class 4)
    5.8     Warnings (Class 8)
    5.9     Warnings (Class 16)
    5.10    Diagnostic Messages Index

6.  Reporting on Bugs and Others
    6.1     MCPP's Bug?
    6.2     malloc() Related Bugs
    6.3     How to Report Bugs
    6.4     Give Us Your Feedback


                              1. Overview

MCPP is a C preprocessor written by kmatsui (Kiyoshi Matsui) based on
the DECUS cpp developed by Martin Minow, and then rewritten entirely.
MCPP means Matusi cpp.  This software is supplied as source codes, and
to use MCPP in any compiler systems, a small amount of compiler system-
specific modifications are required before it can be compiled into an
executable. [1]

This document describes the specification for a MCPP executable that has
been already ported to a certain compiler system.  It also explains
briefly how to port MCPP to a compiler system.  For those who want to
know more about MCPP or want to port it to other compiler systems, refer
to MCPP source and its document "porting.txt".

All these sources and related documents are provided as free software.

Before going into detail, some of the MCPP features are introduced here.
(The sections 1.1 and 1.2 are identical with those of porting.txt.)

Note:
[1] To use a MCPP executable, you must replace a compiler-system-
provided preprocessor with it.  Therefore, the MCPP executable has the
same name with it.  In many case, the name is cpp.


1.1     High Portability

MCPP is portable.  It supports various operating system, including GNU/
Linux, DOS/Windows.  It's source code is also portable.  It can be
compiled by compilers which support Standard C or C++ (ANSI/ISO C or C++)
as well as ancient ones which only support K&R 1st.

It uses only classical library functions.  Some library functions have C
source code appended.  MCPP was developed with emphasis placed on
portability, it will never happen to this Standard C-conformant
preprocessor that it cannot be compiled because, for example, you do not
have the Standard C compiler system or a part of your compiler system is
not Standard-C conformant.

To port MCPP to various compiler systems, in most cases, all you have to
do is to modify some macros in the header files and compile it.  In the
worst case, adding several dozens of lines to the source file, system.c,
would be enough.

As the MCPP executable runs in a small memory, it can be executed even
on an 16-bit system with a small memory space, of course, under
considerable limitation.

MCPP's multi-byte character (Kanji) facility can handle Japanese EUC-JP
and shift-JIS, Chinese GB-2312, Taiwanese Big-5 and Korean KSC-5601 (KSX
1001). The 32-bit or more system ported MCPP can also handle ISO-2022-JP
and UTF-8.  In addition, if the compiler-proper fails to recognize shift-
JIS or Big-5, MCPP can complement it.


1.2     Standard C Mode with Highest Cnformance and Other Modes

By modifying some macros in the header file, system.H, on compilation of
MCPP itself, a preprocessor in various modes is generated: Standard C, K
&R 1st. and "Reiser" model cpp modes.  MCPP also has what I call
"post-Standard mode".   MCPP in the Standard C mode provides an
execution option that allows MCPP to run as C++ preprocessor.

Different from many existing preprocessors, MCPP of the Standard C mode
has the highest conformance to Standards, such as ISO/IEC 9899:1990 and
its Corrigendum 1:1994, and Amendment 1:1995, as well as C99 (ISO/IEC
9899:1999).  The purpose of MCPP is to become a reference model of
Standard C preprocessors.

Even if the compiler-proper fails to conform to the Standard C, MCPP
tries to compensate it whenever possible.  For example, it provides the
functionality to concatenate adjacent string literals on behalf of the
complier proper that does not.

In addition, it provides several useful enhancements: #pragma
__debug_cpp, which traces the mechanism of macro expansion or #if
expression evaluation, and the header file "pre-preprocessing" facility.

The modes mentioned above must be specified on compilation, not
execution of, MCPP.  This is to reduce a MCPP size and to avoid complex
execution options.

However, MCPP also provides several useful execution options, such as
warning level or include directory specification options.

MCPP will never go into an indefinite error loop, or output an
inappropriate message even if it encounters a serious error in source
code.  It always provides an accurate and descriptive diagnostic message
and processes the error condition suitably.  It also issues a warning
message for a portability problem.  Detailed documentation is also
provided.

A disadvantage of MCPP, if any, is slower processing speed.  It takes
twice to three times more time than GNU C/cpp, but seeing that its
processing speed is almost the same as that of Borland C 5.5/cpp and
that it runs a little bit faster when the header file pre-preprocessing
facility is used, it cannot be described as particularly slow.  MCPP
puts an emphasis on standard conformance, source portability and
operability in a small memory space, making this level of processing
speed inevitable.

Validation Suite for Standard C Preprocessing, which is used to test the
extent to which a preprocessor conforms to Standard C, its documentation
cpp-test.txt, which contains results of applying Validation Suite to
various preprocessors, are also released with MCPP.  When looking
through this file, you will notice that so-called Standard C conformant
preprocessors have so many conformance-related problems.

During the course of developing MCPP V.2.3, it was selected as one of
the "Exploratory Software Projects for 2002" by Information-Technology
Promotion Agency (IPA), Japan, along with its Validation Suite.  From
July 2002 to February 2003, the project, financed by IPA, preceded under
advice of Yutaka Niibe project manager.  I asked "HighWell, Inc."
Limited Company, Tokyo, for translation of all the documents.  For
technical details, I revised and corrected the translated documents.

MCPP was continuously adopted to one of the "Exploratory Software
Projects" in 2003 by Hiroshi Ichiji project manager.  The update of MCPP
proceeded into the next version, V.2.4. [1]

ISO/IEC 9899:1990 (JIS X 3010-1993) had been used as C Standard, but in
1999, ISO/IEC 9899:1999 was adopted as a new Standard.  This manual
calls the former C90 and latter C99.  The former is generally called
ANSI C or C89 because it migrated from ANSI X3.159-1989.   ISO/IEC 9899:
1990 + Amendment 1995 is sometimes called C95.

Note:
[1] For details on "Exploratory Software Projects", visit the following
web site:

    http://www.ipa.go.jp/jinzai/esp/

MCPP source, its documentation and Validation Suite, as well as those
under revision, are available at the following CVS repository:

    http://cvs.m17n.org/cgi-bin/viewcvs/?cvsroot=matsui-cpp

You can download a tar-ball from the above site.

They are also available by anonymous ftp from the following site:

    ftp://ftp.m17n.org/pub/mcpp/

The following web page provides guidance.

    http://www.m17n.org/mcpp/

MCPP V.2.2 and Validation Suite V.1.2 are also available at the
following Vector Co.'s sites.  They are also contained in the CD-ROM
called "PACK for WIN GOLD".  Although this software is registered in the
directory called "dos/prog/c", it is not MS-DOS specific.  MCPP source
code is compilable on various OSs, such as UNIX, WIN32/MS-DOS.

    http://download.vector.co.jp/pack/dos/prog/c/cpp22src.lzh
    http://download.vector.co.jp/pack/dos/prog/c/cpp22bin.lzh
    http://download.vector.co.jp/pack/dos/prog/c/cpp12tst.lzh

It seems that http://download.vector.co.jp can be substituted with ftp:/
/ftp.vector.co.jp/.

The text files in these archive files available at Vector use [CR]+[LF]
as a <newline> and encode Kanji in shift-JIS for DOS/Windows.  On the
other hand, those available at m17n.org use [LF] as a <newline> and
encode Kanji in EUC-JP for UNIX.  So, when they are used in other OS
environments, conversion is necessary.  My conversion tool called
"convf" facilitates this conversion by processing all the files with one
operation.  This tool automatically recognizes binary files as such and
copies them without conversion.  Time stamps and the mode are retained
as they are.  However, a MCPP package contains files used to test a
particular multi-byte character encoding.  The encoding contained in
these files must not be converted.  To achieve this, apply "convf" first
to all the files in the MCPP package to convert only <newlines> in batch,
and then apply this tool again only to the doc directory to convert
Kanji encodings.
Convf itself can be compiled by the compiler system MCPP was ported to.
When you migrate these files from DOS/Windows to other OS environments,
be sure to migrate them in an archives file, and then unarchive and
convert under the new OS.   If they are unarchived under MS-DOS or
Windows95, case distinction of file name will be lost.  Convf is found
in the following location (Unfortunately, the document is in Japanese
only.):

    http://download.vector.co.jp/pack/dos/util/text/conv/code/
                                                    convf-1.8.lzh


            2.  Invocation Options and Environment Settings

The <arg> and [arg] shown below indicate required and optional arguments
respectively.  Note that the <,  >,  [, or ] character itself must not
be entered.


2.1     How to Pass Options to MCPP

Before describing how to specify invocation options, this section
explains how to pass them to MCPP.

MCPP invocation options are specified by do_options() in the MCPP source,
"system.c", however, there is not a generic method because compiler
system-specific settings are required.

For a compiler system I have already ported MCPP to, MCPP does not
necessarily implement all the options that the compiler system-specific
preprocessor provides.  I did not implement the options I thought
unnecessary because they can be easily implemented in do_options() if
necessary.  In addition, some options with the same name as those of
acompiler system-specific preprocessor may behave differently.

A compiler system-specific compiler driver cannot pass some options to
the cpp in a normal manner.  However, GCC V.2.6.2 or later at least
provides the -Wp almighty option to allow you to pass any options to the
cpp.  For example, if you specify as follows:

    cc -Wp,-W31,-Q23

The -W31 and -Q23 options are passed to cpp.  The options you want to
pass to cpp must be specified following -Wp with each option delimited
by ", ". [1], [2]

For other compiler systems, if their compiler driver source is available,
it is recommended that this type of an almighty option should be added
to the source.  If you modify the compiler driver source code in the way
that, for example, when -P<opt> is specified, only -<opt> is passed to
cpp, it would be very convenient because any options can be passed.

tool/lcc.dif is a diff file to modify LSI C-86 V.3.30c/lcc.c.  To modify
it, execute the following command in the directory where lcc.c is
installed:

    patch -c -B_ < lcc.dif

Patch is a standard UNIX command and is implemented in DJGPP or DOS/
WIN32.  MS-DOS 5.0 has a standard command with the same name, but it is
a quiet different command.  Once this patch is executed, only -<opts> of
-P<opts> is passed to cpp.  The -P<opts> option can be used repeatedly.
For the options like cpp -W 31, white-space characters are allowed
between the option character and the argument, however, for the options
like lcc -PW31, they are allowed neither immediately after P nor W.

To use MCPP, install it in the directory where the compiler system-
specific preprocessor should be located under an appropriate name.
Before copying MCPP, be sure to change the name of compiler system-
specific one  so that it may not be overwritten.

On FreeBSD 2.* or later, the cpp is installed in the /usr/libexec
directory.  /usr/bin/cc invokes /usr/libexec/cpp.  There exists another
/usr/bin/cpp, which is a shell-script that calls /usr/libexec/cpp by
adding the -traditional option.  This script seems to be provided for
tools that expect old Reiser model's cpp or illegal tools.  When you
just specify "cpp", /usr/libexec/cpp is invoked.  To avoid the -
traditional option, specify a full path name (/usr/libexec/cpp).  I
recommend that you should retain the original /usr/libexec/cpp (GNU C/
cpp) and change its name to, say, cpp_gnuc, and that you should change
the name of a program that invokes /usr/bin/cpp to /usr/libexec/cpp_gnuc.
[3], [4]

For settings on Linux, FreeBSD, or CygWIN, see also 3.9.5.  For settings
in GNU C 3.x, see 3.9.7.

Note:
[1] -Wa and -Wl are almighty options for assembler and linker,
respectively.  The documentation on UNIX/System V/cc describes these
options.  Probably, GNU C/cc provides the -W<x> option for compatibility.
Note that the GNU C V.2.6.0 had not this option.

[2] In GNU C V.3, cpp was absorbed into cc1 (cc1plus). Therefore, the
options specified with -Wp are normally passed to cc1 (cc1plus).  To
have cpp (cpp0), not ccl, preprocess, the -no-integrated-cpp option must
be specified on gcc invocation.

[3] GNU C V.2.95.3 provides cpp0, as well as cpp, under FreeBSD and
Linux.  It is cpp0 that gcc invokes.  Under VineLinux 2.6, cpp is a link
to cpp0.

[4] Under VineLinux 2.6, cpp is installed in the directory /usr/lib/gcc-
lib/i386-redhat-linux/2.95.3.


2.2    How to Specify Invocation Options

MCPP invocation takes a form of:

    cpp [-<opts> [-<opts>]] [in_file] [out_file] [-<opts> [-<opts>]]

Note that you must replace the above "cpp" with other name, depending on
compiler systems or how MCPP is implemented.

When out_file (an output path) is omitted, stdout is used unless the -o
option is specified.  When in_file (an input path) is omitted, stdin is
used.  A diagnostic message is output to stderr unless the -Q option is
specified.

If any of these files cannot be opened, preprocessing is terminated,
issuing an error message.

MCPP uses getopt() to get an option.

Except for the -M option, missing a required option argument causes an
error.

For an option with arguments, white-space characters may or may not be
inserted between the option character and an argument.  In other words,
both of "-I<arg>" and "-I  <arg>" are acceptable.  For options without
arguments, both of -Qi and -Q -i are valid.

If -D, -U, -I, or -W option is specified several times, each of them is
valid.  For -S, -V, or -+ option, only the first one is valid.  For -2,
or -3 option, its specification switches each time an option is
specified.  For other options, the last one is valid.

Principally, MCPP is case sensitive, but if the command processor (shell)
is case insensitive, so is MCPP. [1]

The switch character is '-', not '/', even under DOS/Windows.

When invalid options are specified, a usage statement is displayed.  To
check valid options, enter a command, such as "cpp -?".  In addition to
the usage message, several error messages are displayed, but they are
self-explanatory.  I will omit their explanations.

Note:
[1] OSes, such as MS-DOS, are not case sensitive, but COMMAND.COM is
case sensitive, so is MCPP.


2.3    General Options

This section covers common options across MCPP modes or compiler systems.

-C
    Output also comments in source code.  I hear this option is required
    when the UNIX lint utility is used.  This option is useful for
    debugging even when the lint utility is not used.  Note that a
    comment is moved ahead of a logical source line when output.  This
    is because a comment is processed before macro expansion or
    directive processing, and a comment may appear during a macro
    invocation.

-D <symbol>[=[<value>]]
-D <symbol (a,b)>[=[<value>]]
    Define a macro named symbol.  This option can be used to change the
    definitions of predefined macros other than __STDC__,
    __STDC_VERSION__,  __FILE__,  __LINE__,  __DATE__,  __TIME__, and
    __cplusplus. (__STDC_HOSTED__, C99's predefined macro, is
    exceptionally included because some compiler systems, like GNU C V.3,
    use the -D option to define __STDC_HOSTED__.)  To specify a value,
    use =<value>.  If =<value> is omitted, 1 is assumed. (Note that in
    bcc or bcc32, the macro is defined as zero-token by default.)  Do
    not enter white-space characters immediately before "=".  If a white-
    space character is entered immediately after "=", the macro is
    defined as zero token.  This option can be specified repeatedly.

-E
    A status of "success" is returned on completion even if an error
    other than fatal occurs.  This option can be used when you want to
    proceed with compiler checking after preprocessing since MCPP treats
    an error as a warning.  This option also provides a means to avoid
    Standard C syntax checking.  This does not affect a diagnostic
    message.  For statuses of "success" and "failed", see 4.1.  For
    causes of fatal errors, see 5.3.

-I <directory>
    Specify the first directory in the include directory search path
    order with <directory>.  For a search path, refer to 4.2.  If a
    directory name contains spaces, it must be enclosed with " and ".


-I 1, -I 2, -I 3
    Specify a directory from which MCPP begins searching when it
    encounters a #include "header" directive.  -I1, -I2 and -I3 indicate
    the current directory, the source file (i.e. includer) directory,
    and the both respectively.  For details, see 4.2.

-j
    On outputing a diagnostic message, MCPP displays only one line of
    diagnostic message without additional information, such as source
    lines.  (By default, one line of diagnostic message is followed by a
    source code line having a problem.  If the source code line in
    question is found in a #included file, all the #including file names
    and including line numbers are also displayed in sequence.  For a
    macro expansion, MCPP displays its development.
    When Validation Suite is used in the GNU C testsuite, this option
    must be specified to output a diagnostic message in the same format
    as GNU C/cpp.

-o <file>
    Output the preprocessed source to a file.  If this option is omitted,
    the second argument ([out_file]) is regarded as an output path, so
    this option is not necessary, however, some compiler drivers use
    this option.  I wonder if the purpose is to prevent erroneous
    expansion of a wild card.  In a case insensitive system, -O <file>
    can be specified.

-P
    Do not output line number information to the compiler-proper.  This
    option is specified when you want to use MCPP for purpose other than
    C reprocessing.

-Q
    Output diagnostic messages to the cpp.err file in the current
    directory.  As these messages are appended to this file, it may
    become bigger.  Delete it from time to time.

-U <symbol>
    Disable predefined macro named symbol. This option cannot disable
    __FILE__, __LINE__,  __DATE__,  __TIME__,  __STDC__,
    __STDC_VERSION__ (and __STDC_HOSTED__ for C99 only), as well as
    __cplusplus invoked with -+ options.

-v
    Output the MCPP version and a search order of include directories to
    stderr.

-W <level>
    Specify a warning level with <level>.  "OR" any one or more values
    of 1, 2, 4, 8 or 16, and set <level> to the "ORed" value or 0.  1, 2,
    4, 8, or 16 indicates a warning class.  For example, if -W 5 is
    specified, warnings of classes 1 and 4 are output.  If 0 is
    specified, no warnings are output.  If this option is specified
    several times, all the specified values are "ORed" together.  For
    example, -W 1 -W 4 is equivalent to -W 5.  If this option is omitted,
    -W 1 is assumed.  For warning messages, refer to 5.5 to 5.9.

-Z
    The preprocessing result of the #included files is not output, but
    macros are defined.


2.4     Options by MCPP Mode

On compiling MCPP itself, by setting various macros in system.H, various
types of preprocessors that behave according to various sets of
specifications can be generated.  A set of specifications is called
"mode" (For mode, refer to 4.1.3 of cpp.doc.  The term "mode" may sound
strange, but I decided to call it a mode since more appropriate terms
did not occur to me.)  All the uppercased names below (including
Chapters 3-5) that do not begin with "__", such as MODE, STDC,
TFLAG_INIT, etc, are macros defined in system.H.  These macros are used
only for compiling MCPP itself and a MCPP executable generated does not
contain these macros.  You must understand this point clearly.  This
section only shows a list of various MCPP behaviors by mode, which may
not readable.  Please be patient.

Note that the values of the PRE_STANDARD, STANDARD, and POST_STANDARD
macros are defined as 0, 3 and 9, respectively.

The following options are available for MCPP compiled with MODE ==
STANDARD :

-3
    Reverse initial settings for the trigraph processing.  With
    TFLAG_INIT == FALSE, MCPP recognizes trigraph.  Otherwise, it doesn'
    t.

-c
    Expand recursive macro more than the Standard's specification.  On
    expanding recursive macro, set the range of non-re-replacing of the
    same name narrower than the Standard.
    Refer to cpp-text.txt section 2.4.26 about the specifications of
    recursive macro expansion.  See test-t/recurs.t for a sample of
    recursive macro. [1].

The following options are available for MCPP compiled with MODE >=
STANDARD:

-2
    Reverse initial settings for the digraphs processing.  With
    DIGRAPHS_INIT == FALSE, MCPP recognizes digraph.  Otherwise, it
    doesn't.

-+
    Behave as C++ preprocessor.  MCPP predefines the __cplusplus macro
    (its value is defined in system.H and defaults to 1), interprets the
    text from // to the end of a logical line as a comment and
    recognizes "::", ".*" and "->*" as a single token.   It evaluates
    true and false tokens in a #if expression to 1 and 0, respectively.
    If __STDC__ and __STDC_VERSION__ are defined, they are deleted.  For
    GNU C ported MCPP, __STDC__ is not deleted for compatibility with
    GNU C/cpp.  The predefined macros that do not begin with "_" are
    also deleted.  However, extended characters are not converted to UCN.
    [2], [3] and [4].

-S <n>
    Change the value of __STDC__ to <n> in C.  In C++, this option is
    ignored.  The range of <n> must be 0-9.  With <n> set to 1 or higher,
    the predefined macros that do not begin with "_", such as unix,
    MSDOS, are disabled.  S indicates __STDC__.  If this option is
    omitted, __STDC__ is set to a default value.  (For LSIC, it is set
    to 0, and for others, it is set to 1).  For a GNU C version, -
    pedantic, -pedantic-errors, or -lang-c89 is equivalent to -S1, so
    the next -S is ignored.

-h <n>
    Define the value of __STDC_HOSTED__ macro with <n>.

    The following options are available for MCPP compiled with MODE !=
    POST_STANDARD and TOP_SPACE == TRUE.

-a
    Accept the following notations used in some assembler sources
    without causing an error.

    1.
      #APP
    Even if a line that begins with # does not agree with any of C
    directives, MCPP outputs this line without causing an error.
    2.
      "A very very
      long long
      string literal"
    The above old-fashioned string literals are concatenated into
    "A very very\nlong long\nstring literal".
    3.  Even if token concatenation using a ## operator generates an
    invalid C pp-token, it is not regarded as error.

    These sometimes happen to GNU source code. [5].

Note:
[1] This option is for compatibility with GNU C, Visual C++ and other
major implementations.  'c' means "compatible mode".

[2] C++'s __STDC__ is not desirable and causes many problems.  GNU C
documents says that __STDC__ needs to be predefined in C++ because many
header files expect __STDC__ to be defined.  The header files should be
blamed for this.  For common parts among C90, C99 and C++,
"#if __STDC__ || __cplusplus" should be used.

[3] Different from C99, the C++ Standard makes much of UCN.  So did C
1997/11 draft.  Half-hearted implementation is not permitted.  However,
implementing Unicode in earnest is too much burden for preprocessor.

[4] In C90 MCPP (MODE >= STANDARD) treats // as a comment but issues a
warning.

[5] This is for compatibility with GNU C.


2.5    General Options Except for Some Compiler Systems

-N
    Disable all the predefined macros, including those that begin with
    "_", except for the ones required by C Standard and __MCPP.  The
    predefined macros stipulated by C Standards include __FILE__,
    __LINE__,  __DATE__,  __TIME__,  __STDC__, and __STDC_VERSION__, as
    well as __STDC_HOSTED__ for C99 and __cplusplus for C++.  The
    purpose of excluding __MCPP is to prevent the -undef option from
    making __MCPP undefined.  Previously, GNU C/gcc used by default the
    -undef option.  If you want to disable __MCPP, use the -U option.
    For the Plan 9/pcc ported MCPP, use -n, instead of -N.

The -V option is available for MCPP compiled with MODE >= STANDARD.

-V <value>
    Change the values of the predefined macros __STDC_VERSION__ for C
    and __cplusplus for C++ to <value>.  <value> is of a long type. (In
    ISO/IEC 9899:1990/Amendment 1:1995, C99, and C++ Standard, this
    value is set to 199409L, 199901L and 199711L, respectively.)  With
    __STDC__  set to 0, __STDC_VERSION__  is always set to 0L,
    overriding the -V option.

    If this option is omitted for C, __STDC_VERSION__ is set to the
    value of STDC_VERSION in system.H. (For GNU C V.2.7 - V.2.9x,
    199409L.  For others, 0L.) If specifying -S1 -V199901L results in
    __STDC_VERSION__ >= 199901L, MCPP conforms to the following C99
    specifications: (See 3.7.)

    1. Treats the text from // to the end of a line as a comment. [3] 2.
    Allows the sequence of p+, P+, p-, and P-, as well as e+, E+, e-,
    and E-, in the preprocessing-number.  This is to represent a bit
    pattern of a floating-point number in Hex, like 0x1.FFFFFEp+128.  3.
    Enables the _Pragma operator (A _Pragma( "foo bar") has the same
    effect as specifying a #pragma foo bar.) 4. MCPP compiled with the
    EXPAND_PRAGMA macro set to TRUE will macro-expands an argument on a
    #pragma line that does not begin with STDC. (By default,
    EXPAND_PRAGMA is set to FALSE in MCPP ported to compiler systems
    other than Visual C, so macro expansion does not occur.) 5. For
    compiler-systems with long long, a #if expression is evaluated in
    long long or unsigned long long.  6. Allows an escape sequence for
    Universal-Character-Name (UCN) in identifiers, character constants,
    string literals and pp-numbers.

    Note that although C99 provides for variable argument macros, MCPP
    allows them in the C90 and C++ modes.[4]

    In C++ also, when specifying -V199901L results in __cplusplus >=
    199901L, MCPP will enter the C99 compatibility mode, providing the
    above 2-5 enhancements.  (1 is enabled unconditionally and 6 is
    almost the same.)  These are MCPP's own enhancements that do not
    conform to the C++ Standard.

    On a case insensitive system, this option is -T.  For the Plan 9/pcc
    ported MCPP, use -s, instead of -V.  The -D option cannot be used
    with __STDC__, __STDC_VERSION__, and __cplusplus.  This is to
    distinguish system-defined macros from user-defined ones.

-I-
    Cancel default include directories and enable only ones specified
    with an environment variable and the -I option.  Instead of -I-, GNU
    C ported MCPP uses -nostdinc.  In GNU C, the -I- option provides
    quite different functionality.  See 2.6.

The -m option is available for the MCPP ported to 32 bits or more system.

-m <encoding>
    Change a multi-byte character encoding to <encoding>. For <encoding>,
    refer to 2.8.

The following options are available for MCPP compiled with OK_MAKE ==
TRUE.

When there are several source files and the -M* option is specified for
each of these source files to process and merge the outputs into a file,
dependency description lines are aligned. These dependency description
lines are used for a makefile.  These options are similar to those of
GNU C/cpp, but there are several differences. [1]

-M
    Output lines that describe dependency among source files.  The
    output destination is the file specified in a command line, or
    stdout if omitted.  If a dependency description is too long to fit
    in a line, it is folded over the next lines.  The preprocessing
    result is not output.
-MM
    Almost the same with -M, except that the following header files are
    not output.

    1. Files specified in the format of #include <stdio.h>
    2. Files specified using an absolute path name, such as #include
    "/include/stdio.h".
    3. Files specified in the format of #include "stdio.h" that are
    found not in the current or source directory, depending on compiler
    systems or the -I <n> option, but in system include directories,
    including those specified with the -I <directory> option or with
    environment variables.
-MD [FILE]
    Almost the same with -M, except that the preprocessing result is
    output to the specified file on a command line or stdout.  If [FILE]
    is specified, MCPP outputs dependency description lines to that file.
    Otherwise, they are output to a file having the same base filename
    with the source file and the suffix of ".d" instead of ".c".
-MMD [FILE]
    Almost the same with -MD, except that, like -MM, the files that are
    regarded as system header are not output.  An output file MCPP
    outputs dependency description lines to is same as -MD [FILE].
-MF FILE
    The preprocessing result is output to FILE.  -MF FILE takes
    precedence over -MD FILE or -MMD FILE.
-MP
    "Phony targets" are also output.  Each included file can be written
    as a phony target without a dependency as follows:
        test.o: test.c test.h
        test.h:
-MT TARGET
    The target name is specified as TARGET not foo.o.  -MT
    '$(objpfx)foo.o' outputs the following line.
        $(objpfx)foo.o: foo.c
-MQ TARGET
    Same as -MT, except that a string that has a special meaning to Make
    is quoted as follows:
        $$(objpfx)foo.o: foo.c

Note:
[1] MCPP differs from GNU C/cpp in that:
1. MCPP does not provide the -MG option because its option
specifications are too complicated. (Therefore, I will omit their
explanations.)  The -M option can substitute for the -MG option because
when include files cannot be found using the -M option, MCPP fails but
outputs dependency description lines.
2. MCPP excludes a wider range of header files when using the -MM and -
MMD options.  The GNU C 2/cpp does not exclude the header files shown in
2 and 3 of the -MM option.  The GNU C 3/cpp0 now excludes the header
files shown in 3 that are found in the system header directory.
4.4BSD-Lite has a /usr/bin/mkdep command.  The equivalent command in
FreeBSD or Linux is a shell-script that generates several *.d files and
merges them into a file named ".depend".  This shell-script invokes cpp
-M.


2.6     Options by Compiler System

The following options are available for MCPP ported to GNU C and LSI C-
86.

-b
    Output line number information just like C sources.

    In LSI C, when MCPP passes a #pragma directive to the compiler-
    proper, it normally shortens the spelling to #p.  With this option
    specified, MCPP outputs a #pragma as it is, as long as it is the one
    MCPP itself does not process.  On a case insensitive system, this
    option is -B.

    The format used to pass the line number information from a
    preprocessor to its compiler-proper is usually as follows:

      #line 123 "filename"

    Most compiler systems can use this C source format, but some systems,
    such as LSI C-86, cannot.  The default specification of MCPP is such
    that, when MCPP is ported to the compiler systems that cannot use
    the C source format, MCPP outputs the line number information in a
    format that the compiler-proper can accept it.  In GNU C, MCPP
    outputs this information in the GNU C specific format because,
    whilist the compiler-proper can accept this C source format, some
    cpp versions cannot accept the C format when it is given again, and
    because some tools, like rpcgen, can accept only the GNU C specific
    format.

    However, with this option specified, MCPP ported to the compiler-
    systems that do not accept the C source format outputs the line
    number information in that format.  This option is used with #pragma
    __put_defines (#put_defines) to pre-preprocess a header file.

The following options are available in compiler-systems under MS-DOS.

-m <x>
    Specify a memory model with <x>.  <x> is either t, s, d, p, l, or h
    for LSI C or Lattice C, and is either t, s, c, m, l, or h for other
    compiler systems.  These letters are not case sensitive.  With this
    option specified in Microsoft C, the M_I86mM macro with the "m"
    replaced with the uppercased letter x is pre-set to 1.   In other
    compiler systems, of __TINY__, __SMALL__,  __COMPACT__, __MEDIUM__,
    __LARGE__, and __HUGE__, an appropriate macro is preset to 1.  MCPP
    compiled with OK_SIZE == TRUE evaluates a #if sizeof (type) based on
    this memory model specification.  By default, M_I86SM or __SMALL__
    is set to 1.

The following options are available for Visual C or LSI C-86 ported MCPP.

-w
    Same as -W0.

The following options are available for the LCC-Win32 ported MCPP.

-g <n>
    Define the __LCCDEBUGLEVEL macro as <n>.

-O
    Defines the __LCCOPTIMLEVEL macro as 1.

The following options are available for the Visual C ported MCPP.

-X
    Same as -I-.

-Fl <file>
    Same as -include <file> for GNU C.

-u
    Same as -N.

-Tp
    Same as -+.

-Tc
    Specify that the source is written in C.  The result is same with or
    without this option.

-Wall
    Same as -W17 (-W1 -W16).

-WL
    Same as -j.

The following option is available for the Plan 9/pcc ported MCPP.

-i
    Output a search order of include directories to stderr,

-s
    Same as -V for MCPP ported to other compiler systems.

-n
    Same as -N for MCPP ported to other compiler systems.

The following options (until at the end of the 2.6 subsection) are
available for the GNU C ported MCPP.

Note that since __STDC__  is set to 1 for GNU C, the result is same with
or without the -S1 option.

-trigraphs
    Recognize trigraph (For MCPP compiled with MODE == STANDARD).
    Trigraph specification is also reversed by -3.
-digraphs
    Recognize digraph.  Digraph specification is also reversed by -2.

-nostdinc
    Same as -I- for other compiler systems.
-I-
    Switch the specification of the -I <directory> before and after this
    option; directories specified with the -I options before -I- are
    used to search for header files only in the form of #include
    "header.h"; the directories specified with -I after -I-, if any, are
    used to search for all #include directives.  In addition, during the
    former search, includer's directories are not used.

-pedantic, -pedantic-errors
    Same as -S1.
-Wcomment, -Wcomments
    Same as -W1.  The result is same with or without this option.
-Wtrigraphs
    Same as -W16.
-Wall
    Same as -W17. (With -Wall, MCPP does not issue class 2 and 4
    warnings because these warnings are issued frequently and annoying
    for GNU C standard header files.  Class 8 warnings are generally
    surplus and bothering, but are helpful to confirm portability and
    etc.  When this option is used, be sure to specify gcc -Wp,-W31.)

-undef
    Same as -N.

-lang-c, -x c
    Perform C preprocessing.  The same as not specifying this option at
    all.
-lang-c89, -std=c89, -std=gnu89
    Same as -S1.  Not only C90 but also C95 specifications are used.
    The result is same with or without this option.
-lang-c99, -lang-c9x, -std=c99, -std=c9x, -std=gnu99, -std=gnu9x
    Same as -S1 -V199901L.
-lang-c++, -x c++
    Perform C++ preprocessing.  Same as -+.
-lang-asm, -x assembler-with-cpp
    Same as -a.

-std=iso<n>:<ym>
    Specify a version of C Standard.  To specify C, <n> is 9899 and C++,
    14882.  If <n> is 9899, <ym> is any of 1990, 199409,1999 and 199901.
    If <n> is 14882, <ym> is 199711.  If you enter other value than
    these in <ym>, __STDC_VERSION__  or __cplusplus is set to that value.
    <ym> must be specified in six digits, like 199901.

-dD, -dM
    Output valid macro definitions in the form of #define lines at the
    end of processing.

    With the -dD option specified, the preprocessing result is output.
    Predefined macros are output as comments. With the -dM option
    specified, the preprocessing result is not output, but only the
    standard predefined macros are output as comments.

-include <file>
    #include the <file> before processing the main source file.  This is
    equivalent to writing #include <file> at the beginning of the source
    code.
-isystem <dir>
    Add <dir> to the include path immediately before system-specific
    directories and immediately after site-specific directories.

-A <predicate(answer)>
    MCPP neither makes this option an error nor does anything about it.
    In GNU C, this option is equivalent to writing #assert <predicate
    (answer)> in the source code.  Standard C, does not accept extensive
    directives other than #pragma.  Fortunately, so far, gcc, by default,
    passes an equivalent macro with the -D option, so there are no
    actual problems unless a source program uses #assert, which is a
    rare case.

-traditional
    This option is valid for MCPP compiled with both of
    COMMENT_INVISIBLE and STRING_FORMAL set to TRUE. (Same as not
    specifying it at all.)  In other modes, this option causes an error.

MCPP neither makes the following options an error nor does anything
about them (It sometime issues a warning.)

-$
-g <n>
-iprefix <dir>, -iwithprefix <dir>, -iwithprefixbefore <dir>
-idirafter <dir>
-remap
-noprecomp

Other than the above, MCPP provides additional compiler system-specific
options.  Refer to usage statements.


2.7    Environment Variables

The system include directories MCPP does not set up by default must be
specified using environment variables.  For default system include
directories, refer to 4.8.  For the search order and environment
variable names, refer to 4.2.

Auxiliary system include directories to be used vary depending on Linux/
GNU C versions.  So environment variables must be set up for these
directories.

Different GNU C versions use different auxiliary system include
directories.  The auxiliary system include directories which are not
specified in noconfig.H or config.h must be specified with environment
variables.

In addition, when recompiling of glibc 2 programs, the environment
variable GCC_VERSION is necessary.  To know the value of GCC_VERSION,
use gcc -v or cpp -v.

For example, on Vine Linux 2.5 (i386)/GNU C 2.95.3, if you use bash,
write the following code, say, in ~/.bash_profile.

    C_INCLUDE_PATH="/usr/lib/gcc-lib/i386-redhat-linux/2.95.3/include"
    GCC_VERSION="\"2.95.3 20010315 (release)\""
    export C_INCLUDE_PATH GCC_VERSION

On FreeBSD, if you use tcsh, write the following code, say, in ~/.login.

    setenv GCC_VERSION '"2.95.3 20010315 (release)"'

Normally, FreeBSD uses include directories "/usr/include" and
"/usr/local/include" only, so the C_INCLUDE_PATH setting is not
necessary.  However, if a GNU C version other than the system standard
one is installed, the C_INCLUDE_PATH setting may be required.

For the environment variable LANG, refer to 2.8.


2.8       Multi-Byte Character Encodings

MCPP is enabled for various multi-byte character encodings.  Encodings
that can be used are different between 16-bit system ported MCPP (herein
after called 16-bit MCPP) and 32-bit or more system ported MCPP (herein
after called 32-bit or more MCPP).  There are limits to the encodings 16-
bit MCPP can implement because less memory is available.

The encodings common across 16-bit and 32-bit or more MCPP are as
follows:

    EUC-JP:     Japanese extended UNIX code (UJIS)
    shift-JIS:  Japanese MS-Kanji
    GB-2312 :   EUC-like Chinese encoding (Simplified Chinese)
    Big-Five:   Taiwanese encoding (Traditional Chinesej
    KSC-5601:   EUC-like Korean encoding (KSX 1001)

16-bit MCPP can implement any one of the above encodings.

The following encodings are also available for 32-bit or more MCPP.

    ISO-2022-JP1:   International standard Japanese
    UTF-8:      A kind of unicode encoding

32-bit or more MCPP implements all these encodings at the same time.
The encoding used during execution can be specified as follows (Priority
is given in this order):

  1.  The encoding specified in #pragma __setlocale( "<encoding>") in
    source code. (For Visual C ported MCPP, #pragma setlocale
    ( "<encoding>").)  This directive allows you to specify several
    encodings in one source file.
  2.  The encoding specified with -m <encoding> as run-time option.
  3.  The encoding specified with the LC_ALL, LC_CTYPE and LANG
    environment variables.  Priority is given in this order.
  4.  The default encoding specified when MCPP is compiled.

How to specify a <encoding> is basically same across #pragma __setlocale,
-m, and the environment variables; the encoding on the left-side hand is
specified by the <encoding> on right-hand side; <encoding> is not case
sensitive; '-', '_' and '.' are ignored.  Therefore, EUC_JP, EUC-JP,
EUCJP, euc_jp, euc-jp, eucjp and euc.jp are regarded as same.  '*'
represents any character sequence of zero or more bytes.(ja_JP.utf8 and
en_US.utf8 are equivalent to *utf8.).

    EUC-JP:     eucjp, euc, ujis, jajpuec, jajpeucjp
    shift-JIS:  sjis, shiftjis, mskanji, jajpsjis
    GB-2312:    gb2312, cngb, euccn
    BIG-FIVE:   bigfive, big5
    KSC-5601:   ksc5601, ksx1001, euckr, wansung
    IS0-2022-JP1:   iso2022jp, iso2022jp1, jis, jajpjis
    UTF-8:      utf8, utf, jajputf8
    Not specified:  c, en*, latin*, iso8859*

If any of the following encodings is specified, MCPP is no longer able
to recognize multi-byte characters: C, en* (english), latin* and iso8859
*.  When a non-ASCII ISO-8859 Latin-n single-byte character set is used,
one of these encodings must be specified.  When an empty name is used (#
pragma __setlocale( "")), the encoding is restored to the default.

Only in the Visual C++ ported MCPP, the following encoding name can be
specified with #pragma setlocale .  This is for compatibility with
Visual C++.  It is recommended you should use these names because the
Visual C++ compiler cannot recognize encoding names other than these.
('-' can be omitted for MCPP, but not for the Visual C++ compiler-proper.)

    shift-JIS:  japanese, jpn
    GB-2312:    chinese-simplified, chs
    BIG-FIVE:   chinese-traditional, cht
    KSC-5601:   korean, kor
    Not specified:  C, english

In Visual C++, the default multi-byte character encoding varies,
depending on what language the language parameter and "Region and
Language Option" of Windows are set to.  However, the #pragma setlocale
specification takes precedent over these Windows's settings.

Only in the GNU C ported MCPP, the following encoding names can be
specified with the environment variable LANG.  This is for compatibility
with GNU C.  It is recommended that you should use these names because
the GNU C compiler cannot recognize encoding names other than these.
('-' can be omitted and lowercase letters can be used for MCPP, but for
the GNU C compiler-proper, these names must be specified exactly as
shown below.)

    EUC-JP:         C-EUC
    shift-JIS:      C-SJIS
    ISO-2022-JP1:   C-JIS
    Not specified:  C

Depending on the configuration used to compile the GNU C compiler, the
GNU C compiler sometimes recognizes the environment variable LANG's C-*
specification and sometimes not.  [1]  When the compiler fails to
recognize it, MCPP complements it.

Note:
[1] If the --enable-c-mbchar option is specified when a configure script
is used to compile GNU C, an encoding can be specified using an
environmental variable, such as LANG.  This way of compilation seems to
be available from July 1998 onward, but its implementation is still half-
hearted and has some problems at least at V.3.2.  It is documented that,
besides LANG, environmental variables, such as LC_ALL and LC_CTYPE, can
be used to specify an encoding.  However, the difference between using
LC_ALL and LC_CTYPE or not lies only in their diagnostic messages.  The
same thing can be said with V.2.95 which is implemented in Vine Linux 2.
6 and FreeBSD 4.7.


2.9       How to Use MCPP in One-Path Compilers

Compilers whose preprocessor is integrated into themselves are called
one-path compilers.  These includes Visual C, Borland C, and LCC-Win32.
Such compilers are becoming more popular because they can achieve a
higher processing speed.  However, the time for preprocessing becomes
shorter due to better hardware performance.  In the first place, there
is much point for preprocessing to be a common phase, mostly independent
of run-time environment and compiler systems. It is not desirable that
one-path compilers become more popular.  There will be more compiler
system-specific specifications.

Anyhow, it is impossible to replace the preprocessor of a one-path
compiler with MCPP.  To use MCPP, a source program is preprocessed with
MCPP and then the output is passed to a one-path compiler.  As you see,
preprocessing takes place twice.  It is useless but inevitable.

To use MCPP with a one-path compiler, the procedure must be written in
makefile.  For sample procedures, refer to the makefile recompilation
settings used to compile MCPP itself, such as visualc.mak, borlandc.mak,
and lcc_w32.mak.

Although GNU C 3 now integrates its preprocessing facility into itself,
it provides an option to use an external preprocessor.  Use this option
when MCPP is used. (See 3.9.7.)

2.10       How to Make MCPP Available in IDE

It is difficult to use MCPP in Integrated Development Environment (IDE)
because IDE's GUI follows compiler-system specific specifications and
internal interfaces are not usually made available to third parties.
Furthermore, one-path compilers make it more difficult to insert a phase
in which MCPP is used into IDE.

This subsection describes how to make MCPP available in Visual C++ .net'
s IDE.  I have only version 4 of Borland C++ IDE, which is too old to do
so.  I think I can do the same for LCC-Win32's IDE because LCC-Win32 is
shareware, although it may take time.  I have not tried it.  To use
Borland C and LCC-Win32 ported MCPP, use command lines.

2.10.1      How to Make MCPP Available in Visual C++ .net's IDE

MCPP cannot be used in a normal "project" since the internal
specifications of Visual C++'s IDE are not made available to third
parties and the compiler is a one-path compiler.  However, once a
makefile that specifies MCPP is created, Visual C++'s IDE can recognize
the makefile and you can create a "makefile project" using that file.
This allows you to utilise most of the IDE functions, including source
edition, search, and debugging.

"Creating a Makefile Project" of a Visual C++ .net 2003 document
describes how to make a makefile project.  Perform the following
procedure to create a makefile project. [1]

  1. Log in as a user with debugging privilege [2]
  2. Create a makefile that specifies MCPP. (Refer to visualc.mak.)
  3. Start Visual Studio .net. [3]
  4. Click "New Project" to display the "New Project" window.  Select
    "Makefile Project" and specify "Name" and "Location", and then click
    "OK".
  5. Then the "Makefile Application Wizard" windows appears.  Click
    "Application settings", and enter appropriate values in the
    "Build command line", "Output", "Clean commands", and
    "Rebuild command line" fields.

    Let me explain the appropriate values for these fields by taking an
    example of compiling MCPP itself and assuming that a MCPP executable,
    cpp32_std.exe, is generated.

        "Build command line":   nmake
        "Output":               cpp32_std.exe
        "Clean command":        nmake clean
        "Rebuild command line": nmake PREPROCESSED=1 CPP=cpp32_std

    Since a Makefile project does not provide a 'make install'
    equivalent command, you must write the makefile in such a way that
    the commands you specify in "Build command line" and
    "Rebuild command line" also perform installation.

    If you do not compile MCPP, "Build command line" and
    "Rebuild command line" can be the same.

    When completed, click "Finish".

  6.  Then the Makefile project appears in "Solution Explorer".  Click
    the "Source Files" folder, choose "Add Existing Solution Item" from
    the "Project" menu, select all the source files, and then click "OK".
    Then the source file names appear in Solution Explorer.

You can now use every functions, including Edit, Build, Rebuild and
Debugging.

Note:
[1]  This procedure worked properly under Windows 2000.  However, for
unknown reasons, it did not under Windows XP HE.  When I copied the
project file generated under Windows 2000 to Windows XP, this project
file also worked properly under Windows XP.
[2]  To use the debugging function under Windows 2000, a user must
belong to a group called "Debugger users".  However, Windows XP HE does
not provide such a group, so one must log in as administrator.  In
addition, in order to perform the source level debugging function,
makefile must be written in such a way that cl.exe is called with the
/Zi option appended to generate debugging information.
[3]  If you start Visual Studio .net by selecting "Start" -> "Programs",
environment variables, such as for include directories, are not set when
a built executable is executed for debugging.  In order to have these
variables set, you must open the Visual Studio .NET command prompt to
start Visual Studio. net by typing:
    devenv <Project File> /useenv


                   3.  Enhancements and Compatibility

MCPP has its own enhancements.  Each compiler system-specific
preprocessor has its own enhancements, most of which are not available
in MCPP.  This section covers these enhancements and their compatibility
problems.

Principally, MCPP compiled with HAVE_PRAGMA == TRUE outputs #pragma
lines as they are.  This principle is applied to the #pragma lines
preprocessed by MCPP itself.  This is because the compiler-proper may
interpret the same #pragma in a different way.  However, MCPP outputs
neither #pragma __put_defines nor #pragma __once to avoid duplication of
preprocesses when the header "pre-preprocessing" facility is used.  The
pre-preprocessing facility reprocesses MCPP output.  For pre-
preprocessing, refer to 3.1.    MCPP outputs neither #pragma
__push_macro, #pragma __pop_macro, #pragma push_macro, nor #pragma
pop_macro because they are useless on the later phases.  GNU C ported
MCPP does not output a line of #pragma GCC followed by either poison,
dependency or system_header.

MCPP compiled with HAVE_PRAGMA == FALSE does not output #pragma lines
and issues a warning for those not preprocessed by MCPP itself.


3.1    #pragma __put_defines,
                #pragma __preprocess, #pragma __preprocessed,
                #put_defines, #preprocess, #preprocessed

MCPP compiled with MODE >= STANDARD and MODE == PRE_STANDARD uses #
pragma __put_defines and #put_defines, respectively.  Let me explain by
taking an example of #pragma.

When MCPP encounters #pragma __put_defines directive, it outputs all the
macros defined at that time in the form of #define lines.  Of course,
the #undefed macros are not output.  The macros that cannot be #defined
or #undefed, such __STDC__ and etc, are output in the form of #define
lines, but are enclosed with comment marks.  (Since __FILE__ and
__LINE__ are special macros defined dynamically on a macro invocation,
the replacement list output here means little.)

MCPP compiled with MODE != STANDARD does not memorize parameter names of
function-like macro definitions.  So, these directives mechanically
represent names of the first, second, third parameters as a, b, c, ...
and so on.  If it reaches the 27th parameter, it begins with a1, b1, c1,
..., a2, b2, c2, ... and so on.

If you enter the following directive after invoking MCPP from keyboard
without specifying input and output files, all the predefined macros are
listed.

    #pragma __put_defines

If you invoke MCPP with options such as -S1 or -N, you will see a
different set of predefined macros.

When the MCPP compiled with DEBUG == TRUE encounters with the following
directive:

    #pragma __put_defines

it also outputs a comment to indicate the source file name where each
macro definition is found, as well as its line number.

When such MCPP encounters the following #pragma __preprocess directive,
it outputs the following line:

    #pragma __preprocessed

This indicates that the source file has been preprocessed.

When MCPP encounters a #pragma __preprocessed directive, it determines
that the source file has been preprocessed by MCPP and continues to
output the code it reads as it is, until it encounters a #define line.
When MCPP does encounter a #define directive, MCPP determines that the
rest of the source file are all #define lines and defines macros.  [1]

At this time, the MCPP compiled with DEBUG == TRUE would memorize the
source filename and line number in the comment.  [2]

A #pragma __preprocessed is applied only to the lines that follow the
directive in the source file where the #pragma __preprocessed directive
is found.  If the source file is an #included one, when control is
returned to the #including file, #pragma __preprocessed is no longer
applied.

Note:
[1]  Actual processing is a little more complex.  When MCPP encounters a
#pragma __preprocessed, MCPP outputs lines it has read just as they are,
except for #line lines, which MCPP converts and outputs into a format
that the compiler-proper can accept.  MCPP disregards predefined
standard macro because its #define line is enclosed with comment marks.
[2]  Therefore, information on where a macro definition is found is not
lost during pre-preprocessing.  This means, in turn, the MCPP compiled
with DEBUG == FALSE cannot use this information.

3.1.1   Pre-Preprocessing Header File

With these directives, you can pre-preprocess header files.  Pre-
preprocessing considerably saves the entire preprocessing time.  I think
the explanation so far has already given you an understanding of how to
pre-preprocess header files, but to deepen your understanding, let me
explain it by taking an example of MCPP's own source code.

MCPP source consists of seven *.c files, of which six files include
"system.H" and "internal.H".  No other headers are included.  The source
looks like this:

#if PREPROCESSED
#include  "cpp.H"
#else
#include  "system.H"
#include  "internal.H"
#endif

The system.H includes noconfig.H and configed.H, as well as several
standard header files.  cpp.H is not a source file I provide and is a
"pre-preprocessed" header file I am going to generate.

To generate cpp.H (Of course, after setting up system.H and other
headers), invoke MCPP as follows:

    cpp -S1 > cpp.H

For compiler systems, such as GNU C and LSI C, also specify the -b
option.  The -S1 option can be omitted, but in this case, specifying -S1
would make the processing more efficient.

Enter the following directives from the keyboard:

#pragma __preprocess
#include "system.H"
#include "internal.H"
#pragma __put_defines

Enter "end-of-file" to terminate MCPP.

This has accomplished cpp.H, which consists of the preprocessed system.H
and internal.H and a set of #define lines following them.  Including cpp.
H gives the same effect as including system.H and internal.H, but its
size is one-nth of the original header files containing standard ones.
This is because #if and comments are eliminated.  It takes far less time
to include cpp.H in seven or eight *.c files than to include system.H
and internal.H seven or eight times.  By using #pragma __preprocess,
much more time can be saved.

On compilation, use the -DPREPROCESSED=1 option.

It is recommended that the above procedure should be written in a file
and the makefile should refer to it. The makefile and preproc.c appended
to MCPP sources contain the procedure.  Please refer to it.

Although it is difficult to find a way to use MCPP ported to one-path
compilers like Visual C, Borland C or LCC-Win32 in an independent manner,
the pre-preprocessing facility is useful.

The pre-preprocessing facility of header files is similar to that of the
-dD option of GNU C/cpp, but it differs from it in that:

1. GNU C/cpp outputs line number information not in the form of #line
123 "filename", but in the form of # 123 "filename", which allows GNU C/
cpp to reprocess the information, but the Standard C preprocessor cannot.

2. GNU C/cpp of older version outputs a #define line whenever it
encounters it, but does not output a #undef line.  Therefore,
reprocessing the preprocessed result may produce a different result from
what the original source intends.

3. By using #pragma preprocess, which is not provided by GNU C, MCPP can
provides a higher processing speed.

As far as the pre-preprocessing facility is concerned, MCPP is more
accurate and practical than GNU C/cpp.


3.2    #pragma __once, #pragma once

#pragma __once directive is implemented when MCPP is compiled with MODE >
= STANDARD, that is, __STDC__ is preset to 0 or higher.

#pragma once is also available for GNU C, Visual C and LCC-Win32.

These directives is used when you want to include a header file only
once.  With the following directive in a header file, MCPP includes the
header file only once even if a #include line for that file appears many
times.

#pragma __once

Usually, compiler system-specific standard header files prevent
duplicate definitions by using the following code:

#ifndef __STDIO_H
#define __STDIO_H
/* Contents of stdio.h */
#endif

#pragma __once provides similar functionality to this.  Using macros
always involves reading a header file. (The cpp cannot skip reading the
code as people do and must read the entire header file for #ifs or #
endifs; It must read a comment before it can determine whether a line is
a control line, that is, a line with # at the beginning followed by a
preprocessing directive; To do so, the cpp must identify a string
literal; After all, it must read through the entire header file and
perform most of tokenization. ) #pragma __once eliminates the need of
even accessing to a header file, resulting in a improved processing
speed for multiple includes.

To determine whether two header files are identical, file name
characters, including directory names in a search path, are compared.
DOS/Windows is not case sensitive.  Therefore, "/DIR1/header.h" and
"/DIR2/header.h" are regarded as distinct, but "header.h" and "HEADER.H"
are regarded as the same on DOS/Windows, but distinct on UNIX-like
systems.

I borrowed the idea of #pragma __once from GNU C V.1.*/cpp.  GNU C V.2.*,
and V.3.*/cpp still has this functionality but it is regarded as
obsolete.  The specification of GNU C V.2.*/cpp has been changed as
follows: If the entire header file is enclosed with #ifndef _MACRO, #
define _MACRO, and #endif, the cpp memorizes it and inclusion occurs
only once, even without #pragma once.

However, this GNU C V.2 and V.3/cpp specification sometimes does not
work for commercially available compiler systems that are not based on
the GNU C specification, due to a difference in the standard header file
notation.  In addition, the GNU C V.2 and V.3/cpp specification is more
complex to implement.  For this reason, I decided to implement only #
pragma __once.

As with other preprocessors, it is not advisable to rely only on #pragma
__once when the same header files are used.  It is recommended that #
pragma __once should be combined with macros as follows:

#ifndef __STDIO_H
#define __STDIO_H

#if   __MCPP >= 2
#ifdef __STDC__
#pragma __once
#else
#ifdef __cplusplus
#pragma __once
#endif
#endif
#endif

/* Contents of stdio.h */
#endif

Assuming a pre-ANSI preprocessor is also used, the above code fragment
uses #if, #else, and #ifdef only, and does not use defined operators.
If the pre-ANSI preprocessor is not used at all, the following one-line
coding is enough:

#pragma __once

Note that #pragma __once must not be written in <assert.h>.  For the
reason, see 4.1.2 of cpp-test.txt.  The same thing can be said with
<cassert> and <cassert.h> of C++.

Another problem is that the recent GNU C/glibc system has header files,
like <stddef.h>, which are repeatedly #included by other system headers.
They define  macros, such as __nead_NULL, __nead_size_t, and
__need_ptrdiff_t, and then #include <stddef.h>.  Each time they do so,
definitions such as NULL, size_t, and ptrdiff_t are defined in the
<stddef.h>.  The same thing can be said with <errono.h> and <signal.h>,
and even with <stdio.h>.  Other system headers define macros, such as
__need___FILE, and then #include <stdio.h>.  Each time they do so,
definitions such as FILE may be defined in <stdio.h>.  #pragma __once
can not be used in such header files. [1]

Note:
[1] This is applied at least to Linux/GNU C 2.9x and 3.2/glibc 2.1 and 2.
2.  FreeBSD 4.* has much simpler system headers because it does not use
glibc.

3.2.1   Tool to Write #pragma __once to Header Files

With a small number of header files, writing #pragma __once to them does
not require much effort, but it would be tremendous work if there are
many header files.  I wrote a simple tool to write it automatically to
header files.

tool/ins_once.c is a tool written for relatively old versions of GNU C.
As Borland C 4.0 and LSI C-86 conform to the same standard header file
notation with GNU C, this tool can be used.  However, it is advisable
that this tool should not be used in the systems like Glibc 2 that has
many exceptions shown above.

Even in the compiler systems that can use the tool, some header files do
not strictly conform to the GNU C notation.  Deviations are often seen,
in particular, in DJGPP headers or FreeBSD modified versions for PC-98.
GNU C V.2.*/cpp's read-once functionality does not work properly for
these header files.

Compile ins_once.c and perform the following command in a directory,
such as /usr/include or /usr/local/include, under Unix.

    chmod u+w *.h */*.h */*/*.h

and then execute ins_once as follows:

    ins_once -t *.h */*.h */*/*.h

Ins_once reports header files that do not begin with #ifndef or #if !
defined.  Manually modify these files.  Then, execute ins_once as
follows:

    ins_once *.h */*.h */*/*.h

If the first #directive in each header file is #ifndef or #if !defined,
a #pragma __once line is inserted immediate below the line.  Only a root
user or a user with an appropriate permission is eligible for this
modification. When you modified your access permission, use chmod u-w *.
h */*.h */*/*.h to restore to your original permission.

Ins_once provides the following options.  Select the most appropriate
one for your system.

    -t:  Check whether a header file begins with #ifndef or #if !defined,
        excluding a comment.  This option does not modify the file.
    -p:  Insert a #pragma __once line at the beginning of file.  By
        default, this line is inserted immediate below the #ifndef or #
        if !defined line.
    -o:  Generate the nine lines shown above in preparation for
        preprocessors that cannot accept #pragma.  By default only one #
        pragma __once line is inserted.
    -g:  For a GNU C new system, <stddef.h>, <stdio.h>, <signal.h>,
        <errno.h> are also excluded.  By default, only <assert.h>,
        <cassert> and <cassert.h> are excluded.

ins_once roughly checks to write a #pragma __once line only once in the
same header file even if it is executed several times, but the check is
not very strict.  As this ins_once is of temporary and tentative nature,
it scarcely performs tokenization.  It worked as I expected with FreeBSD
2.0, 2.2.2,2.and 2.7, DJGPP 1.12, Borland C 4.0, and LSI C-86 3.30c, but
it may not work properly for special header files.  So before executing
this tool, be sure to make a backup of an original file.

Have the shell expand a wildcard. (In case of buffer overflow, execute
ins_once several times by specifying some of your system header files.)
In LSI C, have the start up routine expand a wildcard. (Link to the
default startup or expand.obj.)

For use with OS-9/Microware C/09, I have modified a set of standard
header files to insert #pragma __once and uploaded them to NIFTY-SERVE/
FC/LIB 2/header09.lzh.


3.3    #pragma __include_next, #pragma __warning_cpp,
                #include_next, #warning

These directives are provided for compatibility with GNU C/cpp.  GNU C/
cpp provides the #include_next and #warning directives.  Although these
directives are illegal, not only some source programs sometimes use them
but also some Glibc2 standard header files do.  Taking this situation
into consideration, I decided to implement the #include_next and #
warning directives in GNU C ported MCPP to allow compilation of such
source programs, however, MCPP compiled with MODE >= STANDARD issues a
warning when it finds the directives.  Regardless of the compiler
systems MCPP is ported to, MCPP compiled with MODE >= STANDARD also
implements #pragma__include_next and #pragma __warning.

With following directives, MCPP skips the directory where the first
header.h is found and includes the second header.h found during include
directory search.

    #pragma __include_next <header.h>

or

    #include_next  <header.h>

DOS/Windows ignores the distinctions of alphabetical case of header
names.

The following code outputs any message to stderr as a warning message:

    #pragma __warning_cpp  any message
    #warning  any message

Different from #error, this does not cause an error.


3.4     #pragma __push_macro, #pragma __pop_macro,
                #pragma push_macro, #pragma pop_macro
                #pragma __setlocale, #pragma setlocale

When I ported MCPP to Visual C, I implemented these directives in MCPP,
and then made them available for other compiler system ported MCPP.

#pragma __push_macro ("MACRO") and #pragma __pop_macro ("MACRO") are
used to "push" or "pop" a macro definition (MACRO) to the current macro
definition stack.

#pragma push_macro("MACRO") and #pragma pop_macro("MACRO") are also
available for Visual C.

push_macro saves a macro definition to the stack, and pop_macro
retrieves the macro definition.   The pushed macro definition remains
valid after push_macro.  To invalidate it, use #undef or redefine the
macro with a new definition.  push_macro can be used many times.

#pragma __setlocale( "<encoding>") changes the current multi-byte
character encoding to <encoding>.  The argument of setlocale must be a
string literal.  For <encoding>, refer to 2.8.  This directive allows
you to use several encodings in one source file.

In Visual C++, #pragma __setlocale cannot be used.  Use #pragma
setlocale instead.  Encoding specification must be conveyed not only to
MCPP but also to the compiler-proper.  The latter can recognize only #
pragma setlocale.  For other compiler systems, when the compiler-proper
cannot recognize an encoding, MCPP complements it.


3.5    #pragma __debug_cpp, #pragma __end_debug_cpp,
                #debug, #end_debug

MCPP compiled with MODE >= STANDARD implements #pragma __debug_cpp and #
pragma __end_debug_cpp.  MCPP compiled with MODE == PRE_STANDARD
implements #debug and #end_debug.  If MCPP is compiled with DEBUG ==
TRUE and/or DEBUG_EVAL == TRUE, the corresponding debug information is
output.

The #pragma __debug_cpp <args> directive can be written anywhere in a
source program.  <args> specifies a debug information type.  One #
pragma__debug_cpp directive can take several <arg>.  One or more <arg>
must be specified for each directive.  MCPP begins to output debug
information when it finds this directive, and stops it when it
encounters #pragma __end_debug_cpp <args>.  The <args> can be omitted,
in which case all types of debug information is reset.  If <args>
contains an argument that is not supported by MCPP, MCPP issues a
warning and outputs the entire #pragma line, but all the preceding
arguments are regarded as valid.

All the debug information is output to the same path with the
preprocessing output to synchronize with it. Therefore, this directive
prevents compilation.  This directive is normally used without
specifying the output destination, in which case the debug information
is output to screen, so that you can trace it with your eyes.

When you noticed something was wrong with the preprocessing result,
enclose the coding you want to debug with the following directives, for
example:

#pragma __debug_cpp __token __expand
/* Coding you want to debug  */
#pragma __end_debug_cpp

As this directive was originally used for debugging MCPP itself, it was
not developed with end users in mind.  So, you may not understand its
behavior unless you read its source code, and you may sometimes feel it
outputs too much information, but it is useful for tracing the
preprocessing process.  Be patient.

#pragma __debug_cpp and #debug are not implemented unless at least
either DEBUG or DEBUG_EVAL is set to TRUE in system.H.  In case of DEBUG
== TRUE, tokenization or macro expansion can be traced.  If DEBUG_EVAL =
= TRUE, evaluation of a #if line can be traced.

The following debug information types can be specified with <arg>.  For
#pragma, all these labels are prefixed with __.

    For DEBUG == TRUE
        path    Displays the include file search path.
        token   Parses tokens one by one and displays its type.
        expand  Traces a macro expansion process.
        if      Displays the result (true or false) of #if, #elif,
                        #ifdef and #ifndef.
        getc    Traces preprocess 1-byte by 1-byte.
        memory  Displays the status of heap memory used by MCPP.
    For DEBUG_EVAL == TRUE
        expression  Traces #if expression evaluation.

3.5.1   #pragma __debug_cpp __path, #debug path

With these directives, MCPP displays include directories in the search
path (excluding the current and source directories with which search
begins) in the order of priority, starting with the highest one first.

In addition, with a #include directive, MCPP displays all the
directories, including the current one, it actually searched for the #
include file.  When a header file with #pragma __once specified is #
included again, the message to that effect is displayed.

3.5.2   #pragma __debug_cpp __token, #debug token

With these directives, MCPP displays a source line it has read, and then
displays a token and its type on the source line each time it has read.
This token, more specifically, is a preprocessing-token (pp-token).  Not
only pp-tokens on a source line but also ones MCPP reads again
internally during macro expansion are displayed repeatedly.

However, the following 1-byte tokens are not displayed for MCPP program'
s convenience sake:

1. '#' at the beginning of a preprocessing directive line.
2. '(' at the beginning of a parameter list of a function-like macro
definition.
3. ',' delimiting between function-like macro definition parameters.
4. '(' at the beginning of an argument list used for a function-like
macro invocation.

A pp-token has the following types:

    NAM: Identifier                 STR: String literal
    NUM: Preprocessing-number       WSTR: Wide string literal
    OPE: Operator or punctuator     CHR: Character constant
    SPE: Special pp-tokens, such as $ and @
    SEP: Token separator white space    WCHR: wide character constant

Of SEP, other than <newline> are not normally displayed.  Control codes
such as <newline> are displayed as <^J> or <^M>.

3.5.3   #pragma __debug_cpp __expand, #debug expand

With these directives, MCPP traces the expansion process of a macro
invocation.  When MCPP compiled with MODE >= STANDARD encounters a #
pragma __debug_cpp, it behaves as follows:

If there is a macro invocation, MCPP displays the macro definition.
Each argument is read, the argument is substituted for the corresponding
parameter in the replacement list and the replacement list is rescanned.
MCPP displays this whole process.  In case of nested macro definitions,
they are rescanned and expanded one by one.  If an argument has a macro,
MCPP traces the above process recursively before parameter substitution.

Each time control is passed to and returned from a certain set of MCPP
internal functions, MCPP displays the trace information along with the
function name. The following table shows the role of these functions.
Reading MCPP source code will gives you a concrete idea on what each
function is doing.

    expand          Main routine for macro expansion
    replace         Expands a macro one level down.
    collect_args    Collects arguments.
    prescan         Scans a replacement list and processes # and ##
                            operator.
    substitute      Substitutes a parameter with an argument.
    rescan          Rescans a replacement list.

Except for expand, above functions are indirectly recursive with each
other.

For replace and collect_args, MCPP displays data it internally stacks
during macro expansion.  This data is displayed using the following
internal codes:

    <n>         Nth parameter
    <TSEP>      Token delimiter inserted by MCPP
    <MAGIC>     Code that inhibits re-replacement of the macro with the
                        same name
    <RT_END>    Code that indicates the end of a replacement list

It is recommended that #pragma __debug_cpp __token should be also used.

For #debug expand, MCPP compiled with MODE < STANDARD uses internal
routines considerably different from those used for #pragma __debug_cpp.
The explanations are omitted.

3.5.4   #pragma __debug_cpp __if, #debug if

With these directives, MCPP displays #if, #elif, #ifdef and #ifndef
lines and reports their evaluation result (true or false).  As a #if
section is skipped, no report is made.

3.5.5   #pragma __debug_cpp __expression, #debug expression

With these directives, MCPP traces evaluation of a #if or #elif
expression.  DECUS cpp, based on which MCPP has been developed, provides
these directives for the purpose of debugging cpp itself.  I scarcely
modified them.  This directive outputs a very long list of internal
functions, as well as variable names and their values.  Unless you read
the cpp source code, you may not understand these variables.  However,
without the source code, you can manage to understand how the cpp pushes
onto and takes out of a evaluation stack a complex expression value.

3.5.6   #pragma __debug_cpp __getc, #debug getc

With these directives, MCPP outputs detailed data each time it calls get
(), a function to read one byte.  When MCPP compiled with MODE >=
STANDARD scans a pp-token, MCPP calls get() to read only the first byte
of the pp-token.

With a #debug getc, MCPP calls get() during token scan, resulting in a
tremendous amount of data output.

In any way, using these directives outputs a huge amount of data, so you
scarcely need to use them.

3.5.7   #pragma __debug_cpp __memory, #debug memory

With these directives, MCPP reports the status of the heap memory it has
internally allocated or released using malloc(), realloc() or free()
only once.  Only the malloc () I developed and some other types of
malloc () provide this functionality.  Refer to "4.extra" of porting.txt.
In case of other malloc (), MCPP will neither cause an error nor report
a status.

MCPP reports the heap memory status again when it terminates with these
directives on.  The same thing happens when MCPP terminates due to out
of memory.


3.6     #assert, #asm, #endasm

#assert is implemented in MCPP compiled with MODE == PRE_STANDARD.  #
assert provides the functionality equivalent to the #error directive in
the Standard C.  The following code in the Standard C:

#if ULONG_MAX/2 < LONG_MAX
#error Bad unsigned long handling.
#endif

can be expressed as:

#assert LONG_MAX <= ULONG_MAX/2

The argument of #assert is evaluated as a #if expression.  If it
evaluates to true (non-zero), MCPP does nothing and if false (0), it
displays the following message and then the argument line (after
processing line splicing and comments):

    Preprocessing assertion failed

MCPP counts this as error but continues processing.

This #assert is quite different from that of System V or GNU C/cpp.

MCPP regards a block enclosed with the #asm and #endasm directives as
assembler coding.  MCPP compiled with MODE == PRE_STANDARD implements
this functionality for Microware C/09 only.  To implement this
functionality in other compiler systems, do_old () and put_asm () in
system.c must be modified.

For a #asm block, MCPP performs trigraphs conversion and deletes
<backslash> and <newline>, but it neither performs comment processing,
checks tokens or characters, nor deletes white-space characters at the
beginning of a line.  Also, it does not expand a token that happens to
have the same name with a macro and outputs it as it is.  Other
directive lines have no meaning within the #asm block.

These #asm and #endasm directives do not conform to Standard C.  In the
first place, extensive directives in the form other than
"#pragma sub-directive" are not Standard C compliant.  Changing their
directive names to #pragma asm and #pragma endasm does not solve this
problem.  In Standard C, the source code must consist of a C token
sequence (more precisely, a preprocessing token sequence), however, an
assembler program is not a C token sequence.  To use assembly code in
the Standard C, there is no other way but to embed it in a string
literal token.  Then, you have to implement a built-in function that
processes that string literal in the compiler-proper and call it as
follows:

    asm (
        " leax _iob+13,y\n"
        " pshs x\n"
    );

However, this is not suitable for a longer assembly code, in which case,
you had better write the assembly code as a separate file like a library
function, and assemble and link the program.  This seems to be
inconvenient, but it is necessary to separate the assembler portion
completely to write a portable C program.  It is recommended that you
should write assembly code in a separate file rather than using #asm.


3.7    New C99 Features (_Pragma () Operators, Variable Argument
                Macros and others)

These features are implemented in MCPP compiled with MODE >= STANDARD.
The -S1 -V199901L option with __STDC_VERSION__ set to 199901L enables
the following C99's features.  The same thing can be said with C++ for
the -V199901L option with __cplusplus set to 199901L or more.  Although
C++ Standard does not provides for the features other than 1 or 7, MCPP
provides them for better compatibility with C99.  MCPP allows variable
argument macros in the C90 and C++ modes. [1]

1. Treats the text from // to the end of a line as a comment.
2. Enables variable argument macros.
3. Allows the sequence of p+, P+, p-, and P-, as well as e+, E+, e-, and
E-, in the preprocessing-number.  This is to represent a bit pattern of
a floating-point number in Hex, like 0x1.FFFFFEp+128.
4. Enables the _Pragma () operator.
5. MCPP compiled with the EXPAND_PRAGMA macro set to TRUE macro-expands
the arguments on a #pragma line that do not begin with STDC.  (By
default, MCPP is compiled with EXPAND_PRAGMA == FALSE, so they are not
macro expanded.) [2]
6. For compiler-systems with  long long, a #if expression is evaluated
in long long or unsigned long long.
7. Allows an escape sequence for Unicode in the forms of \unnnn and \
Unnnnnnnn in identifiers, character constants, string literals and pp-
numbers.  The value of a #if expression is evaluated as a hex
representation.

A variable argument macro takes a form of:

    #define debug(...)  fprintf(stderr, __VA_ARGS__)

Here is a macro invocation:

    debug( "X = %d\n", x);

This macro is expanded as follows:

    fprintf(stderr, "X = %d\n", x);

"..." in the parameter list corresponds to one or more parameters.  In
the above example, "..." corresponds to __VA_ARGS__ in the replacement
list.  During a macro invocation, several arguments that correspond to
the "...", including ",", are concatenated to be treated as one argument.

_Pragma ( "foo bar") has the same effect as specifying #pragma foo bar.
The argument of the _Pragma() opetator must be one string literal or
wide string literal.  For a wide string, the prefix (L) is deleted, and
for a string literal, " enclosing that string literal is deleted, and "
and \ in that literal is replaced with \"and \\, respectively, before it
is treated as a #pragma argument.

#pragma must be written somewhere in one logical line and its argument
is not macro-expanded at least for C90.  On the other hand, the _Pragma
() operator can be written anywhere in source code (even in a
replacement list), which gives the same effect with #pragma written in a
logical line.  The _Pragma () operator generated during macro expansion
is also valid.  This flexibility provides the pragma directive with a
wide range of portability and allows a header file to absorb the
difference in #pragma among compiler systems. (For this sample, see
pragmas.h and pragmas.t of "Validation Suite".) [3]

C99 stipulates a #if expression is of maximum integer type.  As
"long long" and "unsigned long long" are required types, the type of an
#if expression is "long long" or "unsigned long long" or larger.

Note:
[1] This is for compatibility with GNU C.  It is difficult also for
other compiler systems to implement C99 specifications all at once.
Probably, they will begin to implement them little by little with
__STDC_VERSION__ set to 199409L or so.

[2] All MCPP built-in #pragma directives begin with __, such as __once
or __debug_cpp.  This is to prevent conflict with user-defined macros.

[3] C99 stipulates that a #pragma argument that begins with STDC is not
macro-expanded.  For other #pragma arguments, whether macro is expanded
is implementation-defined.  MCPP compiled with the EXPAND_PRAGMA macro
in the system.H set to TRUE will macro-expand a #pragma argument.


3.8     Asm Statement in Borland C and Other Special Syntaxes

Borland C has the asm keyword.  This keyword is used to write assembly
code as follows:

    asm {
        mov x,4;
        ...;
    }

This is quite irregular and deviates from the C grammar more than #asm.
If there happen to be a token with the same name as a macro, it will be
macro-expanded.  The same can be said with Turbo C and MCPP itself.  It
is recommended that an assembler program should be written in a separate
.ASM file.

MCPP compiled with the TOP_SPACE macro set to FALSE deletes white-space
characters at the beginning of a line, resulting in column misalignment
in a line.

Visual C++ also has the __asm keyword, which provides the similar
functionality  to this.

GNU C and LSI C provide a Standard-compliant built-in function, asm
( " mov x,4\n").

The LSI C specific cpp provides a special built-in macro, __eval__().
The purpose of this macro is to convert an integer constant token
argument in hex or in octal to a decimal token.  I did not implement
this macro in MCPP because __eval__() is too special to call it a macro.
I was not willing to add it to the portable MCPP source.  Implementing
it in MCPP is easy.  If you want, refer to how to implement __LINE__.
Use evalnum() and sprintf() for conversion.


3.9     Compatibility with GNU C/Cpp

Although I tried to develop MCPP in such manner that the GNU C ported
MCPP compiled with MODE == STANDARD provides compatibility with GNU C/
cpp to the extent that it does not hinder practical use, it is still
incompatible in many aspects.

First of all, as shown in Chapter 2, there are many differences in
execution options.  MCPP implements neither -A option nor Standard
incompliant directives, including #assert and #ident. [1]

Fortunately, there seems to be quite few sources that cannot be compiled
due to a lack of this compatibility.

It is more problematic that there are some sources that assume special
specifications of old preprocessors.  Most of such source code receives
a warning when -pedantic is specified in GNU C.  MCPP compiled with MODE
== STANDARD, by default, provides almost the same functionality as GNU C
/cpp's -pedantic or -pedantic-errors since it implements Standard
compliant error checking.  However, since GNU C/cpp, by default, allows
such Standard violations without issuing a diagnostic, there are some
sources that take advantage of this.

It is very easy to rewrite such illegal code to Standard-compliant code,
so it is meaningless to take the trouble to write illegal code only to
impair portability and, what is worse, to provide a hotbed of bugs.
When you find such code, do not hesitate to correct it.

Note:
[1] The functionality of #assert and #ident should be implemented using
#pragma, if necessary.  The same can be said with #include_next and #
warning, but these directives seem to be sometimes used in GNU C, so I
grudgingly implemented them in MCPP, however, a warning is issued when
they are used.

3.9.1   Preprocessing FreeBSD 2/Kernel Source

Taking FreeBSD 2.2.2-R's kernel source code as an example, this section
explains some preprocessing problems.  All the directories that appear
in this section are installed in /sys (/usr/src/sys).   Of the items I
point out below, 3.9.1.7 and 3.9.1.8 are not necessarily Standard
violations and work properly in MCPP, but MCPP issues a warning because
their coding is confusing.  3.9.1.6 is an enhancement and C99 provides
the same functionality, but it differs from GNUC/cpp in notation.

3.9.1.1     Multi-Line String Literal

Assembly codes are embedded by the following manner in i386/apm/apm.c,
i386/isa/npx.c, i386/isa/seagate.c, i386/scsi/aic7xxx.h, dev/aic7xxx/
aic7xxx_asm.c,  dev/aic7xxx/symbol.c, gnu/ext2fs/i386- bitops.h, pc98/
pc98/npx.c:

asm("
    asm code0
#ifdef PC98
    asm code1
#else
    asm code2
#endif
    ...
");

When no " closing a string literal appears by the end of line, GNU C/cpp,
by default, interprets that the string literal ends at the end of line.
The above coding is based on this specification.  In addition, the
compiler-proper seems to interpret the whole content of asm () as a
string literal spreading across lines.

I think that assembler source code should be written in an separate file,
but if you want to embed it in ".c" file by all means, write it in the
following manner, instead of using the confusing coding shown above.

asm(
    "  asm code0\n"
#ifdef PC98
    "  asm code1\n"
#else
    "  asm code2\n"
#endif
    "  ...\n"
);

Standard C compliant preprocessors will accept it.

3.9.1.2     #else junk, #endif junk

The following line appears in ddb/db_run.c, netatalk/at.h, netatalk/aarp.
c, net/if-ethersubr.c, i386/isa/isa.h, i386/isa/wdreg.h, i386/isa/tw.c,
i386/isa/b004.c, i386/isa/matcd/matcd.c, i386/isa/sound/sound_calls.h,
i386/isa/pcvt/pcvt_drv.c, pci/meteor.c, and pc98/pc98/pc98.h:

#endif MACRO

This line should be changed to:

#endif /* MACRO */

3.9.1.3     #ifdef 0

To my surprise, i386/apm/apm.c contains the following strange line:

#ifdef 0

Of course, this should be written as:

#if 0

This code must have been neither debugged nor used.

3.9.1.4     Duplicate Definition of Macro

gnu/i386/isa/dgb.c has a duplicate definition of the following macro:

#define DEBUG

Some of header files have a  macro definition conflicting with this.

The Standard C regards duplicate definitions as "undefined", but how
they are treated depends on compiler systems; some make the first
definition valid after issuing an error message and others, like GNU C 2
/cpp, make the last definition valid without issuing any messages.  To
make the last definition valid, the following code should be added
immediately before the last definition.

#undef DEBUG

3.9.1.5     #warning

i386/isa/if_ze.c, and i386/isa/if_zp.c have the #warning directive.
This is the only Standard violation directive I found in the kernel
source.  To conform to the Standard C, there is no way but to comment
out this line.

MCPP V.2.3 or higher accepts #warning.

3.9.1.6     Variable Argument Macros

gnu/ext2fs/ext2_fs.h and i386/isa/mcd.c have the following macro that
takes variable number of arguments:

#define MCD_TRACE(fmt, a...)        \
{                                   \
    if (mcd_data[unit].debug) {     \
        printf("mcd%d: status=0x%02x: ",    \
            unit, mcd_data[unit].status);   \
        printf(fmt, ## a);          \
    }                               \
}

#   define ext2_debug(fmt, a...)  { \
        printf("EXT2-fs DEBUG (%s, %d): %s:",   \
            __FILE__, __LINE__, __FUNCTION__);  \
        printf(fmt, ## a);          \
        }

This is a GNU C/cpp-specific enhanced specification and cannot be
applied to other compiler systems.  The above "## a" can be simply
written as "a".  With ## and in the absence of an argument corresponding
to "a..." in a macro invocation, the preceding comma is deleted.  C99
also provides for variable argument macros, but their notation differs
from that of GNU C/cpp.  The above example is written as follows in C99:

#define MCD_TRACE( ...)             \
{                                   \
    if (mcd_data[unit].debug) {     \
        printf("mcd%d: status=0x%02x: ",    \
            unit, mcd_data[unit].status);   \
        printf( __VA_ARGS__);       \
    }                               \
}

#  define ext2_debug( ...)     {    \
            printf("EXT2-fs DEBUG (%s, %d): %s:",   \
                __FILE__, __LINE__, __FUNCTION__);  \
            printf( __VA_ARGS__);   \
            }

The most annoying difference is that in C99 requires one or more
arguments on a macro invocation corresponding to "..." while GNUC/cpp
requires 0 or more arguments corresponding to "a...".  To handle this,
when there is no argument corresponding to "...", MCPP V.2.3 or higher
issues a warning, instead of making it an error.  Therefore, you can
change the above code as follows:

#define MCD_TRACE(fmt, ...)         \
{                                   \
    if (mcd_data[unit].debug) {     \
        printf("mcd%d: status=0x%02x: ",    \
            unit, mcd_data[unit].status);   \
        printf(fmt, __VA_ARGS__);   \
    }                               \
}

#  define ext2_debug(fmt, ...)     {    \
            printf("EXT2-fs DEBUG (%s, %d): %s:",   \
                __FILE__, __LINE__, __FUNCTION__);  \
            printf(fmt, __VA_ARGS__);   \
            }

This is simpler with one-to-one correspondence.  However, this way of
writing has a disadvantage that a comma immediately before an empty
argument remains, resulting in, for example, printf( fmt, ).  In this
case, there is no other way but to write a macro definition in
accordance with C99 specifications, or avoid using an empty argument in
a macro invocation.  Harmless tokens, such as NULL or 0, are used to
write, for example, MCD_TRACE(fmt, NULL).[1]

Note:
[1] To use MCPP, source code must be rewritten in this way.  In addition,
with the -Q option, a huge amount of warnings is output not to the
screen but to the cpp.err file.  With the -E option, cc is not
terminated even with a preprocessing error, allowing you to roughly
check the entire source code.  By doing this, #ident and #assert can be
ignored substantially.  GNU C 2.95.3 or later also implements variable
argument macros based on the C99 syntax.  It is recommended to use this
syntax in the future.  GNU specific one provides the flexibility of
allowing for zero number of variable argument macros, but its notation
is bad in that (1) for the "args..." parameter, a white space must not
be inserted between "args" and "...", but such a pp-token is not
permitted, and that (2) it is not desirable that the notation for a
token concatenation operator is used to indicate a variable argument in
a replacement list.  It is desirable to allow zero number of variable
arguments based on the C99 notation. GNU C 3 introduced a notation for
variable argument macros that is a mixture of GNU C 2's traditional
notation and C99 one.  For details, refer to 3.9.6.3.

3.9.1.7     Empty Arguments during Macro Calls

The following macro invocations appear in nfs/nfs.h, nfs/nfsmount.h, nfs
/nfsmode.h, netinet/if_ether.c, netinet/in.c, sys/proc.h, sys/socketvars.
h, i386/scsi/aic7xxx.h, i386/include/pmap.h, dev/aic7xxx/scan.l, dev/
aic7xxx/aic7xxx_asm.c, kern/vfs_cache.c, pci/wd82371.c, vm/vm_object.h,
and vm/device/pager.c.  So do in /usr/include/nfs/nfs.h.

    LIST_HEAD(, arg2)
    TAILQ_HEAD(, arg2)
    CIRCLEQ_HEAD(, arg2)
    SLIST_HEAD(, arg2)
    STAILQ_HAED(, arg2)

The first argument is empty.  C99 approved empty arguments but C90
regarded them as undefined.  Taking it consideration that an argument
may happen to be empty during a nested macro invocation, empty arguments
should be approved, however, it is neither necessary nor desirable to
write an empty argument in source code.  Note that for a one-argument
macro, there is syntax ambiguity between an empty argument and a lack of
argument.

Taking everything into consideration, the following notation is
recommended:

#define EMPTY

    LIST_HEAD(EMPTY, arg2)
    TAILQ_HEAD(EMPTY, arg2)
    CIRCLEQ_HEAD(EMPTY, arg2)
    SLIST_HEAD(EMPTY, arg2)
    STAILQ_HAED(EMPTY, arg2)

Any Standard C compliant preprocessor will accept this notation.

By the way, some of the header files (in the nfs directory) shown in the
previous page neither have the macro definitions shown above nor #
include any other header files.  This is because such header files
assume that these macro definitions exist in sys/queue.h and that *.c
programs will #include sys/queue.h first.  These files arise ambiguity.

kern/kern_mib.c has the following macro definitions:

    SYSCTL_NODE(, arg2, arg3, arg4, arg5, arg6, arg7, arg8, arg9)

In this case, the first argument cannot be changed to EMPTY.  Because
the corresponding macro definition in the sys/sysctl.h is as follows:

#define SYSCTL_NODE(parent, nbr, name, access, handler, descr)
    extern struct linker_set sysctl_##parent##_##name;
    SYSCTL_OID(parent, nbr, name, CTLTYPE_NODE|access,
        (void*)&sysctl_##parent##_##name, 0, handler, "N", descr);
    TEXT_SET(sysctl_##parent##_##name, sysctl__##parent##_##name);

In other words, these arguments are not macro-expanded.  The arguments
of the SYSCTL_OLD macro shown above, including the first one, are not
macro expanded.  In this case, there is no way but to leave the empty
argument as it is.  [1]

Note:
[1] C99 approves empty arguments as legitimate.  Taking macros, such as
SYSCTL_NODE () and SYSCTL_OID (), into consideration, the EMPTY macro is
not almighty and using empty arguments has some reason.  In addition,
even if EMPTY is used, a nested macro invocation may cause empty
arguments.  However, for source readability, using EMPTY is recommended
whenever possible.

3.9.1.8     Object-Like Macros Replaced with Function-like Macro Name

i386/include/endian.h, as well as /usr/include/machine/endian.h, has the
following macro definitions. (There are four same kinds of definitions.)

#define __byte_swap_long(x) (replacement text)
#define NTOHL(x)            (x) = ntohl ((u_long)x)
#define ntohl               __byte_swap_long

The problem is the ntohl definition.  Although ntohl is an object-like
macro, it is expanded to a function-like macro name, then rescanned with
subsequent text, and is expanded as if it were a function-like macro.
This way of macro-expansion has been regarded as an implicit
specification since K&R 1st, and the Standard C tentatively approved it
as legitimate.  However, as I discuss in other documents, it is this
specification that makes macro-expansion unnecessarily complicated and
brings confusion to Standard documents.  This is a bug specification.

The following example shows a function-like that looks like an object-
like macro.  You had better define ntohl like a function-like macro that
it is:

#define ntohl(x)    __byte_swap_long(x)

This causes no problem.

i386/isa/sound/os.h has the same kind of macro definitions:

#define INB         inb
#define INW         inb

This should be written as follows:

#define INB(x)      inb(x)
#define INW(x)      inb(x)

Note:
[1] ISO 9899:1990 Corrigendum 1:1994 regarded the notation as undefined.
C99 replaced this article with other.  However, Standard documents are
still confusing about this.  For details, see 1.7.6 of cpp_test.txt.

3.9.1.9     Preprocessing .S File

Some kernel sources are contained in several ".S" files, that is, they
are written in assembler.  These sources contain #includes or #ifdefs,
which require preprocessing.  To preprocess them, in FreeBSD 2.2.2-R, cc
is called with the -x assembler-with-cpp option, and cc calls /usr/
libexec/cpp with the -lang-asm option and then calls as.

Of course, this ways of using .S files is illegal.  This assembler
source code must not contain a token that happens to have the same name
with a macro.  White spaces between tokens and at the beginning of a
line must be retained during preprocessing..  In addition, if the first
token at the beginning of a line is a # indicating an assembler comment,
special processing is required on the cpp side.  This not only
considerably limits available cpps but also increases the possibility of
unknowingly introducing bugs.  So, using .S files in this way is not
recommended.[1]

To preprocess source code for use with several types of machines, the
code should be written in the following manner and be saved in not ".S"
but ".c" file.  4.4BSD-Lite actually adopts this way of coding.

asm(
    "  asm code0\n"
#ifdef Machine_A
    "  asm code1\n"
#else
    "  asm code2\n"
#endif
    "  ...\n"
);

Note:
[1] In FreeBSD 2.0-R, these kernel sources are contained not in *.S but
in *.s file.  The Makefile is so defined as to call cpp, instead of cc,
to process them.  Then the cc calls as.  When the cpp is called, /usr/
bin/cpp is invoked.  /usr/bin/cpp is a shell-script that calls /usr/
libexec/cpp -traditional.  This method was more convenient in that it
provides a way to change preprocessors to be used by modifying the
script.

3.9.2   Preprocess of FreeBSD 2/libc Source

I have recompiled all the source files in /usr/src/lib/libc of FreeBSD 2.
2.2R.  There was no problem, probably because most of them comes from 4.
4BSD-Lite without modification.  It is quite rare and surprising that a
huge amount of source files in excellent quality is gathered together.

Only at one place, I found the following coding in gen/getgrent.c.  Of
course, ";" at the end of line is surplus.

#endif;

3.9.3   Problems Concerning GNU C 2/cpp

As seen so far, writing a Standard-compliant source code with better
portability in a more secure manner neither requires much effort nor
provides any demerits.  In spite of it, why does source code less
compliant to Standards still exist at all?

When comparing the FreeBSD 2.0-R kernel sources with those of 2.2.2-R,
Standard-incompliant ones do not decrease in number.  The problem is
that newer sources are not necessarily more compliant to the Standards.
There are few Standard-incompliant sources in 4.4BSD-Lite.  This is
probably because the 4.4BSD sources were rewritten to become compliant
to the Standard C and POSIX.  However, during the process of
implementing these sources to FreeBSD, the old writing style revived in
some sources.  For example, although the ntohl shown above is written as
ntohl (x) in 4.4BSD-Lite, it is written as ntohl in FreeBSD.  Why did
the notation once put away revive?

I blame GNU C/cpp for this revival, which passes these Standard-
incompliant sources without issuing a diagnostic.  If -pedantic had been
a default behavior, the old style source would have never revived.  If -
pedantic-errors had been a default behavior, although, GNU C/cpp would
not have been put into practical use because too many sources failed to
be compiled.  The gcc's man page describes the -pedantic option as:
"There is no reason to use this option except for satisfying pedants."[1]
Now that eight years have already passed since Standard C was
established, it is a high time that  GNU C/cpp should set -pedantic as
default, not go so far as to set -pedantic-errors.

In FreeBSD 2.0-R, nested comments were sometimes found, but in 2.2.2-R,
they disappeared. This is because GNU C/cpp no longer allowed them.
This has nothing to do with -pedantic, but I want to say how influential
cpp's source checking is.

Without #pragma __once in a header file, MCPP is slow in processing
multiple includes.

Execute the following command in each directory under /sys:

    ins_once *.h */*.h */*/*.h

Note:
[1] I wrote 3.9.3 in 1998.  After that, gcc's man page or info deleted
this expression, however, the specification remains the same.

3.9.4   Preprocessing Linux/glibc 2.1 Source

I recompiled glibc 2.1.3 sources on Vine Linux 2.1 (i386).  Different
from those of FreeBSD libc, I found many problems.  Some sources are
written based on GNU C/cpp's undocumented specifications, in which case
it took me a lot of time to identify them.

3.9.4.1     Multi-Line String Literal

sysdeps/i386/dl-machine.h and stdlib/longlong.h have many multi-line
string literals as shown below:

#define MACRO asm("
    instr 0
    instr 1
    instr 2
")

Some string literals are very long.  compile/csu/version-info.h created
by make also has a multi-line string literal.  Of course, it is Standard
non-conforming, but GNU C treats it as a string literal with embedded
<newline>.

The -lang-asm (-x assembler-with-cpp, -a) option allows MCPP V.2.3 to
convert a multi-line string literal into the following code:

#define MACRO asm("\n  instr 0\n  instr 1\n  instr 2\n")

However, this option cannot work properly for a string literal with a
directive inserted in the middle as shown in 3.9.1.1, in which case
there is no way but to rewrite the source.

These GNU C's too illegal specifications seem to have been made obsolete
by as late as GNU C V.3.

3.9.4.2     #include_next, #warning

#include_next appears in the following files:

catgets/config.h, db2/config.h, include/fpu_control.h, include/limits.h,
include/bits/ipc.h, include/sys/sysinfo.h, locale/programs/config.h, and
sysdeps/unix/sysv/linux/a.out.h

sysvipc/sys/ipc.h has #warning.

Although these directives are not approved by the Standard C, #
include_next, in particular, becomes indispensable for glibc 2.  So,
MCPP V.2.3 or higher for GNU C implements #include_next and #warning.

The problems concerning #include_next is that it is not only a standard
violation but also that what headers are actually included depends on
the setting of include directories and a search order, which are
controlled by users via environment variables.

When glibc is installed, some files in glibc's include directory are
copied to the /usr/include directory.  These files are used as system
header files.  That these header files contain #include_next means
system headers become patchy.  It seems to be time to reorganize them.

3.9.4.3     Variable Argument Macros

The following files contain definitions of macros with variable number
of arguments based on the GNU C/cpp specification, as well as macro
invocations:

elf/dl-lookup.c, elf/dl-version.c, elf/ldsodefs.h, glibc-compat/nss_db/
db-XXX.c, glibc-compat/nss_files/files-XXX.c, linuxthreads/internals.h,
locale/loadlocale.c,  locale/programs/linereader.h,  locale/programs/
locale.c, nss/nss_db/db-XXX.c, nss/nss_files/files-XXX.c, sysdeps/unix/
sysdep.h, sysdeps/unix/sysv/linux/i386/sysdep.h, and sysdeps/i386/fpu/
bits/mathinline.h

This is a deviation from the C99 Standard.  You must rewrite the source
code  before you can use MCPP.

3.9.4.4     Empty Argument During Macro Calls

The following files have macro invocations with empty arguments:

catgets/catgetsinfo.h, elf/dl-open.c, grp/fgetgrent_r.c, libio/
clearerr_u.c, libio/rewind.c, libio/clearerr.c, libio/iosetbuffer.c,
locale/programs/ld-ctype.c, locale/setlocale.c, login/getutent_r.c,
malloc/thread-m.h, math/bits/mathcalls.h, misc/efgcvt_r.c, nss/nss_files
/files-rpc.c, nss/nss_files/files-network.c, nss/nss_files/files-hosts.c,
nss/nss_files/files-proto.c, pwd/fgetpwent_r.c, shadow/sgetspent_r.c,
sysdeps/unix/sysv/linux/bits/sigset.h, sysdeps/unix/dirstream.h

math/bits/mathcalls.h, in particular, contains as much as 79 empty
arguments.  This header file is installed in /usr/include/bits/mathcalls.
h and is #included by /usr/include/math.h.  Even with an EMPTY macro,
nested macro invocations generate a lot of empty arguments.  Are there
any other ways to write macros more clearly?

3.9.4.5     Object-Like Macros Replaced with Function-like Macro Name

The following files contain object-like macro definitions replaced with
function-like macro names:

argp/argp-fmtstream.h, ctype/ctype.h, elf/sprof.c, elf/dl-runtime.c, elf
/do-rel.h, elf/do-lookup.h, elf/dl-addr.c, io/ftw.c, io/ftw64.c, io/sys/
stat.h, locale/programs/ld-ctype.c, malloc/mcheck.c, math/test-*.c, nss/
nss_files/files-*.c, posix/regex.c, posix/getopt.c, stdlib/gmp-impl.h,
string/bits/string2.h, string/strcoll.c, sysdeps/i386/i486/bits/string.h,
sysdeps/generic/_G_config.h, sysdeps/unix/sysv/linux/_G_config.h

Of these, some function-like macros, like math/test-*.c , are first
replaced with an object-like macro name and then further replaced with a
function-like macro name.  Why did these macros have to be written in
this way?

3.9.4.6     Macros Expanded to 'defined'

sysdeps/generic/_G_config.h, sysdeps/unix/sysv/linux/_G_config.h, and
malloc/malloc.c contain the following macro definition expanded to the
"defined" pp-token.

#define HAVE_MREMAP defined(__linux__) && !defined(__arm__)

The intention of this macro definition is that with the following
directive,

#if HAVE_MREMAP

, the above line is expected to be expanded as follows:

#if defined(__linux__) && !defined(__arm__)

However, the behavior is undefined in Standard C when a #if line has a
"defined" pp-token in a macro expansion result.  Apart from it, this
macro definition is strange in the first place.

The HAVE_MREMAP macro is first replaced with the following,

defined(__linux__) && !defined(__arm__)             (1)

, and then the identifiers "defined", "__linux__" and "__arm__" are
rescanned for more macro replacement.  If any of them is a macro, it is
expanded.  In this case, "defined" cannot be defined as a macro
(Otherwise, it causes another undefined result), and if __linux__ is
defined as 1 and __arm__ is not defined, this macro is finally expanded
as follows:

defined(1) && !defined(__arm__)

defined(1), of course, is a syntax error of a #if expression.

However, GNU C/cpp stops macro expansion at (1) and regards it as the
final macro expansion result of the #if line.  Since this is "undefined"
anyhow, this GNU specification cannot be described as wrong, but it
lacks of consistency in that how to expand a macro differs between
macros in a #if line and in other lines.  At least, it lacks of
portability. [1]

The above code should be written as follows:

#if defined(__linux__) && !defined(__arm__)
#define HAVE_MREMAP 1
#endif

I hope this kind of confusing code be eliminated as early as possible.

Note:
[1] GNU C 2/cpp internally treats "defined" in a #if line as a special
macro.  For this reason, when GNU C/cpp rescans the following sequence
of tokens for macro expansion, it evaluates it as a #if expression,
instead of macro expanding it.  In other word, distinction between macro
expansion and #if expression evaluation is ambiguous.

    defined(__linux__) && !defined(__arm__)

This problem relates to GNU C/cpp' own program structure.  GNU C 2/cpp
has a de facto main routine rescan(), which is a macro rescanning
routine.  This routine reads and processes source code from the
beginning to the end, during the course of which, it calls a
preprocessing directive processing routine.  Although implementing
everything using macros is a traditional program structure of a macro
processor, this structure can be thought to cause mixture of macro
expansion and other processing.

3.9.4.7     Preprocessing .S File

The files named *.S contain assembler source code requiring
preprocessing.  Some of these files have preprocessing directives, such
as #include, #define, and #if.  In addition, the file named compile/csu/
crti.S generated by Make contains the following lines:

#APP

or

#NO_APP

From a syntax point of view, preprocessors cannot tell whether these
lines are invalid preprocessing directives or valid assembler comments.
GNU C seems to leave these lines as they are during preprocessing and
treat it as assembler comments.

Concatenation of pp-tokens using the ## operator sometimes generates an
invalid pp-token.  GNU C/cpp outputs these pp-tokens without issuing a
diagnostic.

For compatibility with GNU C/cpp, I reluctantly decided that, with the -
lang-asm (-x assembler-with-cpp, -a) option, MCPP V.2.3 does not treat
these illegal directives and invalid pp-tokens generated by ## as error,
and outputs them as they are and issues a warning.

Essentially, these sources should be processed with an assembler macro
processor.  GNU seems to provide a macro processor called gasp, but it
seems to be scarcely used for some reason.

3.9.4.8     Problems of rpcgen and -dM Option

When invoked with the -dM option, GNU C/cpp outputs only macro
definitions, which is used by stdlib/isomac.c in make check routine.

The problem of the isomac.c is that it accepts only GNU C/cpp's macro
definition file format and regards a comment or a blank line as an error.
In order for isomac.c to use this macro definition file output by MCPP
with the -dM option, you must modify the stdlib/isomac.c using the
appended isomac.dif.

Glibc make sometimes uses a program called rpcgen.  The problem of
rpcgen is that it accepts only GNU C/cpp's output format of preprocessor
line number information as follows:

#123 "filename"

Rpcgen does accept neither:

#line 123

nor

#line 123 "filename"

Rpcgen regards them as error.

I reluctantly decided that MCPP V.2.3 and later for GNU C uses the GNU C
/cpp format by default.  Rpcgen's specification is poor in that it is
based on a particular compiler system's format and cannot accept the
standard one.

3.9.4.9     -include, -isystem and -I- Options

Glibc 2 make often uses the -include option and sometimes uses -isystem
and -I- options.  The former can be substituted with #include at the
beginning of source code.  The latter is less necessary; it is only used
to update system headers.

Only MCPP V.2.3 for GNU C implements these two options, but I would like
these less necessary options to be made obsolete. [1]

Note:
[1] GNU C/cpp provides several more options that specify include
directories and their search orders, such as -iprefix, -iwithprefix, and
-idirafter.  It also provides the -remap option that specifies mapping
between long-file-names and MS-DOS 8+3 format filenames.  On CygWIN
systems, specs files contain these options, but it is not necessary to
use these options because include directories can be specified with
environment variables and because such mapping is no longer necessary on
CygWIN.

3.9.4.10    Undocumented Predefined Macros

The following macros are GNU C/cpp predefined macros although their
names do not appear in documentation.

__VERSION__,  __SIZE_TYPE__,  __PTRDIFF_TYPE__, and __WCHAR_TYPE__

On Vine Linux 2.1 (egcs-1.1.2) systems, __VERSION__ is set to
"egcs-2.91.66 19990314/Linux (egcs-1.1.2 release)".  On many systems,
including Linux/i386, the values of other three macros have types
unsigned int, int, and long int, respectively.  However, on FreeBSD,
CygWIN, and DJGPP systems, their types are slightly different from them
(I do not know why).  Why does those predefines macros remain
undocumented?

Although GNU C/cpp has the __VERSION__ value in its own source code,
MCPP uses the environment variable GCC_VERSION to set the value.

3.9.4.11    Undocumented Environment Variables

The most strange thing is the undocumented environment variable named
SUNPRO_DEPENDENCIES. sysdeps/unix/sysv/linux/Makefile contains the
following script:

    SUNPRO_DEPENDENCIES='$ (@:.h=.d)-t $@'
    $ (CC) -E -x c $ (sysinclude) $< -D_LIBC -dM |
    ...
    etc.

The intent of this script is to specify a file name with the environment
variable SUNPRO_DEPENDENCIES, and to have cpp output macro definitions
in source code and dependency description lines between source files to
that file.

I had no other way but to read the GNU C/cpp source code (egcs-1.1.2/gcc
/cccp.c) to know how this environment variable works.

In addition, there is another environment variable, DEPENDENCIES_OUTPUT,
which has a similar function.  The difference between the two is that
SUNPRO_DEPENDENCIES also outputs dependency description lines among
system headers, but DEPENDENCIES_OUTPUT does not.

Only MCPP V.2.3 for the GNU C enables these two environment variables,
but I would like these undocumented specifications to be made obsolete
as early as possible.

3.9.4.12    Other Problems

Linux (i386)/GNU C appends the -Asystem(unix), -Acpu(i386) or -Amachine
(i386) to cpp invocation options by using specs file.  As long as the
glibc 2.1.3 for Linux/x86 is concerned, there seems to be no source code
that utilizes this functionality.

It is a big problem that glibc's system headers have become patchy and
very complicated.  A small difference in settings may result in a big
difference in preprocessing results.

On the other hand, Glibc 2.1.3 did not contain #else junk, #endif junk,
or duplicate macro definitions that were found in FreeBSD 2.2.2/kernel
sources.  In some aspects, Glibc 2 source is better organized than
FreeBSD 2/kernel source.

However, as a whole, there were not a few sources that are based on GNU
C-specific specifications in glibc 2, which impairs portability to other
compiler systems although such sources form only a small portion of
several thousand source files.  Dependence on GNU C local specifications
is not desirable for program readability and maintainability.  I hope
that GNU C V.3 will make obsolete these local specifications and that
all the source code based on them will be completely rewritten.

3.9.5   To Use MCPP with GNU C 2

You must modify some source code as follows before you can use MCPP V.2.
3 or higher to compile glibc 2 programs:

1. Macro definitions with variable number of arguments: Modify the 14
files in 3.9.4.3 as shown in 3.9.1.6.  Of course, you had better save
the original files under other name such as *.orig.

2. Macros contained in the three files shown in 3.9.4.6 that has
"defined" in its replacement list:  /usr/include/_G_config.h is a file
generated when sysdeps/unix/sysv/linux/_G_config.h is installed and has
the same contents with this.  You had better modify /usr/include/
_G_config.h.

3. stdlib/isomac.c in 3.9.4.8

As environment variables, such as C_INCLUDE_PATH and GCC_VERSION, are
used by other programs, write these variables, say, in ~/.bash_profile
for bash and, say, in ~/.login for tcsh.

In addition to the options specified in Makefile or specs file, you must
specify the -lang-asm (-xassembler-with-cpp) option to process *.S files
containing multi-line string literals or assembler comments before you
can invoke MCPP.  Normally, you can leave this option specified when
preprocessing other files.

When you want to use GNU C/cpp or MCPP, or change the default options,
you had better perform the following steps:

1. Login as a super-user to move to the directory where MCPP resides.
(On Vine Linux 2.1, go to /usr/lib/gcc-lib/i386-redhat-linux/egcs-2.91.
66).  Let me assume that this directory has GNU C/cpp installed under
the name of cpp and MCPP cpp_std.

2. Create a file called cpp.sh with the following contents. [1], [2], [3]

#!/bin/sh
/usr/lib/gcc-lib/i386-redhat-linux/egcs-2.91.66/cpp_std -v -Q \
    -lang-asm $@

The -v and -Q options are optional, however, I recommend that you should
use -Q to record a large amount of diagnostic messages and -v to confirm
successful invocation of MCPP.  (Similarly, I recommend you to create a
shell-script that invokes gcc with -v.

3.  Enter the following commands:

    chmod a+x cpp.sh
    mv cpp cpp_gnuc
    ln -sf cpp.sh cpp

These commands execute cpp.sh linked to cpp when gcc calls cpp, and cpp.
sh calls cpp_std using the above options before the ones specified by
gcc.

4.  To change default options, modify cpp.sh or call cpp_std directly.
Use GNU C/cpp in the following manner:

    ln -sf cpp_gnuc cpp

With the above settings, staying linked to MCPP may cause a problem;
Some programs call cpp with the -traditional option.  However, MCPP in
STANDARD and POST_STANDARD modes does not implement the -traditional
option.  So, you had better modify cpp.sh to  automatically select GNU C
/cpp when the -traditional option is specified, and MCPP when not: [4]

#!/bin/sh
for i in $@
do
  case $i in
    # Note "$@" instead of $@, some tools needs this fashion.
    "-traditional*")
      /usr/lib/gcc-lib/i386-redhat-linux/2.91.66/cpp_gnuc "$@"
      exit
      ;;
  esac
done
/usr/lib/gcc-lib/i386-redhat-linux/2.91.66/cpp_std -Q -lang-asm $@

Another problem of using MCPP is that it issues a huge amount of warning
messages.  You can redirect them to a file using the -Q option, but when
you uses with the -W3 option to preprocess a large amount of source code,
such as glibc, several hundred MB of cpp.err is created, so it is
impossible for you to look through the whole file.

Taking a close look at cpp.err, you will find same warnings being issued
repeatedly.  This is because the same *.h files are #included by many
source programs.  To make the file more readable, perform the following
procedure:

1.  To find error messages, enter the following command:

    grep fatal `find . -name cpp.err`
    grep error `find . -name cpp.err`

2:  To sort warning messages, enter the following command:

    grep warning `find . -name cpp.err` | sort -k4 -u | less

3.  To find all the source lines causing a warning, enter the following
command:

    grep warning `find . -name cpp.err` | sort -k4 | uniq | less

4.  To find a particular type of warnings, enter the following command,
for example:

    grep 'warning: Replacement' `find . -name cpp.err` | sort -k4
        | uniq | less

After you get an overall idea of what source lines are causing what
kinds of errors or warnings, you can see a particular cpp.err by less
and then, if necessary, see the source file in question.

In addition, you can sandwich the source code in question with #pragma
__debug_cpp __expand and #pragma __end_debug_cpp and preprocess it again
to see the output, in which case I recommend you to invoke MCPP in the
following manner so that preprocessing results and diagnostic messages
are output to the same file:

    cpp_std <-opts> in-file.c > in-file.i 2>&1

When you use "make", you must temporarily change the above shell-script.

Note:
[1] VineLinux 2.5/GNU C 2.95.3 has cpp0 and cpp files in the /usr/lib/
gcc-lib/i386-redhat-linux/2.95.3 directory.  cpp is linked to cpp0.  In
other word, cpp0 is a real preprocessor.  The gcc calls cpp0 directly.
Does this mean other preprocessors like cpp1 and cpp2 are created in the
future?  Anyway, in this case, it is necessary to change cpp to cpp0 in
the above mv or ln command.

FreeBSD 4.4/GNU C 2.95.3 also has cpp and cpp0 in /usr/libexec.

[2] VineLinux 2.5/ GNU C 2.95.3 has the following cpps, each of which is
a link to other cpps, except the last one (/usr/bin/cpp-2.95.3), which
is an executable.  To make the matter more complicated, the last one
calls /usr/lib/gcc-lib/i386-redhat-linux/2.95.3/cpp0.   What brought
about this confusing situation?
    /lib/cpp ->/etc/alternatives/libcpp
    /etc/alternatives/libcpp ->/usr/bin/cpp-2.95.3
    /etc/alternatives/cpp ->/usr/bin/cpp-2.95.3
    /usr/bin/cpp ->/etc/alternatives/cpp
    /usr/bin/cpp-2.95.3

[3] When the Validation Suite is applied to MCPP in GNU C/testsuite, use
the -23j option to enable digraph and trigraph, and to output a message
in the similar format to GNU, which does not contain any additional
information, such as source lines.  Other options must not be specified
in this shell-script.  For more information on using Validation Suite in
GNU C/testsuite, refer to 2.2.3 of cpp-test.txt.

[4] This shell script is for use with VineLinux 2.1/GNU C 2.91.66.  For
use with other FreeBSD, Linux and CygWIN versions, change the name of
the directory where cpp resides and cpp names (cpp or cpp0).  GNU C 3
has tradcpp0, a separete preprocessor in the traditional mode, so when
the -traditional option is used to invoke gcc, tradcpp0 is invoked.
Therefore, this shell script is not used for GNU C 3.

3.9.6     Preprocessing GNU C 3.2 Source

I first compiled GNU C 3.2R sources on Linux and FreeBSD, then I used
the generated gcc to compile MCPP and then I recompiled GNU C 3.2
sources using MCPP for preprocessing.

New GNU C compilers are bootstrapped during various phases of make; gcc
and cc1, etc generated in an earlier phase are used to recompile
themselves, and those generated compiler drivers and compiler-propers
are used again to recompile themselves, and so on.  During the bootstrap,
gcc exists under the name of xgcc.

Other than cc1 and cc1plus, GNU C 2 has a separate preprocessor called
cpp.  In GNU C 3, cpp was absorbed into cc1 and cc1plus.  However, there
still exists a sepatate cpp (or cpp0).  To have cpp0 preprocess, the -no-
integrated-cpp option must be specified when you invoke gcc or g++.
Therefore, to have MCPP preprocess, you must use a shell-script that
have gcc (xgcc) or g++ invoke MCPP first then invoke cc1 or cc1plus. [1]

In the GNU C compiler system, the settings of system headers and their
search order are becoming very complex.  So, a small difference in
settings may result in a difference in preprocessing results.  Even
successful compilation was often difficult to attain.  In addition,
compilation and tests require a lot of other software.  Older versions
of such software may cause failure in compilation or tests.  Actually,
compilation sometimes failed due to some hardware problems.

Actually, I failed to compile GNU C 3.2 source under FreeBSD 4.4R.  I
had to upgrade FreeBSD to 4.7R and changed software packages to those
for FreeBSD 4.7R before I was able to succeed in compilation. [2]

I use VineLinux 2.5 on two PCs.  Although compilation of GNU C 3.2
sources using GNU C 2.95.3 was successful on one PC (K6/200MHz),
recompilation of GNU C 3.2 sources using the generated GNU C 3.2/cc1
failed, and caused many segmentation faults.  Then I changed CPU from K6
to AthlonXP.  This time, recompilation was successful; no segmentation
faults occurred.  Hardware may have caused the problem.

When I compiled GNU C 3.2 sources using GNU C 2.95.4 under FreeBSD on K6,
"make -k check" of the generated gcc was almost successful.  When I
recompiled GNU C 3.2 itself using the generated GNU C 3.2, in
"make -k check" of g++ and libstdc++-v3, about 20 percent of testsuite
was unsuccessful.  However, when using AthlonXP, instead of K6,
everything went OK.  Hardware may have caused the problem.

On both VineLinux PCs, when I recompiled GNU C 3.2 sources using GNU C 3.
2 itself and MCPP, "make -k check" of the generated gcc was successful.
However, in "make -k check" of g++ and libstdc++-v3, 20 percent of
testsuite failed.  [3], [4], [5]

In anyway, the cause of this testsuite failure seems to lie not in the
generated compilers themselves, such as gcc, g++, cc1 and cc1plus, but
in the header file settings.

MCPP cannot be described as completely compatible with GNU C/cpp, but is
highly compatible.  So, MCPP and GNU C/cpp can be used interchangeably.

GNU C 3.2 sources were compiled in the following environment:

    OS              make        library     CPU
    VineLinux 2.5   GNU make    glibc 2.2.4 Celeron/1060MHz
    VineLinux 2.5   GNU make    glibc 2.2.4 K6/200MHz, AthlonXP/2.0GHz
    FreeBSD 4.7R    UCB make    libc.so.4   K6/200MHz

Only C and C++ were compiled.

My Validation Suite V 1.3 has the new edition to be used in GNU C
testsuite.  Validation Suite V 1.3 allows you to perform detailed and
systematic preprocessor tests using "make -k check" or "runtest".
Validation Suite V 1.3 checks not sources but preprocessor's behaviors.
For details, see of 2.2.3 of cpp-test.txt

Note:
[1] I had to do this for each bootstrap stage.  Since makefile is too
large and too complex to change, I employed an inelegant method; I kept
on sitting in front of PC screen during the entire process of bootstrap.
At each end of the stages, I entered ^C and replaced xgcc and others
with shell-scripts.

[2] Due to dependency between packages, the system falls into confusion
unless appropriate versions are installed.  Actually, for this reason,
my FreeBSD temporarily failed to invoke kterm.

[3] "make -k check" cannot be used with MCPP because diagnostics of MCPP
are different from those of GNU C/cpp.

[4] "make -k check" seems to require an English environment, so the LANG
environment variable must be set to C.

[5] All the testsuite failures were caused by inability of the pthread_*
functions, such as pthread_getspecific and pthread_setspecific, to be
linked in the library i686-pc-linux-gnu/libstdc++-v3/src/.libs/libstdc++.
so.5.0.0.  When a correctly generated library was installed,
"make -k check" was successful.  On FreeBSD, this problem never happened.
This is probably because of small differences in settings.

3.9.6.1     Multi-Line String Literal

This very old way of coding was no longer found in GNU C 3.2 sources.
Multi-line string literals were made obsolete as late as at GNU C 3.
GNU C 3 processes a source with a multi-line string literal as you
expect, but issues a warning.

3.9.6.2     #include_next and #warning

limits.h and syslimits.h in build/gcc/include generated during the
course of make have #include_next.  When GNU C 3.2 is installed, these
header files are copied to limits.h and syslimits.h in lib/gcc-lib/i686-
pc-linux-gnu/3.2/include.

GNU C 3.2 sources does not have #warnings.

3.9.6.3     Variable Argument Macros

GNU C 3.2 sources have some variable argument macros, but most of them
are found in testsuite and they are nothing but test samples.  Although
GNU C 3.2 still supports variable argument macros in GNU C 2 notation,
the ones using __VA_ARGS__(in C99 notation) are more frequently found in
GNU C 3.2 sources.

In GNU C 3, variable argument macros in a mixed notation of GNU C 2 and
C99 are found:

    #define eprintf( fmt, ...)   fprintf( stderr, fmt, ##__VA_ARGS__)

According to the GNU C 3 specification, in the absence of an argument
corresponding to "...", the comma immediately before "..." is deleted.
So, this is expanded as follows:

    eprintf( "success!\n")  ==>  fprintf( stderr, "success!\n")

As far as this example is concerned, this specification seems to be
convenient, but is not desirable in that (1) a comma in a replacement
list of a macro definition is not always used to delimit parameters, (2)
it allows a token concatenation operator (##) to have other
functionality, (3) it makes rules more complex by allowing exceptions.
MCPP does not implement this functionality.  MCPP does not regard this
macro definition as an error, but it does not delete the comma
immediately before the empty argument in a macro invocation.

3.9.6.4     Empty Arguments in Macro Invocation

Apart from #includeded system headers, such as /usr/include/bits/
mathcalls.h and /usr/include/bits/sigset.h, empty arguments in a macro
invocation are found only in gcc/libgcc2.h of GNU C 3.2 sources
themselves. [1]

Note:
[1] These two header files are copied into the system header directory
when glibc is installed.  They do not exist on FreeBSD because glibc is
not used.

3.9.6.5     Object-Like Macros Replaced with Function-Like Macros

gcc/fixinc/gnu-regex.c and libiberty/regex.c have object-like macros
that are replaced with function-like macro name.  /usr/lib/bison.simple,
a #included file, also has such macros.  These macros are all relevant
to alloca.  For example, libiberty/regex.c has the following macro
definitions.

    #define REGEX_ALLOCATE  alloca
    #define alloca( size)   __builtin_alloca( size)

This should be written as follows:

    #define REGEX_ALLOCATE( size)   alloca( size)

Why did they omit (size)?

In addition, regex.c also has another alloca, which is defined as
follows:

    #define alloca  __builtin_alloc

Their writing style is inconsistent.

Furthermore, regex.c has a #include "regex.c" line, which is including
itself.  regex.c is a strange and unnecessarily complicated source.

3.9.6.6     Macros Expanded to 'defined'

GNU C 3.2 sources do not have macros expanded to 'defined'.  According
to GNU C 3.2 documents, this type of macro is preprocessed in the same
way as GNU C 2/cpp, but GNU C 3.2 issues a warning to indicate
"may not portable".  However, when I tested, GNU C 3.2 did not seem to
issue a warning to an example shown in 3.9.4.6.

3.9.6.7  Preprocessing of .S Files

The gcc/config directory has several *.S files.

3.9.6.8     rpcgen and -dM Option

Make of GNU C 3.2 uses neither rpcgen nor -dM option.  However,
specifications for rpcgen and the -dM option do not seem to change from
the previous versions.

3.9.6.9     -include, -isystem and -I- Options

These options are frequently used in make of GNU C 3.2.  Sometimes, the
-isystem option is used to specify several system include directories at
one time.  Is it inevitable to use the option during software
compilation that updates system headers themselves?  I think they had
better use an environment variable to specify all the system include
directories.

On the other hand, GNU C 3/cpp documents discourage to use the -
iwithprefix and -iwithprefixbefore options.  GNU C provides many options
to specify include directories.  Does GNU C 3.2 move toward
reorganization or reduction in number of them? [1]

Note:
[1]  GNU C 3.2 Makefile uses the -iprefix option in a standalone manner
(without using -iwithprefix or -iwithprefixbefore), although the -
iprefix option makes sence only when used with one of these two options
following it.

3.9.6.10    Undocumented Predefined Macros

GNU C 2 did not document predefined macros, such as __VERSION__,
__SIZE_TYPE__, __PTRDIFF_TYPE__ and __WCHAR_TYPE__.  Even with the -dM
option, their existence was unknown.  GNU C 3 not only documents them
but also enhances -dM to show their values.

3.9.6.11    Undocumented Environment Variables

GNU C 3 documents the SUNPRO_DEPENDENCIES environment variable GNU C 2
did not.  (I do not know why this environment variable is needed.)

3.9.6.12    Other Problems

GNU C 3/cpp implements following #pragmas:

    #pragma GCC poison
    #pragma GCC dependency
    #pragma GCC system_header

Of these, GNU C 3.2 sources use poison and system_header.  MCPP does not
support these #pragmas because I do not think them necessary. (I omit
explanation of their specifications.)

GNU C 3 deprecates assertion directives, such as #assert, although gcc,
by default, specifies the -A option.

In GNU C 2, the -traditional option is implemented in one and the same
cpp, result in a strange mixture of very old specifications and C99 ones.
In GNU C 3, its preprocessor was divided into two: non-traditional cpp0
and tradcpp0. The -traditional option is valid only for gcc.  cpp0 does
not provides it.  gcc -traditional invokes tradcpp0 for preprocessing.

tradcpp0 is getting closer to a true traditional preprocessor before C90.
They say that they no longer maintain tradcpp0 except for serious bugs.

The strange specifications of GNU C 2/cpp seem to have been
significantly revised.

3.9.7   To Use MCPP with GNU C 3

As seen above, as far as preprocessing is concerned, GNU C 3.2 sources
have been much improved than glibc 2.1.3 sources in that the traditional
way of writing has been almost eliminated and that meaningless options
are no longer used.

GNU C 3.2/cpp itself is also much superior to GNU C 2/cpp in that it
regards traditional specifications as obsolete and articulates the token-
based principle.  Undocumented specifications have been significantly
reduced.  Although these improvements are not still sufficient, GNU C/
cpp is certainly moving toward the right direction.

However, GNU system headers become so complex that it is difficult to
grasp their entire structure, which may one of the biggest causes of
problems in the GNU system.

Another pitiful fact is that the preprocessor is absorbed into the
compiler-proper.   Therefore, to use MCPP, the -no-integrated-cpp option
must be specified when invoking gcc or g++.  If you compile a large
amount of source files with complicated or many makefiles, or if some
program automatically invoke gcc, you must create a shell-script that
invokes gcc or g++ with the -no-integrated-cpp option automatically
specified.

Let me take an example of this.  Place the following shell-scripts in
the directory where gcc and g++ reside (on my Linux, /usr/local/gcc-3.2/
bin), under the names of gcc.sh and g++.sh, respectively.

    #!/bin/sh
    /usr/local/gcc-3.2/bin/gcc_proper -no-integrated-cpp $@

    #!/bin/sh
    /usr/local/gcc-3.2/bin/g++_proper -no-integrated-cpp $@

Move to this directory and enter the following commands:

    chmod a+x gcc.sh g++.sh
    mv gcc gcc_proper
    mv g++ g++_proper
    ln -sf gcc.sh gcc
    ln -sf g++.sh g++

In the directory where cpp is located (on my Linux, /usr/local/gcc-3.2/
lib/gcc-lib/i686-pc-linux-gnu/3.2), create a script that executes MCPP
when cpp or cpp0 is invoked, as you did for GNU C 2 (See 3.9.5).  By
doing this, gcc or g++ first invokes MCPP and then invokes cc1 or
cc1plus with the -fpreprocessed option appended.  -fpreprocessed
indicates the source has been preprocessed.

In addition, as with other GNU C versions, you must set up environment
variables (See 3.9.5).  To use MCPP in GNU C 3.2, additional include
directory settings for C++, as well as PATH settings, are required.
Note that when a GNU C version other than the system standard one is
installed, additional include directory settings may be required.  MCPP
V.2.4 now contains these settings when MCPP itself is compiled, thus
eliminating the need to set them with environment variables.  On my
Linux, the /usr/local/gcc-3.2/lib line has been added to /etc/ld.so.conf,
and the following settings have been added to ~/.bash_profile.

    PATH=/usr/local/gcc-3.2/bin:$PATH
    #GCC_VERSION="\"2.95.3 20010315 (release)\""
    GCC_VERSION="\"3.2\""
    export PATH GCC_VERSION

If possible, I want to replace the ccplib source, the preprocessing part
of cc1 or cc1plus, with MCPP.  The source files that define the internal
interface between cpplib and ccl or ccpplus, as well as the external
interface between cpplib and user programs that use it, amount to as
much as 46KB.  It is impossible to replace.  Why is the interfaces so
complex?  It is pity.


3.10    Visual C++ .net System Header Problems

I used MCPP to preprocess some sample programs provided by Visual C++ .
net 2003.  The system headers seems to have only a few compatibility
problems shown below.  These problems are often seen in other compile
systems and do not have a serious impact on preprocessing.

  1.  Although Visual C++ .net system header scarcely implements the C99
    specifications, // comments are often used in source code.
  2.  Object-like macro definitions that are expanded into function-like
    macro names are sometimes found.
  3.  There is one erroneous macro definition in limits.h. (See Note 2
    in 4.1.3.1 of cpp-test.txt)

Although the Linux and glibc system headers often contain GNU C local
specification based coding, Visual C++ system headers scarcely has
Visual C++ local coding.

3.10.1  Comment Generating Macro?

I found only one outrageous macro in Visual C++.  Vc7/PlatformSDK/
Include/WTypes.h has the following macro definition:

    #define _VARIANT_BOOL   /##/

This macro definition is used in oaidl.h and propidl.h in Vc7/
PlatformSDK/Include/ as follows:

    _VARIANT_BOOL bool;

What does this macro aim at?

This macro seems to expect _VARIANT_BOOL to be expanded into // and the
line to be commented out.  Actually, this expectation is met in Visual C
cl.exe !

In the first place, // is not a token (preprocessing-token).  Macro
definitions should be processed and expanded after source are parsed
into tokens and a comment is converted into one space.  Therefore, it is
irrational for a macro to generate comments.  When this macro is
expanded into //, the result is undefined because // is not a valid
preprocessing-token.

In order to use these header files with MCPP, comment out these macro
definitions and change many _VARIANT_BOOL occurrences as follows:

    #if !__STDC__ && (_MSC_VER <= 1000)
        _VARIANT_BOOL bool;
    #endif

If you use only Visual C 5.0 or higher, this line can be commented out
as follows:

    // _VARIANT_BOOL bool;

This macro is, indeed, out of question, however, it is Visual C/cl.exe,
which allows such an outrageous macro to be preprocessed as a comment,
should be blamed.  This example reveals the following serious problems
this preprocessor has:

  1.  Preprocessing is not token-based but character-based.
  2.  The macro expansion result is treated as comment, which indicates
    the translation phases are pretty confusing.

Probably, the cl.exe preprocessor was developed based on a very old
character-based preprocessor.  It is easy to presume that the
preprocessor has been upgraded by repeating partial revision to the old
preprocessor.

There are many preprocessors which presumably have a very old program
structure.  GNU C 2/cpp, shown in 3.9, is one of such preprocessors.
Repeated partial revision of such a preprocessor will only makes its
program structure more complicated.  However much such revision may be
made, there are limits to quality such preprocessor can achieve.  Unless
a old source is given up and completely rewritten, a clear and well-
structured preprocessor cannot be obtained.

At GNU C 3/cpp, a total revision was made to GNU C 2; the entire source
code was rewritten.  So, GNU C 3/cpp has become quite different from GNU
C 2. Although MCPP was initially developed based on the source of an old
preprocessor, DECUS cpp, the source code was totally rewritten soon.


                  4.  Implementation-defined Behaviors

I have neither time nor space to write all the C preprocessor
specifications here.  For details on Standard C preprocessing, refer to
cpp-test.txt.  For MCPP behaviors in each pre-Standard mode, refer to 4.
1.3 of porting.txt.  This chapter covers several preprocessor-related
specifications, including those called implementation-defined by
Standard C.  For more details on MCPP specifications for each compiler
system, see Chapter 5, "Diagnostic Messages".


4.1     Status Value on Exit

The header file internal.H defines values returned by MCPP to a parent
process.  MCPP returns 0 on success, and errno for errno != 0 and 1 for
errno == 0 on error.  Success means by default that no error has
occurred, but it means no fatal error if MCPP is invoked with the -E
option.


4.2     Include Directory Search Path

This section explains the order in which MCPP searches directories for
an include file when it encounters a #include directive.


1. If a #include directive argument take a form of neither "file-name"
nor <file-name>, and is a macro, the macro is expanded.  The resulting
filename must take a form of either "file-name" or <file-name>.
Otherwise, it causes an error.

2. If the resulting filename, either in form of "file-name" or <file-
name>, is a full path name, MCPP tries to open it.   If it fails, it
causes an error.

3. If the resulting filename is not a full path but takes a form of
"file-name", MCPP regards it as a filename relative from the current
directory or source file directory, and begins searching from that
directory.  The former is a directory from which MCPP was invoked and
the latter is a directory where the source file that includes the
"file-name" resides.   Depending on the specified options and compiler
systems, MCPP begins searching directories as follows:

  If -I1 is specified, search begins from current directory.
  If -I2 is specified, source file directory.
  If -I3 is specified, current first and then source file directory.
  By default, MCPP ported to UNIX compiler systems, GNU C, Visual C or
    LSI C begins searching from the source file directory.
  For other MCPP, search begins at the current directory.  However, for
    MCPP ported to the Borland C of BC 4 or lower, search begins at
    current directory.  For BC 5, current first and then source file
    directory.
  If MCPP fails to find the desired file, it begins searching as shown
    in step 4.

In case of a nested #include, if search begins at current directory, the
base directory is always the same.  If search begins at a source file
directory, the base directory changes each time a header file resides in
other directory.

4. If the resulting filename is not a full path name but takes a form of
<file-name>, MCPP searches directories in the following order:

1. Directory (s) specified with the -I <directory> option on MCPP
invocation.  If several directories are specified, they are searched in
the order in which specified (with the left first).
2. For GNU ported MCPP, directories specified with the -isystem option.
If several directories are specified, they are searched in the order
specified (from the left).
3. Directories specified with an environment variable.
ENV_C_INCLUDE_DIR in system.H defines environment variable names.  In C+
+, ENV_CPLUS_INCLUDE_DIR, if defined, takes precedence over
ENV_C_INCLUDE_DIR.  GNU C ported MCPP uses C_INCLUDE_PATH (and also
CPLUS_INCLUDE_PATH for C++) as default environment variable.  The Plan 9
ported MCPP uses include as default. Other MCPP uses INCLUDE (and also
CPLUS_INCLUDE for C++) as default.  If an environment variable specifies
several directories with each separated with a delimiter, they are
searched in the order in which specified.  DOS/Windows, Plan 9 and other
OSs use ";", space and ":" as delimiter, respectively.
4. Implementation-specific directories defined by the C_INCLUDE_DIR
macros in system.H.  In C++, if CPLUS_INCLUDE_DIR is defined, its
directory is searched first.
5. Site-specific and system-specific directories defined by setsysdirs()
in system.c (For UNIX systems, /usr/local/include and /usr/include).

If any of the above directories are specified as a relative path using
"..", then MCPP regards it as a relative directory from the current one.
If MCPP fails to find or open the desired file after searching all the
directories in the above order, it causes an error.

The total number of include directories above must be equal to or less
than the number specified with NINCLUDE in system.H.  With the -I-
option (-nostdinc option for GNU C ported MCPP and -X for Visual C
ported MCPP), the directories specified in 3 and 4 of Step 4 are not
searched.

ANSI C Rationale says the ANSI committee intends to define a current
directory as base directory.  I think this is acceptable, in that the
base directory is always constant and that the specification is clearer.
However, some implementations, such as UNIX, seem to define a source
file directory as base one at least for #include "header".


4.3     How to Construct Header Name

This section explains how to construct a header-name pp-token and
extract a file name from it.

1. If source code contains a header file name in the string literal
format, MCPP regards it as a header-name and removes the " at the both
ends to construct a filename.  This can be applied to a string literal
resulting from macro expansion in source code.

2. If source code contains a header file name in the <filename> format,
MCPP regards it as a header-name and removes the < and > at the both
ends to construct a filename.

3. If source code contains a macro and it is expanded to <filename>,
MCPP removes the < and > at the both ends, as well as all the spaces, to
construct a filename.

4. In any case, MCPP converts \ to /, although both of "\" and "/" can
be used as path delimiters under DOS/Windows.

5. Under DOS/Windows, all the uppercased letters in file names are
converted into lowercased letters.

4.4     Evaluation of #if Expression

If both of host and target implementations have type "unsigned long"
(HAVE_ULONG == TRUE), MCPP evaluates a #if expression in "long" or
"unsigned long", as follows:

1. An integer constant token with a U suffix, including character
constants, is evaluated in unsigned long. (Note that MCPP compiled with
MODE == PRE_STANDARD does not recognize the U suffix).
2. Otherwise, a token within the range of positive long is evaluated in
long.
3.  Otherwise, a token within the range of unsigned long is evaluated in
unsigned long.
4.  Otherwise, it is evaluated to an out of range error.
5.  In a binary operation, if either operand is unsigned long, both are
converted to unsigned long.  Otherwise, an operation is performed in
signed long.

However, in the implementations having type "long long", if
___STDC_VERSION__ is set to 199901L or higher using the -S1 -V199901L
option, MCPP evaluates a #if expression in "long long" or
"unsigned long long", according to the C99 specification.  This can be
applied to C++ when MCPP is invoked with the -V199901L option.  In this
case, please read "long" and "unsigned long" hereinafter, until the end
of 4.5, as "long long" and "unsigned long long", respectively.

Visual C and Borland C 5.5 do not have a "long long" type, but have an
__int64 type of the same length.  So, a #if expression for C99 is
evaluated as __int64/unsigned __int64.  (However, since LL and ULL
suffixes cannot be used in Borland C 5.5, these suffixes must not be
used in coding other than #if lines.)

In addition, when you invoke MCPP with the -+ option for C++
preprocessing, MCPP evaluates pp-tokens 'true' and 'false' in a #if
expression to 1L and 0L, respectively.

Anyway, an integer constant token always has a positive value.  If
either host or target implementation does not have type unsigned long,
an integer constant token is evaluated within the range of positive long.
A token beyond that range is evaluated as an out of range error.  All
the operations are performed within the range of long.

If both of host and target implementations have type unsigned long and
the range of unsigned long of the host is narrower than that of the
target, a token beyond that host range is evaluated to an out of range
error.

If an operation using constant tokens produces a result out of range of
long, an out of range error occurs.  If it produces a result out of
range of unsigned long, a warning is issued.  This can be applied to
intermediate operation results.

Since a bitwise right shift of a negative value or a division operation
using it does not provide portability, MCPP issues a warning.  If an
operation using a mixture of unsigned and signed operands converts a
signed negative value to an unsigned positive value, a warning is also
issued.  How these values are processed depends on the specification of
the compiler-proper of the host implementation.

C90 makes it a rule that a preprocessor evaluates a #if expression in
long/unsigned long (in C99, the maximum integer type is used).  These
specifications are rougher than those of compiler-propers.  A (#)if
expression is often evaluated differently between preprocessor and
compiler-proper, especially when sign extension is involved.

In addition, since keywords are not used during Standard C preprocessing,
sizeof or cast cannot be used in a #if expression.  Of course, neither
variables, enumeration constants, nor floating point numbers can be used
there.  MCPP compiled with MODE >= STANDARD allows the "defined"
operator in a #if expression as well as the #elif directive.  Except for
these differences, MCPP evaluates a #if expression in accordance with
priority of and the associative law among operators, just as compiler-
propers do.  In a binary operation, an arithmetic conversion often takes
place to equalize the types on both-hand sides; If one operand is
unsigned long and the other is long, the both are converted to unsigned
long.

The Standard C <limits.h> shows the range of the types long and unsigned
long, however, the MCPP source code does not use <limits.h>.  This is
because so-called Standard C conforming compiler systems sometimes use
Standard incompliant <limits.h>, and in order to allow pre-Standard
compiler systems to compile MCPP.


4.5    Character Constant Evaluation in #if Expression

Constant tokens in a #if expression includes identifiers (macros and non-
macros), integer tokens and character constants.  How to evaluate
character constants is implementation-defined and lacks of portability.
Even (#)if 'const' is sometimes evaluated differently between
preprocessor and compiler-proper.  Note that Standard C does not even
guarantee that (#)if 'const' is evaluated to the same.

MCPP compiled with MODE == POST_STANDARD does not evaluate a character
constant in a #if expression, which is almost meaningless, and makes it
an error.

Like other integer constant tokens, MCPP evaluates a character constant
in a #if expression within the range of long or unsigned long.

A multi-byte character or a wide character is generally assumed to be
two bytes, except for the UTF-8 encoding, which is assumed to be four
bytes..  MCPP does not support EUC's 3 byte encoding scheme. (A 3-byte
character is recognized as 1 byte + 2 bytes.  The implementations using
3-byte encoding must have wchar_t of 4 bytes, which is different from
the type a wide character constant is normally assumed to have.  However,
a 3-byte multi-byte character is evaluated in 1 byte + 2 byte,
irrelevant to wchar_t, so it is evaluated correctly.)  This size
indicates a number of bytes on source code.  Although there are some
implementations using the 2-byte encoding scheme on source that treats
wchar_t as 4-byte, MCPP does not refer to wchar_t. The following
paragraphes describe two-byte multi-byte character encodings.

Multi-byte character constants, such as '', are evaluated to ((First
byte value << CHARBIT) + Second byte value).  CHARBIT has the value of
CHAR_BIT in <limits.h>.

Let me take an example of multi-character character constants, such as
'ab', '\x12\x3', and '\x123\x45'.  'a', 'b', '\x12', '\x3' and '\x123'
are regarded as one byte.  When a multi-character character constant is
evaluated, each one byte, starting from the highest one, is evaluated
within the range of [0, UCHARMAX] and combined by shifting it to left by
CHARBIT.  If the value of one escape sequence exceeds UCHARMAX, an out
of range error occurs.  Therefore, in the implementation with CHARBIT ==
8 and the ASCII character set, the above three tokens are evaluated to
0x6162, 0x1203 and error, respectively.

L'' is evaluated to the same value as ''.  Let me take an example of
multi-character wide character constants, such as L'ab', L'\x12\x3', and
L'\x123\x45'.  L'a', L'b', L'\x12', L'\x3', L'\x123', and L'\x45' are
regarded as one wide character.  When a multi-character wide character
constant is evaluated, each wide character, starting from the highest
one, is evaluated within the range of [0, (UCHARMAX << CHARBIT) |
UCHARMAX] and combined by shifting it to left by CHARBIT *2.  If the
value of one escape sequence exceeds the maximum value of an unsigned 2-
byte integer, an out of range error occurs.  Therefore, in the
implementation with CHARBIT * 2 == 16 and the ASCII character set, the
above three tokens are evaluated to 0x00610062, 0x00120003, and
0x01230045, respectively.

If the values of a multi-character character constant and a multi-
character wide character constant exceed the range of unsigned long, an
out of range error occurs.

With __STDC_VERSION__ or __cplusplus set to 199901L or higher, MCPP
evaluates a Universal Character Name (UCN) in the form of \uxxxx and \
Uxxxxxxxxa as a hex escape sequence. (I know this evaluation is nonsense
but no other way.)

If the compiler-proper of the target compiler system uses a signed char
or signed wchar_t, a character constant in a (#)if expression may be
evaluated differently between MCPP and compiler-proper.  The range that
causes a range error may also differ between them.  In addition,
evaluation of multi-character character constants and multi-byte
character constants varies even among preprocessors and among compilers.
Standard C does not stipulates whether, with CHAR_BIT set to 8, 'ab' is
evaluated to 'a' * 256 +'b' or 'a' + 'b' * 256.

In general, character constants should not be used in an #if expression,
as long as you have an alternative method.  I think an alternative
method always exists.


4.6     #if sizeof (type)

Standard C stipulates that preprocessing is a process independent of run-
time environments or compiler-proper specifications, thus prohibiting it
from using sizeof and cast in an #if expression.  However, MCPP compiled
with MODE == PRE_STANDARD allows sizeof (type) in a #if expression with
OK_SIZE set to TRUE in system.H.  This was done as a part of my effort
to add necessary modifications to the original version, such as adding
long long and long double processing, while retaining its original
functionality.  As to cast, I neither implemented nor had a will to do
so because it would require troublesome work.

How sizeof is evaluated depends on run-time environments, so you have to
be careful when using this.  Just like under MS-DOS, sizeof (type)
evaluation  varies from memory model to memory modes.  In addition, MCPP
does not recognize compiler system-specific modifiers, such as near and
far.  For other OSs with things like memory models, mem_model () in
system.c must be modified to enable sizeof.

A series of macros beginning with S_, such as S_CHAR, in evel.c define
the size of each type.  Under cross implementation, these macros must be
modified to specify size of the types, in integer values, used in the
target implementation.  The HAVE_LLONG and HAVE_LDBL macro only
indicates whether the compiler system allows sizeof (long long) or
sizeof (long double).  If the host and target implementations handle
these types differently, the macros, such as S_LLINT, and S_PLLINT, must
be modified in accordance with those of the target, regardless of macros
such as HAVE_LLONG.

I have to admit that MCPP does not provide the full functionality of #if
sizeof.  MCPP just ignores the letter of "signed" or "unsigned"
preceding char, short, int, long, and long long when it appears in a #if
sizeof.  So, when it encounters "signed" or "unsigned" in an
implementation without those types, such as unsigned char and unsigned
long, MCPP does not make it an error and returns the size of char, long.
In Visual C and Borland C 5.5, __int64 is expressed as "long long".  I
know this is a half-hearted implementation but I do not want to increase
the number of flags in system.H in vain for this Standard non-conforming
function.  I initially thought of removing the sizeof code from the
original version because I did not intend to support cast at all, but on
the second thought, I decided to make a small amount of modifications to
make use of the existing code.


4.7     How to Handle White-Space Sequence

MCPP compresses a white-space sequence, excluding <newline>, as a token
separator into one space character during tokenization in the
translation phase 3. It also deletes a white-space sequence at the end
of a line.  How to handle a white-space sequence at the beginning of a
line varies among implementations; some compress and other delete.

This compression and deletion occurs during the intermediate phase.  The
next phase 4 involves macro expansion and preprocessor directive line
processing.  Macro expansion may sometimes produce several space
characters before and after the macro.  Of course, the number of space
characters does not affect compilation results.

Standard C stipulates that whether cpp compresses a white-space sequence
into one space character during the translation phase 3 is
implementation-defined, but you usually do not have to worry about this.
<Vertical-tab> or <form-feed> in a preprocessor directive line may
adversely affect portability, since this is undefined in Standard C.
MCPP converts it to one space character.


4.8     Default Specifications for MCPP Executables

mcpp-2.3-bin.lzh, which is to be uploaded to Vector, contains several
MCPP executables for the following combinations of OSs and compiler
systems along with this manual:

This subsection describes the specifications of MCPP executables
generated when DIFfile and makefile for each compiler system in the
noconfig directory are used to compile MCPP with default settings.  When
a configure script is used to compile MCPP, the generated MCPP may
differ, depending on configure's results, however, as long as OS and
compiler system versions are same, generated MCPPs would be same except
for include directories.

DIFfiles and makefiles for the following combination of compiler systems
and OSs were used to compile MCPP in each compiler system:

    FreeBSD 4.7-R (GNU C V.2.95.4)
    Vine Linux 2.6-i386 (GNU C V.2.95.3 and GNU C V.3.2R)
    CygWIN 1.13 (GNU C V.2.95.3)
    DJGPP V.1.12 maint 4(GNU C V.2.7.1)
    LCC-Win32 V.3.2
    Visual C++ .net 2003
    Borland C++ V.5.5J/Win32
    Borland C++ V.4.02J/MS-DOS
    Trial version of LSI C-86 V.3.30c

MCPP executable for FreeBSD is dynamically linked to libc.so.4.0.  MCPP
executable for Linux is dynamically linked to libc.so.6.  MCPP for DJGPP
requires go32 of V.1.12 maint 3 or higher.  Borland C ported MCPP has
two versions: 32-bit version for BCC 5.5 and 16-bit version for BCC 4.0.
The 16-bit version is compiled to work with the large model.  LSI C-86
ported MCPP is compiled to work with the small model.

MCPP V.2 allows you to generate various MCPP executables that behave
according to various sets of specifications by changing macro
definitions in the system.H source.  Of these macros, MODE is most
important in that it determines the basic behaviors of MCPP.  This
subsection describes two types of MCPP executable: one compiled with
MODE == STANDARD and one with MODE == POST_STANDARD.  The STANDARD mode
is the mode in which MCPP behaves according to the Standard C
specifications, and the POST_STANDARD mode is a set of preprocessing
specifications I created in order to reorganize the Standard C
preprocessing specifications by eliminating or correcting ambiguous,
confusing, or irregular ones.  For details, see 4.1.3.1 of porting.txt.
You can use MCPP in the POST_STANDARD mode as Standard C preprocessor
for normal programs.

Of all the macros defined in noconfig.H and system.H, the settings of
those mentioned below are identical among every MCPP executable,
regardless of their compiler systems and modes (STANDARD or
POST_STANDARD).

HAVE_PRAGMA is set to TRUE since their compiler-propers accept the #
pragma line.

These executables being Standard C conforming preprocessors,
OLD_PREPROCESSOR, COMMENT_INVISIBLE, STRING_FORMAL, OK_IF_JUNK, OK_SIZE,
and OK_UNTERM_STRING were set to FALSE on compilation.  DOLLAR_IN_NAME
is also set to FALSE, so $ cannot be used in names.

CONCAT_STRINGS is set to FALSE, so string literals are not concatenated.

MCPP compiled with OK_DIGRAPHS == TRUE and DIGRAPHS_INIT == FALSE
enables digraph when the -2 (-digraphs) option is specified.

With OK_PRAGMA_OP set to TRUE, the -V199901L option enables the _Pragma()
operator.  With OK_MBIDENT set to FALSE, multi-byte-characters cannot be
used in identifiers.

With DEBUG and DEBUG_EVAL set to TRUE, the #pragma __debug_cpp directive
outputs various debug information.

With OK_TRIGRAPHS == TRUE and TFLAG_INIT == FALSE in MODE == STANDARD,
trigraph is enabled with the -3 (-trigraphs) option.

With OK_UCN set to TRUE in MODE == STANDARD, Universal Character Name
(UCN) can be used in C99 and C++.

In MODE == POST_STANDARD, neither trigraphs nor UCN can be used.

With TOP_SPACE set to TRUE in MODE == STANDARD, a white-space sequence
at the beginning of a line is basically compressed into one space
character.  On the other hand, with TOP_SPACE set to FALSE in MODE ==
POST_STANDARD, a white-space sequence at the beginning of a line is
removed.

The settings of the macros below are different among compiler systems.
( BC4/16 refers to the 16 bit version of Borland C 4.0.)

    STDC (Initial value of __STDC__)
        LSI C:  0
        Others: 1
    STDC_VERSION (Initial value of __STDC_VERSION__)
        GNU C 2: 199409L
        Others: 0L
    HAVE_DIGRAPHS   (Is digraphs output as it is?)
        GNU C and Visual C: TRUE
        Others: FALSE
    EXPAND_PRAGMA   (Is a #pragma line macro-expanded in C99?)
        Visual C: TRUE
        Others : FALSE
    OK_MAKE (Is option to output lines for a makefile implemented?)
        LSI C: FALSE
        Others :   TRUE
    IDMAX (Valid length of identifier)
        LSI C : 63
        BC4/16: 255
        Others: 1024
    NMACPARS (Maximum number of macro arguments)
        LSI C : 127
        Others: 255
    NEXP (Maximum number of nested levels of #if expressions)
        LSI C : 64
        Others: 256
    BLK_NEST (Maximum number of nested levels of #if section)
        LSI C : 64
        Others: 256
    RESCAN_LIMIT (Maximum number of nested levels of macro rescans)
        LSI C : 32
        Others: 64
    NINCLUDE (Maximum number of include directories)
        LSI C, BC4/16: 16
        Others: 32

Both of NBUFF, which specifies the maximum length of a source line after
converting a comment into a space character and line splicing, and NWORK,
which specifies the maximum length of an output line, are set to the
following values (They are set to the same values):

    LSIC : 1024
    BC4/16: 4096
    DJGPP : 16384
    Others: 65536

NMACWORK, which specifies the size of internal buffers used for macro
expansion, is set to four times as long as NBUFF and NWORK.

No restrictions are imposed on the number of nested levels of #include,
which means it can exceed the limit imposed by OS of the number of
simultaneously opened files.

MCPP for LSI C, which works for the small model, has lower translation
limits.

GNU C V.2.6.2 implemented neither __STDC_VERSION__ nor digraphs but V.2.
7.1 did, and __STDC_VERSION__ is now set to 199409L. However, in GNU C V.
3.x, __STDC_VERSION__ is no longer predefined by default and is now
defined in accordance with an execution option.  GNU C ported MCPP
enables both of __STDC_VERSION__ and digraphs.

If STDC is set to 0, MCPP predefines __STDC__  as 0.  If STDC_VERSION is
set to 0L, MCPP predefines __STDC_VERSION__ as 0L.  So, specifying the -
S1 -V199409L option sets __STDC__ and __STDC_VERSION__ to 1 and 199409L,
respectively and allows only predefined macros that begin with '_',
resulting in MCPP in the strictly C95 conforming mode.  The -V199901L
option specifies C99 mode.

In C99 mode, MCPP predefines __STDC_HOSTED__ as 1.

MCPP itself predefines neither __STDC_ISO_10646__, __STDC_IEC_559__ nor
__STDC_IEC_559_COMPLEX__.  These values are compiler system-specific.
In glibc 2/ x86, the system header defines __STDC_IEC_559__ and
__STDC_IEC_559_COMPLEX__ as 1.  Other compiler systems do not define
them.

With HAVE_DIGRAPHS set to FALSE, digraph is output after conversion.

Include directories are set as follows:

System-specific or site-specific directories under UNIX or UNIX-like OSs
are as follows:
        FreeBSD, Linux and CygWIN:  /usr/include, /usr/local/include
        Plan 9:     /sys/include/ape, /$objtype/include/ape

Implementation-specific directories that vary among compiler systems and
their versions are as follows:

    C_INCLUDE_DIR1 macro is set to:
        FreeBSD / GNU C 2.95.4 :
                /usr/lib/gcc-lib/i386-unknown-freebsd4.7/2.95.4/include
        Linux / GNU C 2.95.3   :
                /usr/lib/gcc-lib/i386-redhat-linux/2.95.3/include
        Linux / GNU C 3.2  :
          /usr/local/gcc-3.2/lib/gcc-lib/i686-pc-linux-gnu/3.2/include
        CygWIN / GNU C 2.95.3-5 :
                /usr/lib/gcc-lib/i686-pc-cygwin/2.95.3-5/include
        BCC5.5 :   /BCC55/INCLUDE
        BCC4   :   /BC4/INCLUDE
        LSIC   :   /LSIC86/INCLUDE
    C_INCLUDE_DIR2 macro is set to:
        Linux / GNU C 3.2  :   /usr/local/gcc-3.2/include
    CPLUS_INCLUDE_DIR1 macro is set to:
        FreeBSD :   /usr/include/g++
        Linux  :   /usr/include/g++-3
        CygWIN :   /usr/include/g++-3
    CPLUS_INCLUDE_DIR2 macro is set to:
        Linux / GNU C 3.2  :   /usr/local/gcc-3.2/include/c++/3.2

The macros that define these directories are found in noconfig.H or
config.h.

MCPP for Visual C, LCC-Win32 or DJGPP does not use these macros but uses
the following environment variables:
    Visual C, LCC-Win32:    INCLUDE, CPLUS_INCLUDE
    DJGPP:      C_INCLUDE_PATH, CPLUS_INCLUDE_PATH

If these default settings do not suit you, change settings to recompile
MCPP, or use environment variables or the -I option.

In LSI C, the lcc command uses settings in the _lcc file.  _lcc usually
has include directory settings.   If the settings agree with MCPP or you
implemented MCPP in the way that they agree, you do not need to write
settings in _lcc.  Under MS-DOS, a command line length is likely to
exceed the command.com limit, so it is recommended that the setting in
_lcc should be deleted.  Using environment variables also eliminates the
need to set _lcc, in which case you must ensure that the same
environment variable is not used by other compiler systems.

When the length of a preprocessed line exceeds NWORK-1, MCPP divides it
into several parts so that each part length becomes equal to or less
than NWORK-1.  A string literal length must be equal to or less than
NWORK-2.

Again for confirmation, the macros mentioned above are used only to
compile MCPP, and are not built-in macros in a MCPP executable.

MCPP has the following built-in macros.  The (=value) below indicates
that the macro is set to the value.  Macros without (=value) are defined
as 1.

With __STDC__ set to 1 or higher, the macros that do not begin with _
are deleted.  The -N (-undef) option deletes all the macros other than
__MCPP.  After -N, you can use -D to defines macro symbols over again.
When you use a different compiler system version from those specified
here, -N and -D allow you to redefine your version macro without
recompiling MCPP.  The -D option allows you to specify a particular
macro without using -N or -U.  This is enough for redefining other BC
version macros.

When you invoke MCPP without an input file and enter #pragma
__put_defines, the following built-in macros are displayed:

    FreeBSD / GNU C 2.95:   __i386__, unix, __unix, __unix__,
                __FreeBSD__ (=4), __GNUC__ (=2), __GNUC_MINOR__ (=95),
                __MCPP (=2), __SIZE_TYPE__ (=unsigned int),
                __PTRDIFF_TYPE__ (=int), __WCHAR_TYPE__ (=int)
    Linux / GNU C 2.95:     __i386__, unix, __unix, __unix__, __linux__,
                __GNUC__ (=2), __GNUC_MINOR__ (=95), __MCPP (=2)
                __SIZE_TYPE__ (=unsigned int), __PTRDIFF_TYPE__ (=int),
                __WCHAR_TYPE__ (=long int)
    Linux / GNU C 3.2:  __i386__, unix, __unix, __unix__, __linux__,
                __GNUC__ (=3), __GNUC_MINOR__ (=2), __MCPP (=2)
                __SIZE_TYPE__ (=unsigned int), __PTRDIFF_TYPE__ (=int),
                __WCHAR_TYPE__ (=long int)
    CygWIN 1.3.10:  __i386__, __CYGWIN__, __CYGWIN32__,
                __GNUC__ (=2), __GNUC_MINOR__ (=95), __MCPP (=2),
                __SIZE_TYPE__ (=unsigned int), __PTRDIFF_TYPE__ (=int),
                __WCHAR_TYPE__ (=short unsigned int)
    DJGPP V.1.12:   __i386__, __GO32__, __DJGPP__, __GNUC__ (=2),
                __GNUC_MINOR__ (=7), __MCPP (=2),
                __SIZE_TYPE__ (=long unsigned int),
                __PTRDIFF_TYPE__ ( = long int), __WCHAR_TYPE__ ( = int)
    LCC-Win32:  __i386__, __WIN32__, WIN32, _WIN32, __FLAT__,
                __LCC__, __LCCDEBUGLEVEL (=0), __LCCOPTIMLEVEL (=0),
                __MCPP (=2)
    VC .net 2003:   __i386__, __WIN32__, _WIN32,  WIN32, __FLAT__,
                _MSC_VER (=1310), _MSC_EXTENSIONS, _M_IX86 (=600),
                __MCPP (=2)
    BC 5.5 / 32 bits:   __i386__, __WIN32__, WIN32, __FLAT__,
                __BORLANDC__ (=0x0550), __TURBOC__ (=0x0550),
                __MCPP (=2)
    BC 4.0 / 16 bits:   __8086__, MSDOS, __MSDOS__, __SMALL__,
                __BORLANDC__ (=0x0452), __TURBOC__ (=0x0452),
                __MCPP (=2)
    LSI C-86:   __8086__, MSDOS, __MSDOS__, __SMALL__, __LSIC__, LSI_C,
                __MCPP (=2)
    Plan 9 / pcc:   unix, __unix, __unix__, _PLAN9, __MCPP (=2)

When you use the -+ (-lang-c++) option to specify C++ preprocessing,
__cplusplus is predefined with its initial value of 1L.  In addition,
the following macros are also predefined:

    GNU C V.2.*:    __GNUG__ (=2)
    GNU C V.3.*:    __GNUG__ (=3)

    BC 4.0:     __BCPLUSPLUS__ (=0x0320)
    BC 5.5:     __BCPLUSPLUS__ (=0x0550)

Although there are many predefined macros in GNU C, those predefined by
GNU C/cpp are few.  Most of them are passed from gcc to cpp by the -D
option.  So, it is not necessary for MCPP to define them to avoid
duplicate definition.  However, MCPP predefines these macros for being
used in a stand-alone manner, such as pre-preprocessing.  (MCPP does not
go so far as to predefine __STRICT_ANSI__ even with the -S1 option). [1]

With the -S1 option, MCPP for LSI C undefs LSI_C, so it is recommended
that you should define that macro in _lcc.

Since FreeBSD, Linux, CygWIN/ GNU C, LCC-Win32 and Plan 9 / pcc have a
type long long, with the -V199901L option, an #if expression is
evaluated in long long or unsigned long long.  Visual C and Borland C 5.
5 do not have a "long long" type but __int64 and unsigned __int64
instead.  These types are used.  In other compiler systems, it is
evaluated in long or unsigned long, just as in C90.

In the above compiler systems with type long ranges:
    [-2147483647-1, 2147483647 ([-0x7fffffff-1, 0x7fffffff])
and
unsigned long ranges:
    [0, 4294967295 ([0, 0xffffffff]).

In the compiler systems with type long long ranges:
    [-9223372036854775807-1, 9223372036854775807
   ([-0x7fffffffffffffff-1, 0x7fffffffffffffff]),
and
unsigned long long ranges:
    [0, 18446744073709551615 ([0, 0xffffffffffffffff]).

All the compiler-propers of the above compiler systems internally
represent a signed integer as two's complement number.  So do bit
operations.  This can be applied to MCPP's #if expression.

Right shift of a negative integer is an arithmetic shift.  This can be
applied to MCPP's #if expression. (Right shifting an integer by one bit
halves the value with the sign retained)

In an integer division or modulus operation, if either or both operands
are negative values, an algebraic operation like Standard C
's ldiv () function is performed.  This can be applied to MCPP's #if
expression.

These OSs use the ASCII basic character set.  So does MCPP.  The default
multi-byte character encoding is EUC-JP for FreeBSD or Linux ported MCPP,
UTF-8 for Plan 9 ported MCPP and shift-JIS for other MCPP.

There is a memory management routine, kmmalloc, that I developed.  This
routine has malloc(), free(), realloc() and other memory handling
functions.  If kmmalloc is installed in systems other than CygWIN,
kmmalloc is static linked when the MALLOC=KMMALLOC (or -DKMMALLOC=1)
option is specified in make; kmmalloc is compiled in the _MEM_DEBUG mode
and linked to a heap memory debugging routine.  (The XMALLOC macro is
also defined as 1.)  MCPP for Linux, DJGPP, and LCC-Win32 uses EFREEP,
EFREEBLK, EALLOCBLK, EFREEWRT and ETRAILWRT with an errno of 2120, 2121,
2122, 2123 and 2124 assigned, and other MCPP uses 120, 121, 122, 123,
and 124. (Refer to 4.extra of cpp.doc.)  [2]

On the systems other than GNU and Visual C, you must preset the
environment variable TZ to JST-9. Or, the __DATE__ and  __TIME__ macros
are not set correctly.

Note:
[1] The macros predefined by FreeBSD 2.2.7/ GNU C V.2.7.2.1/cpp are
__FreeBSD__ , unix, and i386 only.  For DJGPP 1.12/GNU C V.2.7.1/cpp,
GO32, MSDOS, unix, and i386 only.  Even these macros are so defined that
they are overridden when the -undef option is passed from gcc.

[2] CygWIN 1.13 provides malloc () that has an internal routine named
_malloc_r() which is called by other library functions.  So this malloc()
cannot be replaced with my memory.c.


                        5.  Diagnostic Messages

5.1     Diagnostic Messages Format

This section covers diagnostic messages issued by MCPP, as well as their
meaning.  By default, these messages are output to stderr.  With the -Q
option, they are redirected to the cpp.err file in the current directory.
A diagnostic message is output in the following manner:

1. "filename: line: " is followed by "fatal error: ", "error: " or
"warning: " and then by any of the diagnostic messages shown in sections
5.3 to 5.9.  Although the specification that a diagnostic message must
fit in one line that begins with "filename: line:" seems to lack of
flexibility, I followed, although unwillingly, because it is a
traditional way of implementing messages in C on UNIX and because
various tools have already assumed that.  Some MCPP messages do not fit
in a line.

2. If an error occurs during macro expansion, the macro invocation is
displayed.  For nested macro invocations, MCPP traces back the nested
macros in the reverse order to display each macro names.  The MCPP
compiled with DEBUG == TRUE displays the macro definition, as well as
the source filename and line number where the macro definition is found.

3. The source file name, the line number and the line5 at which an error
has occurred are displayed.  If an error has occurred in an included
file, the names, line numbers and the #include lines6 of all the
including files are displayed.  Usually, a logical line with comments
replaced with a space character is displayed.  The logical line is
constructed from one or more physical lines with \ at the line end.  If
a comment spreads over several lines, several logical lines are
concatenated into one, which is displayed as the line.  In this case,
the line number of the last concatenated physical line is displayed
Note that if an error occurs during the translation phase before
processing a comment, the line in the phase is displayed.

If the -j option is specified, MCPP outputs neither the above 2 nor 3.

Diagnostic messages are divided into three levels:

fatal error: Indicates an error is so serious that it is no longer
    meaningful to continue preprocessing.
error: Indicates there is a syntax or usage error.
warning: Indicates code lacks of portability or may contain a bug.

Warnings are further divided into five classes:

Class 1: Source code may contain a bug.
Class 2: Probably, source code will present no problem in practical use,
    but is problematic in terms of Standard conformance.
Class 4: Probably, source code will present no problem in practical use,
    but is problematic in terms of portability.
Class 8: Rather surplus warnings to #if groups skipped, sub-expression
    in #if expression whose evaluation is skipped, concatenation of
    string literals, and etc.
Class 16: Warning to trigraphs and digraphs.

Warnings other than Class 1 or 2 are rather specific to MCPP.

MCPP has various types of diagnostic messages.  For example, MODE ==
STANDARD MCPP provides the following types of diagnostics for each level
and class. (This number will increase if they are further subdivided
into smaller categories.)

    fatal error: 19 types
    error: 75 types
    warning class 1: 34 types
    warning class 2: 8 types
    warning class 4: 15 types
    warning class 8: 23 types
    warning class 16: 2 types

Principally, these messages point the coding in question.  The
diagnostic messages below have a sample value embedded in a token or a
numeric value from source code.  For the messages with a macro name
embedded, a value the macro is expanded into is shown in real messages.

Depending on cases, a same message is issued as warning or error, in
which case, this manual gives the first occurrence a detailed
description.  For the subsequent occurrences, the message is only listed.

Note that under DOS/Windows, MCPP converts all path-lists and file names
in diagnostic messages into lowercased letters for normalization.


5.2     Translation Limits

Of all the errors shown below, some errors, such as a buffer overflow,
occur due to MCPP specification restrictions.  Some macros in system.H
define translation limits, such as a buffer size.  Expand the buffer
size and recompile MCPP if necessary, however, be careful not to expand
it too much.  A large buffer in a system8 with a limited amount of
memory may cause an "out of memory" error frequently.


5.3     Fatal Errors

A fatal error occurs and preprocessing is terminated when it is no
longer possible to continue preprocessing due to an I/O error or a
shortage of memory, or it is no longer meaningful to do so due to a
buffer overflow.  A status value of failure is returned to a parent
process.

5.3.1   MCPP's Own Bugs

Bug:
    This message has several types.  Should it be issued, it would
    indicate MCPP's own bug.  The only MCPP compiled with DEBUG and/or
    DEBUG_EVAL set to TRUE issues this message.  I think this message is
    rarely issued, but should it be issued, do not hesitate to let me
    know the situation.

5.3.2   Physical Errors

File read error
    An error has occurred during reading a source file.  Disk may have
    been damaged.

File write error
    An error has occurred during writing to a file.  Disk may have been
    damaged or full.

Out of memory (required size is 0x123 bytes)
    Runs short of memory.  MCPP tried to obtain memory of 0x123 bytes
    from the heap, but in vain. This error occurs when there are too
    many long macro definitions on a system with a small amount of
    memory.  Divide your source file to decrease the number of macro
    definitions in one translation unit.

5.3.3   Translation Limits and Internal Buffer Errors

Too long header name "long-file-name"
    The length of the full path name of a file to include (file name
    concatenated with the specified directory path) has exceeded
    FILENAMEMAX or NWORK

Too long source line
    The length of a physical line in source file has exceeded NBUFF-2.
    The source code may not be written in C.

Too long logical line
    The length of a logical line, which is constructed from the several
    physical lines with \ at the line end, has exceeded NBUFF-2.  This
    error may occur when a defined macro is too long.  The code should
    be written not as a macro but as a function.

Too long line spliced by comments
    The length of a preprocessed line with a comment replaced with a
    space character has exceeded NBUFF-2.  This error occurs when
    several lines are concatenated into one if a comment spreads over
    several lines.  Divide the comment into several parts and write each
    on a separate line.

Too long output line
    The length of a preprocessed line has exceeded NWORK - 2.  Several
    long macro calls may be contained in a line.  Divide the line.

Too long token
    A preprocessed line has a token with a length more than NWORK - 2.
    MCPP compiled with NWORK < NMACWORK tries to divide the preprocessed
    line into a length that the compiler-proper can accept.  However, if
    a line contains a too long token, it sometimes fails to do so.

    The following four errors may also be caused by a buffer overflow at
    a token that is not so particularly long during macro expansion, in
    which case, you must divide the macro invocation.  A buffer overflow
    may also occur during concatenation of string literals, in which
    case you must divide the line.

Too long quotation "long-string"
    A string literal, character constant or header-name is too long.  In
    case of a string literal, divide it.  Standard C conforming compiler
    systems concatenate adjacent string literals for you. (Although,
    MCPP compiled with CONCAT_STRINGS == TRUE does it.)

Too long pp-number token "1234567890toolong"
    A preprocessing-number token is too long.  MODE >= STANDARD MCPP
    issues this error.

Too long number token "12345678901234......"
    A number token is too long.  MODE == PRE_STANDARD MCPP issues this
    error.

Buffer overflow scanning token "token"
    A buffer overflow has occurred during token scan.  This message is
    issued to tokens other than string literals, character constants,
    header-names and pp-numbers.

More than BLK_NEST nesting of #if (#ifdef) sections
    The depth of nested #if, #ifdef, and #ifndef has exceeded BLK_NEST.
    (In real message, the macro name BLK_NEST is replaced with an actual
    numerical value. This is applied to all the messages below with a
    macro name embedded.)  Divide the #if section.

Too many include directories "dir"
    The number of include directories specified has exceeded NINCLUDE.

Too many include files
    The number of #included header files preprocessed in one source file
    has exceeded NINCLUDE*4.  Duplicated included header files are
    counted as one.

5.3.4   #pragma __preprocessed Related Errors

This is not the preprocessed source
    Although the "#pragma __preprocessed" directive is found, this is
    not a source preprocessed by MCPP.

This preprocessed file is corrupted
    This seems to be a source preprocessed by MCPP, but cannot be used
    because it is destroyed.


5.4     Errors

MCPP issues an error message when it found a grammatical error.
Standard C stipulates that a compiler system should issue a diagnostic
message when they encounter a violation of syntax rules or constraints.
Principally, MODE == STANDARD MCPP issues an error message to this type
of violation, but sometimes issues a warning.

MCPP issues an error message to most of undefined items in Standard C,
and a warning to some.  However, MCPP issues neither an error nor a
warning to the following undefined items:

1.' or /* in a header name in the form of a string literal - MCPP
regards them as characters, resulting in a file open error. (' or /* in
a header name enclosed with < and > is regarded as the beginning of a
character constant or a comment, resulting in some errors.)  Although
how to treat \ in a header name is undefined in Standard C, MCPP does
not check it because it may eventually cause an error when MCPP actually
tries to open the file.  DOS/Windows ported MCPP issues a class 2
warning to \ and converts it to /.

2. #undef defined - Although #undefing a name "defined" yields an
undefined result, MCPP does not issue a message because, in the first
place, MCPP does not allow definition of a macro name "defined", so it
does not think of revoking the definition.

3.Illegal multi-byte character sequence in a comment - Although how to
deal with such character sequence is undefined in Standard C, MCPP does
not issue a message because it does no harm. (MCPP issues a warning to
an illegal multi-byte character sequence in string literals, character
constants and header names.)

4. Identifiers that begin with _ (Reserved for compiler systems) -
Although using these identifiers in a user program will cause an
undefined result, MCPP does not check it because MCPP does not always
have a means to decide whether these identifiers are used in a user
program or the compiler-system.

5.__STDC_ISO_10646__, __STDC_IEC_559__, and __STDC_IEC_559_COMPLEX__ -
Although #defining or #undefing these optional C99 predefined macros
yields an undefined result, MCPP does not check it because MCPP does not
always have a means to determine whether these macros appear in a user
program or the compiler-system. (These macros are most likely to be
defined in a header file of a compiler system.)

6. UCN equivalent sequence - Although it is undefined in C99 how to deal
with a UCN equivalent sequence generated by deleting <backslash><newline>
during the translation phase 2 or by concatenating string literals, MCPP
does not issue a message and regards it as a UCN.

For details on what is a violation of syntax rule or constraint,
undefined, unspecified or implementation-defined in Standard C
preprocessing, refer to cpp_test.txt.

Even if an error occurs, MCPP continues preprocessing as long as they
are not fatal one.  MCPP shows the number of errors and returns the
status of failure to the parent process when it exits.  MCPP invoked
with the -E option always returns the status of success.

5.4.1   Character and Token Related Errors

Illegal control character 0x1b, skipped the character
    A control code other than a white space character is found in a
    string literal, character constant, header name or comment.  MCPP
    skips it and continues preprocessing.

The following five messages are all token-related errors.  For the first
four, MCPP skips the line in question and continues preprocessing.  The
first three are string literal or other token-related errors, indicating
that a closing quotation mark is not found by the end of the logical
line.  This type of error occurs when you write a text that does not
take a form of a preprocessing-token sequence in neither a string
literal nor comment, as shown below:

    #error I can't understand.

As processing-tokens are not so strictly defined as C tokens in the
compiler-proper, most character sequences are regarded as token
sequences, as long as they belong to a source character set.  Therefore,
it is only this type of coding that causes a preprocessing-token error.
Pp-token errors may occur in a skipped #if group.

Unterminated string literal "string
    A string literal is unterminated.  A string literal cannot spread
    over several logical lines.  If necessary, write a string literal on
    each of several lines and have the compiler concatenate them.  This
    error will occur during conversion into a string by a #operator, in
    which case the line in question is not skipped.  MCPP compiled with
    MODE == PRE_STANDARD and OK_UNTERM_STRING == TRUE does not make an
    unterminated string literal an error. (Instead, it regards the line
    end as literal end.)  Nor does MCPP compiled with MODE !=
    POST_STANDARD and TOP_SPACE == TRUE and invoked with the -a (-lang-
    asm, -x assembler-with-cpp) option (it issues a warning); it regards
    an unterminated string literal as a literal spreading over several
    lines and concatenates a line with the next by inserting \n.

Unterminated character constant 't understand.
    A character constant is not terminated.  MCPP compiled with
    OK_UNTERM_STRING does not make it an error. (Instead, it regards the
    line end as literal end.)

    Unterminated header name <header.h A header-name is not terminated.
    " or ' in a header-name enclosed with < and > causes the above two
    errors, not this one.  If /* is found in a header-name enclosed with
    < and >, MCPP regards it and the following text as a comment.

Empty character constant ''
    A character constant is empty.

Illegal UCN sequence
    MCPP compiled with MODE == STANDARD and invoked with
    __STDC_VERSION__ set to 199901L or in C++ mode can recognizes UCN.
    This message is issued when the number of orders of a hex sequence
    that begins with \u and \U in an identifier is less than four and
    eight, respectively.  (If this occurs in a character constant in a #
    if expression, an undefined escape sequence warning results.  Other
    tokens are not checked by MCPP.)

UCN cannot specify the value "0000007f"
    UCN cannot specify a hex value in the ranges of 0 to 9f, except for
    0x24 ($), 0x40 (@) and 0x60 (`), and of d800 to dfff.  The former
    range agrees with the range of the basic source character set.  The
    latter range falls into the reserved area for special characters.
    Note C++ does not have the latter restriction. (Specifications
    slightly differ among Standards for an unknown reason.)  However,
    when MCPP is invoked as C++ with -V199901L to preset the __cplusplus
    macro to 199901L or higher, MCPP behaves in accordance with the C99
    specifications in this respect.

Illegal multi-byte character sequence "XY"
    MCPP compiled with MODE == STANDARD and OK_MBIDENT == TRUE allows
    for a multi-byte character in an identifier in C99, however, it will
    cause an error when it finds a character sequence that cannot be
    regarded as a multi-byte character.

5.4.2   Unterminated Source File Related Errors

This section covers messages issued when a source file ends with an
unterminated line, comment, #if section or macro invocation.  If the
file (not included file) marks the end of input, the message
"End of input", not "End of file", is issued.

These diagnostic messages are issued as an error or warning, depending
on MCPP modes.

MCPP compiled with MODE >= STANDARD issues these messages as error, in
which case MCPP skips the line or macro invocation in question and
restores relationship between paired directives in a #if section to that
of when the file is initially included.

On the other hand, MODE == PRE_STANDARD MCPP does not check a source
file that ends with \.  It regards a file that ends with no <newline> or
an unterminated comment as an error.  For a file that ends with an
unterminated #if section, #asm block, or macro invocation, MCPP compiled
with OLD_PREPROCESSOR set to TRUE issues a warning, while MCPP with
OLD_PREPROCESSOR set to FALSE issues a warning at the end of the
included file and an error at the end of input.  The latter MCPP ignores
these errors and warnings to continue processing until it reaches the
end of input as if the included file were actually in the including file,
causing many unexpected results.

End of file with no newline, skipped the line
    A file must be terminated with a newline.

End of file with , skipped the line
    A file must not be terminated with <backslash><newline>.

End of file with unterminated comment, skipped the line
    A comment is not terminated.

End of file within #if (#ifdef) section started at line 123
    #if (#ifdef or #ifndef) on the line 123 does not have a
    corresponding #endif.

End of file within macro invocation started at line 123
    A macro invocation that begins at the line 123 is not terminated by
    the end of the file.  This error may occur when an argument has an
    ill-balanced parenthesis, or when a token error occurs between
    opening and closing parentheses, in which case, MCPP continues to
    read tokens for a corresponding parenthesis until it reaches to the
    file end. (Probably, a buffer overflow may occur before reaching
    there.)  In addition, since macro expansion specifications vary
    among modes, a macro that is successfully expanded in a mode may not
    in other modes.

End of file with unterminated #asm block started at line 123
    #asm on the line 123 does not have a corresponding #endasm.

5.4.3   Ill-Balanced Preprocessing Group Related Errors

This sections covers errors caused by ill balanced directives of #if, #
else and etc. Even if MCPP finds ill balance among these directives, it
continues processing, assuming that the processing group so far still
continues.  MCPP checks to see if directives are balanced even in a
skipped #if group.

The #if (#ifdef) section is a block between #if (#ifdef or #ifndef) and
#endif.  The #if (#elif, #else) group is a smaller block, say, between #
if (#ifdef or #ifndef) and #elif, between #elif and #else, or between #
else and #endif within the #if (#ifdef) section.

Already seen #else at line 123
    Another #else (#elif) is found after #else at the line 123.  #endif
    may be missing.

Not in a #if (#ifdef) section
    #else (#elif, #endif) is found without #if (#ifdef or #ifndef).

Not in a #if (#ifdef) section in a source file
    An included file has #else (#elif or #endif) without #if (#ifdef or
    #ifndef).  If the included file in question had been in the
    including source file, this error would never have occurred.  In
    other words, each of these directives contained in a separate file
    is not balanced by itself.  The only MODE >= STANDARD MCPP issues
    this error. (MODE == PRE_STANDARD MCPP issues a warning.)

The following two errors occur when #asm and #endasm are not balanced.
These messages are issued only by MCPP compiled with MODE ==
PRE_STANDARD and ported to a particular compiler system.

In #asm block started at line 123
    A #asm block that begins at the line 123 has another #asm.  #asm
    cannot be nested.  Maybe, the programmer forgot to write #endasm.

Without #asm
    #endasm is found in a non #asm block.

5.4.4   Simple Syntax Errors on Directive Lines

This section covers simple syntax errors on control lines that begin
with #.  The errors hereinafter discussed until 5.4.12 do not occur
within a skipped #if group.  MCPP invoked with the -W8 option issues a
warning to an unknown directive.

When MCPP finds a control line with a syntax error, it ignores the line
and continues processing, in which case, it neither regards #if as the
beginning of a section nor changes line numbers even with a #line.  If a
#include or #line line has a macro argument, MCPP expands the macro and
checks the syntax.  MODE == PRE_STANDARD MCPP do not expand the macro.

Although the messages below do not show the directive name in question,
the source line that follows the message show it. (A control line with a
comment converted in a space character always becomes one line, which is
called "preprocessed line" here.)

Illegal #directive "123"
    A token that immediately follows # is not a name.  The token must be
    a directive name. (MCPP compiled with OLD_PREPROCESSOR == TRUE
    regards #123 as #line 123.)

Unknown #directive "pseudo-directive"
    The directive "pseudo-directive" is not implemented.  MCPP invoked
    with the -a (-lang-asm or -x assembler-with-cpp) option issues a
    warning, not an error.

No argument
    #if, #elif, #ifdef, #ifndef, #assert or #line has no arguments.

No header name
    A #include line does not have an argument, or expansion of a macro
    argument of a #include line results in no token.

Not a header name "UNDEFINED_MACRO"
    The specified argument is not a header name.  This message is issued
    when a macro that should define a header name is not defined.  A
    header name must be enclosed with < and >, or ", ".

Not an identifier "123"
    #ifdef, #ifndef, #define or #undef requires an identifier as an
    argument, but 123 is not an identifier.

No identifier
    #define or #undef does not have an argument.

No line number
    #line has a macro argument, but its expansion has resulted in no
    token.

Not a line number "name"
    The first argument of a #line is not a numeric token (preprocessing
    number).

Line number "0x123" isn't a decimal digits sequence
    The first argument of a #line must be a decimal integer.  MODE >=
    STANDARD MCPP issues this message.  In other modes, hex and octal
    integer tokens are allowed although a warning is issued.

Line number "2147483648" is out of range of [1,2147483647]
    The first argument of a #line must be within the range of 1 to
    2147483647.  0 is regarded as an error.  This is applied to MODE >=
    STANDARD.  With __STDC_VERSION__ < 199901L or __cplusplus < 199901L,
    the valid range will be 1 to 32767, but the range between 32768 and
    2147483647 is not regarded as error and a warning is issued.

Not a file name "name"
    The second argument of a #line, if any, must be a string literal.
    An identifier or wide string literal is not allowed here.

The following errors occur only in MODE >= STANDARD and these directives
are ignored.  MCPP compiled with MODE == PRE_STANDARD and OK_IF_JUNK ==
TRUE issues neither an error nor a warning.  Other MCPP issues a warning
and continues preprocessing as if there had been no "junk" text.

Excessive token sequence "junk"
    #else, #endif, #asm, or #endasm line has a junk text, or such text
    follows a valid argument of #ifdef, #ifndef, #include, #line or #
    undef line.

5.4.5   Syntax Related Errors in #if Expressions

This section covers syntax-related errors in #if, #elif and #assert
directives.  If a #if (#elif) line has these errors, MCPP evaluates it
to false, skips the #if (#elif) group, and continues processing.

For a skipped #if (#ifdef, #ifndef, #elif or #else) group, MCPP checks
validity of C preprocessing tokens and balance of these directives, but
not other grammatical errors.

A #if line has a sub-expression whose evaluation is skipped.  For
example, in case of #if a || b, if "a" is evaluated to true, "b" is not
evaluated at all.  However, the following 13 types of syntax errors or
translation limit errors are checked, even if they are located in a sub-
expression whose evaluation is skipped.

More than NEXP*2-1 constants stacked at "12"
    The number of constants in the stack has exceeded NEXP*2-1 when MCPP
    tried to evaluate "12" in a #if expression.  The depth of nested #if
    expressions is too deep.

More than NEXP*3-1 operators and parens stacked at "+"
    The total number of operators and parenthesis in the stack has
    exceeded NEXP*3-1 when MCPP tried to evaluate + in a #if expression.
    (A pair of parentheses is counted as two.)  The depth of nested #if
    expressions is too deep.

Misplaced constant "12"
    A #if expression has a constant 12 where no constant should be found.
    This error occurs when casting, such as (int)0x8000, is used in a #
    if expression, where casting is not allowed.  In this case, (int)
    0x8000 is evaluated to (0)0x8000, causing this error.  The int is
    regarded as an identifier that is not defined as macro and is
    evaluated to 0.

Operator ">" in incorrect context
    A #if expression has a > operator where no > should be found.  If a
    macro MACRO is defined as 0 token, #if MACRO > 0 will be expanded to
    #if > 0, causing this error, which is indicated by the preceding
    warning -- Macro "MACRO" is expanded to 0 token.

Unterminated expression
    A #if expression is not terminated.  This error is caused by, for
    example, #if a || MACRO with MACRO defined as 0 token.

Excessive ")"
    A #if expression has a ")" that does not corresponds to "(".

Missing ")"
    A #if expression does not have a ")" that corresponds to "(".

Misplaced ":", previous operator is "+"
    : without a corresponding ?.

Bad defined syntax
    A #if defined has a syntax error.  This error is caused by an
    unbalanced parenthesis or missing identifier in an argument.  When a
    macro expansion causes this error, the MCPP compiled with DEBUG ==
    TRUE displays this message followed by an expansion result.

Can't use a string literal "string"
    A string literal is not allowed as a constant in a #if expression.

Can't use a character constant 'a'
    In POST_STANDARD, a character constant, or a wide character constant
    is not allowed as a constant in a #if expression.

Can't use the operator "++"
    A #if expression has an illegal operator, such as = or ++.

Not an integer "1.23"
    Only integers, including character constants, are allowed as a
    constant in a #if expression.

Can't use the character 0x24
    A #if expression contains an illegal character (code 0x24), which is
    not any of the preprocessing tokens: identifiers, operators,
    punctuators, string literals, character constants, and preprocessing
    numbers. (Control codes are excluded since they had been checked
    before.)  To avoid this error, MCPP ported compiler systems that
    allows $ as an identifier must be compiled with OK_DOLLAR == TRUE.
    Of course, this is not checked in a skipped group.

The following error messages are relevant to #if sizeof.   The only MCPP
compiled with OK_SIZE == TRUE issues this error.

sizeof: Syntax error
    A #if sizeof has a syntax error.  This error is caused by an
    unbalanced parenthesis or missing arguments.

sizeof: No type specified
    Like sizeof(*), the "type" of #if sizeof (type) is not specified.
    Note that sizeof ((*)()) is a valid syntax to determine the size of
    a pointer to a function.

5.4.6   #if Expression Evaluation Errors

The following errors do not occur in a sub-expression whose evaluation
is skipped. (MCPP invoked with the -W8 option issues a warning.)

Constant "123456789012" is out of range
    An integer constant has a value that exceeded the range of unsigned
    long. (For compiler systems without unsigned long, this range will
    be long, and so forth.)  With __STDC_VERSION__ or __cplusplus set to
    199901L or higher, this range will be unsigned long long in compiler
    systems with long long, and so forth.

Integer character constant 'abcde' is out of range
    A character constant 'abcde' has a value that exceeded the range of
    unsigned long (or unsigned long long).

Wide character constant L'abc' is out of range
    A wide character constant L'abc' has a value that exceeded the range
    of unsigned long.  Two UTF-8 encoded characters exceeds the range of
    long because one character is evaluated to four bytes.  When the -
    V199901L option is used to indicate C99 conformance, up to two
    characters fall within the range of long long.  This error occurs
    only in MODE == STANDARD.

CHARBIT bits can't represent escape sequence 'x123'
    An escape sequence in a character constant has exceeded the range of
    CHARBIT bits ([0, UCHARMAX]).

CHARBIT*2 bits can't represent escape sequence L'x12345'
    An escape sequence in a wide character constant has exceeded the
    range of CHARBIT*2 bits (CHARBIT*4 bits for UTF-8).  This error
    occurs only in MODE == STANDARD.

Division by zero
    A #if expression contains a division by zero.  A division can be
    expressed using / or %.  This error may be caused by a #if dividend/
    divisor with the divisor not defined as a macro.  To avoid this
    error, "#if defined divisor && (dividend/divisor ..)" is recommended.

Result of "op" is out of range
    An operation result using the operator op is out of range of long.
    Op is any of binary operators: *, /, %, +, and -.  When two's
    complement representation is used, the unary operator - will cause
    an overflow with -LONG_MIN (For C99, with -LLONG_MIN).  Unsigned
    long (or unsigned long long) will never cause an overflow, so it
    does not cause this error.  If the result of an algebraic operation
    is out of range, a warning is issued.

The following errors are relevant to sizeof.  They are not issued in a
sub-expression whose evaluation is skipped (The -W4 option issues a
warning).  MCPP compiled with OK_SIZE == TRUE issues these messages.

sizeof: Unknown type "type"
    The "type" of #if sizeof (type) is unknown.

sizeof: Illegal type combination with "type"
    A type combination, like #if sizeof (long float), is invalid.  What
    type combinations are valid, such as long long, depends on MCPP
    settings.

5.4.7   #define Related Errors

This section covers #define related errors.  A macro will not be defined
if an error occurs at #define.  The # and ## operator related errors
occurs in MODE >= STANDARD.  __VA_ARGS__ related errors also occur in
MODE >= STANDARD.  Although variable argument macro is a C99
specification, MCPP allows these macros to be used in C90 and C++ modes
for compatibility with GNU C.  (A warning is issued.)

"defined" shouldn't be defined
    A macro name "defined" cannot be defined.  MODE >= STANDARD MCPP
    checks this.

"__STDC__" shouldn't be redefined
    The __STDC__ macro cannot be #defined.  The same can be said with
    __STDC_VERSION__, __FILE__, __LINE__, __DATE__ and __TIME__
    (__STDC_HOSTED__ in C99 mode, and __cplusplus when MCPP is invoked
    with -+ option).  MODE >= STANDARD MCPP checks these macros.

"__VA_ARGS__" shouldn't be defined
    C99 allows a variable argument macro with the __VA_ARGS__ parameter
    in the replacement list, but this identifier cannot be defined as a
    macro.  MODE >= STANDARD MCPP checks this.

More than NMACPARS parameters
    The number of parameters of a macro definition has exceeded NMACPARS.

Empty parameter
    A macro definition has an empty parameter.

Illegal parameter "123"
    A token other than an identifier is used in a parameter of a macro
    definition.  In MODE >= STANDARD, even an identifier __VA_ARGS__
    cannot be used.

Duplicate parameter name "a"
    A macro definition has a duplicate parameter name "a".

Unterminated parameter list "(a,b"
    A macro definition does not have a parenthesis ")" that closes a
    parameter list.

No token before ##
    No token precedes the ## operator in the replacement list of a macro
    definition.

No token after ##
    No token follows the ## operator in the replacement list of a macro
    definition.

## after ##
    The replacement list of a macro definition has a token sequence of
    "## ##".  Some may do not regard this as error, but since
    concatenation of ## with other token always generates an invalid
    token, when this happens in macro expansion, it always causes an
    error.  MCPP makes it an error when it finds this in a macro
    definition.

Not a formal parameter "id"
    A function-like macro definition has a # operator whose operand id
    is not a parameter name.

"..." isn't the last parameter
    "..." must be the last parameter of a macro definition.  In
    PRE_STANDARD mode, "..." causes an illegal parameter error.

    "__VA_ARGS__" without corresponding "..." "__VA_ARGS__", an
    identifier in a replacement list, can be used only when it has a
    corresponding "..." parameter.

5.4.8   #undef Related Errors

This section covers #undef related errors.

"__STDC__" shouldn't be undefined
    The __STDC__ macro cannot be #undefined. The same can be said with
    __STDC_VERSION__, __FILE__, __LINE__, __DATE__ and __TIME__
    (__STDC_HOSTED__ in C99 mode, and __cplusplus when MCPP is invoked
    with -+ option).  MODE >= STANDARD MCPP checks these macros.

5.4.9   Macro Expansion Errors

This section covers macro expansion errors.

The MCPP compiled with DEBUG == TRUE displays a macro definition, as
well as the source filename and line number where it is found.

Less than necessary N argument (s) in macro call "macro( a)"
    A macro invocation has an insufficient number of arguments.  This
    macro requires N number of arguments.  MCPP assigns a zero token to
    missing arguments and continues to process.  MCPP does not regard a
    macro that takes only one parameter with zero number of arguments
    specified as error because it cannot distinguish an empty argument
    from a missing argument.

More than necessary N argument (s) in macro call "macro( a, b, c)"
    A macro invocation has too many arguments.  The macro should take N
    number of arguments.  MCPP ignores surplus arguments and continues
    processing.

Not a valid preprocessing token "+12"
    The ## operator has concatenated two pp-tokens, resulting in an
    invalid token "+12".  The token may be separated at a later time.
    MCPP continues processing.  MCPP invoked with the -lang-asm (-x
    assembler-with-cpp, -a) option issues a warning.

Not a valid string literal "\\"str\""
    When a # operator tried to convert macro invocation's argument into
    a string, a token sequence of "\\"str"" has resulted, instead of a
    single valid string literal.  \ that precedes or follows the literal
    cause the error. (When MCPP tries to convert such an argument into a
    string, it may or may not cause an unterminated string literal error.)
    MCPP tries to continue processing but maybe an error occurs again in
    the compilation phase.  This error may not occur in MODE ==
    POST_STANDARD. (An unterminated string literal error may occur).

When the following errors occur, the macro invocation will be skipped.

Buffer overflow expanding macro "macro" at "something"
    A buffer overflow has occurred at "something" during macro expansion.
    Divide the macro.

Unterminated macro call "macro( a, (b,c)"
    A macro invocation is not terminated.  This error usually occurs
    when a macro invocation on the control line is not terminated at
    that line.  In MODE >= STANDARD, a macro in an argument is expanded
    before argument substitution, in which case, the macro invocation
    must be terminated in the argument.  In MODE == POST_STANDARD, a
    macro invocation unterminated in a replacement list also causes this
    error.

Rescanning macro "macro" more than RESCAN_LIMIT times at "something"
    The depth of nested macros is so deep that the number of rescans has
    exceeded RESCAN_LIMIT at "something" during expansion.  This error
    occurs only in MODE >= STANDARD but it is quite rare.

Recursive macro definition of "macro" to "macro"
    A macro definition is recursive.  This error occurs only in MODE ==
    PRE_STANDARD.  When the number of rescans has exceeded RESCAN_LIMIT,
    MCPP regards it as a recursive macro definition.

5.4.10  #error and #assert

#error
    A #error directive has been executed.  Following this message, the #
    error line is displayed.  If an argument itself contains a token
    error, such as unterminated strings, #error is not executed.  The
    only MODE >= STANDARD MCPP has #error.

Preprocessing assertion failed:
    A #assert directive has been executed.  Following this message, #
    assert line arguments are displayed.  If any of the arguments
    contains an error, MCPP regards assertion has failed.  The only MODE
    == PRE_STANDARD MCPP has #assert.

5.4.11  Failure of #include

Can't open include file "file-name"
    This error occurs when a file to include does not exist.  Probably,
    this is due to wrong spelling of the file name or an
    "include directory" should have been specified.

5.4.12  Other Errors

Operand of _Pragma() is not a string literal
    The _Pragma() operator must take an argument of one string literal
    or wide string literal.  This is checked when MCPP compiled with
    MODE >= STANDARD and OK_PRAMGA_OP == TRUE is invoked with the -S1 -
    V199901L option.  The same thing can be said when MCPP is invoked
    with the -V199901L option in C++ mode.


5.5     Warnings (Class 1)

A warning is issued when code, although syntactically correct, possibly
contains some coding mistakes or has a portability problem.  Warnings
are divided into five classes: 1, 2, 4, 8, and 16.  These classes are
enabled when the -W <n> option is specified on MCPP invocation.  <n>
specifies a ORed value of any of 1, 2, 4, 8, and 16.  Class 4, for
example, can be specified explicitly with -W<4>, and implicitly with -W
<n>, where <n> is 5, 6, 7, 12, 13, 14, 15, 20, 21, 22, 23, 28, 29, 30,
or 31 because the ANDed value of <n> and 4 is 4 (true).

MODE == STANDARD MCPP issues an error message to most of the source code
that causes a Standard C undefined behavior, but a warning to some code.

Likewise, MODE == STANDARD MCPP always issues a warning to the source
code based on Standard C unspecified specifications, except for the
following:

1.  Evaluation order of sub-expressions in a #if expression - Although
the evaluation order of the operands of operators other than ||, &&, ? ,
and : is unspecified in Standard C, MCPP does not issue a warning
because #if expression does not cause side-effects and therefore the
evaluation order does not affect results.  MCPP always evaluates integer
constant tokens from left to right in the order they appear and performs
an operation using the tokens in accordance with an operator grouping
rule when their values are needed.

MODE == STANDARD MCPP issues a warning to many implementation-defined
behaviors, except for the following:

1.Directories the #include directive searches for a file to include and
how to construct a header-name pp-token from #include's argument. - MCPP
does not issue a warning because there will be too many warnings if it
actually does.  Unless a header name is a macro, the source token
sequence, including spaces, are used as it is.  If it is a macro, the
expanded result, including spaces, is used.  (In MODE == POST_STANDARD,
a space character is inserted between pp-tokens during macro expansion.
A header-name is constructed by concatenating the resulting pp-tokens
from < to > and then by removing space characters.  In POST_STANDARD, a
header-name enclosed with < and > is obsolete.)  When MCPP encounters #
pragma __debug_cpp __path or #debug path, it displays a search path,
instead of issuing a warning.

2. Evaluation of a single byte character constant, such as 'a', and of a
wide character constant that consists of only one multi-byte character,
such as L'', in a #if expression. - MCPP does not issue a warning
because even with the same basic character set used, there are an
unlimited number of factors that limits the portability, such as single
byte Katakana, presence or absence of a sign, encoding scheme of Kanji,
and etc.  The same thing can be said with UCN.

3. Bit operations using negative numbers in a #if expression - Although
bit operation results depend on internal representation of an integer on
a machine, most of the machines use two's complement representation,
thus causing no portability problem.  However, MCPP issues a warning to
a right bit shift operation of a negative value and a division operation
involving either or both of negative operands because they lack of
portability.

4. A sequence of several white space characters as a token separator -
Standard C is clear that it is implementation-defined whether a sequence
of white space characters is compressed into one space character in the
translation phase 3, but you do not have to worry about this.
Portability becomes an issue only when a preprocessing directive line
has <vertical-tab> or <form-feed>.  MCPP converts it into one space
character and issues a warning.  For a sequence of several space
characters and tabs, MCPP compresses it into one space character without
a warning.

5. Compiler system's own built-in macros will not cause warning.

6. #pragma sub-directive - Principally, MCPP does not issue a warning to
#pragma sub-directive, however, MCPP ported compiler systems that do not
recognizes #pragma issues a warning to them, except for #pragma __once,
#pragma __put_defines, #pragma __preprocess, #pragma __preprocessed, #
pragma __debug_cpp, and #pragma __end_debug_cpp, which MCPP itself
processes.  MCPP issues a warning to these #pragma sub-directives if
they have an invalid argument.  In addition, MCPP issues a warning to
GNU C V.3's #pragmas, such as #pragma GCC poison (dependency,
system_header), that compiler-specific preprocessors process but MCPP
does not.

Doubled \ - Although it is implementation-defined in C99 whether a
single \ is changed into double \ (\\) when the # operator converts a
UCN sequence into a string, MCPP does not issue a warning to this.  MCPP
does not double \.

As you see, MCPP can perform almost all the portability checks necessary
at a preprocessing level.

MODE == POST_STANDARD is identical with MODE == STANDARD except for some
specification differences described in porting.txt 4.1.3.

Regardless of the number of warnings, MCPP always returns a status of
success.  MCPP invoked with the -W0 option does not issue a warning.

5.5.1   Character, Token and Comment Related Warnings

Illegal control character 0x1b in quotation
    A string literal, character constant or header name has a control
    code other than a white space character, which may cause an error in
    the compiler-proper.  This way of coding is not desirable.  A
    control code in string literals or character constants should be
    written using an escape sequence.

Illegal multi-byte character sequence "XY" in quotation
    The first byte (X) of "XY" in a string literal, character constant,
    or header name is the first byte of a multi-byte character (Kanji),
    while the second byte (Y) is not the second byte of the character.
    "XY" may not be displayed correctly.  MCPP does not regard "XY" as a
    single multi-byte character.  It treats the first byte as a single-
    byte character and the second byte as the next character.

    MCPP does not issue a warning to a character in an external
    character set, as long as it is within the proper range.  Even
    within the proper range, there are some holes (no corresponding
    characters).  MCPP does not check whether such a character is
    defined or not.  The following table shows the range of each multi-
    byte character set:

        Encoding        first byte              second byte
        shift-JIS       0x81-0x9f, 0xe0-0xfc    0x40-0x7e, 0x80-0xfc
        EUC-JP          0x8e, 0xa1-0xfe         0xa1-0xfe
        KS C 5601       0xa1-0xfe               0xa1-0xfe
        GB 2312-80      0xa1-0xfe               0xa1-0xfe
        Big Five        0xa1-0xfe               0x40-0x7e, 0xa1-0xfe
        ISO-2022-JP     0x21-0x7e               0x21-0x7e

    Beside chracter codes, ISO-2022-JP has a shift sequence.  Apart from
    the shift sequence, all the multi-byte characters other than UTF-8
    are two bytes.

    In UTF-8, multi-byte characters are two bytes or three bytes.  Kanji
    is encoded in three bytes.  The first byte is within the range of
    0xc2 to 0xef, second and third 0x80 to 0xbf.  Details are omitted
    here.  Anyway, all these bytes must fall within the appropriate
    ranges.

    Note that since MCPP is unable to recognize EUC-JP's three byte
    encoding (JIS X 0213), it regards 0x8f + 0xa1-0xfe + 0xa1-0xfe not
    as one character but as two characters of 0x8f and 0xa1-0xfe + 0xa1-
    0xfe.  As a result, MCPP does not issue a warning to the three byte
    encoding and can evaluate it correctly, except for a wide character
    constant in a #if expression.  In EUC-JP, a character with the first
    byte of 0x8e (a half-width Katakana) is encoded in two bytes, and
    treated as a multi-byte character.  This warning is not issued in a
    skipped #if group.

"/*" in comment
    A comment has a sequence of /*.  Unless it is intended, the
    programmer may have forgot to enclose the comment.  A comment cannot
    be nested.

Parsed "//" as comment
    A text from // to the end of the line is interpreted as a comment.
    This is a legal notation in C99 and C++.  In C90 mode MCPP treats it
    as a comment after issuing a warning.  This is applied to MODE >=
    STANDARD.

Too long identifier, truncated to "very_long_identifier"
    Since the length of an identifier has exceeded IDMAX, it is
    truncated to IDMAX.

Illegal digit in octal number "089"
    An octal numeric token contains 8 or 9.  The only MODE ==
    PRE_STANDARD MCPP issues this warning.  MODE >= STANDARD MCPP does
    not check whether a numerical token on lines other than #if
    directives is correct or not.  If a #if expression has an octal
    numeric token of 8 or 9, it will cause a "Not an integer" error.

Unterminated string literal, catenated to the next line
    Although an unterminated string literal in a logical line is
    normally regarded as an error, MCPP invoked with the -lang-asm (-x
    assembler-with-cpp, -a) option regards it as a multi-line string
    literal and concatenates the line with the next by inserting '\n'.
    This way of writing has no advantage.  Using a function to
    concatenate adjacent string literals is preferable.

5.5.2   Unterminated Source File Related Warnings

The following warning messages are issued to included files by MCPP
compiled with OLD_PREPROCESSOR == TRUE and MODE == PRE_STANDARD.  MCPP
compiled with MODE >= STANDARD issues an error.  MCPP of MODE ==
PRE_STANDARD and OLD_PREPROCESSOR == FALSE issues an error when it finds
the end of input.

End of file within #if (#ifdef) section starting from line 123
End of file with unterminated #asm block starting from line 123
End of file within macro invocation starting from line 123

5.5.3   Directive Line Related Warnings

The macro is redefined
    The MCPP compiled with DEBUG == TRUE displays this message followed
    by the source filename and line number where this macro definition
    is found.  The macro has been redefined with a different contents.
    Source must not be well organized.  The following conditions must be
    met for macro definitions with the same name to exist. Or, a warning
    is issued.

    1.  Have the same number of parameters.
    2.  Have the same replacement list (one or more white space
    character between tokens are regarded as one.  In POST_STANDARD, the
    difference of the token separators does not matter because any
    number of space characters is changed into one, regardless of the
    presence or absence of the token separators.)
    3. In MODE == STANDARD, parameter names must be the same.  Other
    MODE, including POST_STANDARD, they are not checked.

The following messages are issued only by MODE == STANDARD MCPP.

No space between macro name "MACRO" and repl-text
    There is no space between macro name and replacement list of a #
    define line.  Normally, this does not happen, but it does happens
    when an illegal character is used in a macro name as follows:

        #define THIS$AND$THAT(a, b)   ((a) + (b))

    MCPP interprets this as follows:

        #define THIS  $AND$THAT(a, b) ((a) + (b))

    and issues a warning.  Of course, this is a quite rare case.

The following four warnings are issued only by MODE >= STANDARD MCPP.

No sub-directive
    A #pragma line does not have any argument.  The line is ignored.

Unknown argument "name"
    There is no such directive as #pragma name.  The only MCPP compiled
    with PRAGMA == FALSE issues this warning.

Unknown encoding "encoding"
    The encoding name, encodingh, specified with #pragma __setlocale
    ( "encoding") is not implemented.

Too long encoding name "encoding"
    The encoding name, "long-long-encoding", specified with #pragma
    __setlocale( "long-long-encoding") exceeds 19 bytes.  MCPP ignores
    it.

Bad push_macro syntax
Bad pop_macro syntax
    There is a syntax error in #pragma __push_macro, #pragma __pop_macro,
    #pragma push_macro or #pragma pop_macro.  To use these #pragma
    directives, first enclose a macro name in an argument with ", " and
    then further enclose with ( ).  For example, ("MACRO").

"MACRO" has not been defined
    MACRO in the argument,("MACRO"), for #pragma __push_macro, #pragma
    __pop_macro, #pragma push_macro, or #pragma pop_macro is not defined
    as a macro.

"MACRO" is already pushed
    MACRO of #pragma __push_macro ("MACRO") has been pushed and then
    further #undefed.  Without redefining the MACRO, push would not be
    possible.

"MACRO" has not been pushed
    MACRO in #pragma __pop_macro( "MACRO") has not been pushed.  It may
    have been already poped.

The GNU C ported MCPP issues the following warnings:

Ignored #ident
Ignored #sccs
    #ident or #sccs lines are ignored.

For the following three #pragma related warnings, MCPP compiled with
PRAGMA == TRUE outputs the line in question as it is.

Unknown argument "name"
    There is no such an argument of #pragma __debug_cpp or #debug as
    "name".

No argument
    A #pragma __debug_cpp or #debug does not have an argument.

Not an identifier "123"
    The argument of a #pragma __debug_cpp or #debug is not an identifier.

The above three warnings are issued only by MCPP compiled with DEBUG ==
TRUE and/or DEBUG_EVAL == TRUE.

GNU C ported MCPP issues a Class 2 warning to a line with #pragma GCC
followed by either poison, dependency or system_header and does not
output the line.  GNU C V.3 specific preprocessor process the line but
MCPP does not.

The following warnings are issued only by MODE == PRE_STANDARD MCPP.
MODE >= STANDARD MCPP regards them as errors.

Not in a #if (#ifdef) section in a source file
    Line number "0x123" isn't a decimal digit sequence.

The following warning is issued only by MCPP compiled with OK_IF_JUNK ==
FALSE.  MODE >= STANDARD MCPP issues the same warning to #pragma __once,
#pragma __put_defines, #pragma __push_macro, #pragma __pop_macro, #
pragma push_macro, #pragma pop_macro, #pragma __debug_cpp, and #pragma
__end_debug_cpp; for other directives, MCPP issues an error.  MCPP
compiled with OK_IF_JUNK == TRUE issues neither an error nor a warning.

"Excessive token sequence "junk"

5.5.4   #if Expression Related Warnings

The following three warnings are relevant to an argument of #if, #elif,
or #assert:

Macro "MACRO" is expanded to "defined"
    The MACRO macro in a #if expression has been expanded to "defined".
    MCPP treats this strange macro not as identifier but as operator.
    How to treat it is undefined in Standard C.

Macro "MACRO" is expanded to "sizeof"
    The MACRO macro in a #if expression has been expanded to sizeof.
    MCPP treats this strange macro not as identifier but as operator.
    MCPP compiled with OK_SIZE == TRUE issues this warning.

Macro "MACRO" is expanded to 0 token
    The MACRO macro has been expanded to zero token.  If this happens in
    a #if expression, it almost always causes an error.  The purpose of
    this warning is to indicate the cause of an error.

The followings warnings are relevant to an argument of #if, #elif or #
assert.  They are not issued in a sub-expression whose evaluation is
skipped. (MCPP invoked with the -W8 option issues them.)

Undefined escape sequence '\x'
    There is no such escape sequence as \x.  \x is evaluated to a two
    byte sequence. (Of course, an escape sequence of "\x" followed by a
    hex string is valid)  This warning is also issued to a UCN with an
    insufficient number of orders.

The followings warnings are relevant to operations in a constant
expression on #if, #elif or #assert lines.  No warnings are also issued
in a skipped sub-expression. (MCPP invoked with -W8 issues them.)  How
MCPP performs an operation in a constant expression depends on the
specifications of the compiler-proper with which MCPP was compiled.

Negative
    A mixture of signed and unsigned operations results in conversion of
    a signed negative value into an unsigned positive value.  This is
    not an error, but indicates source code may contain a bug.  For both
    operands of a binary operator, such as *, /, %, +, -, <, >, <=, >=,
    ==, !=, &, ^ and | , and the second and third operands of a ternary
    operator, such as ? and :, if one operand is unsigned and the other
    is signed, the signed one is always converted into unsigned.

Illegal shift count "-1"
    The value of the right operand of a bit shift operator, << or >>, is
    a negative number or has exceeded the bit width of long.  Probably,
    this is also a bug in source code.

"op" of negative number isn't portable
    If an operation using a binary operator (op) results in either or
    both of negative operands, it lacks of portability. "Op" is any of /,
    %, and >>.  The >> operator with a negative left operand provides
    portability across compiler systems on computers having an
    arithmetic shift command, where a one-bit shift means a division by
    2.  Otherwise, it does not provide portability.

Result of "op" is out of range
    An operation result using "op" is out of range of unsigned long.
    "Op" is any of binary operators (*, /, %, +, or -), or a unary
    operator (-).  Unsigned long will never cause an overflow, so it
    never causes an error.  If the result of an algebraic operation is
    out of range, MCPP issues a warning.

5.5.5   Macro Expansion Related Warnings

The MCPP compiled with DEBUG == TRUE displays a macro definition
followed by the source filename and line number where the macro
definition is found.

Macro started at line 123 swallowed directive-like line
    MCPP has read a line that begins with # as an argument of the macro
    that begins at the line 123.  Maybe, the macro invocation has a bug.
    If it had not been for the macro, the line that begins with # would
    have been interpreted as a directive line.  The same thing could be
    said if the macro had been located in a #if group whose evaluation
    is skipped, because such macro is never expanded.

Replacement text "sub(" of macro "head" involved subsequent text
    Rescanning of the replacement list "sub(" of the macro "head" has
    involved the text succeeding the macro invocation.  K&R 1st. to
    Standard C did not regard this as an error, but if you used this
    type of macro without having these standards in mind to receive this
    warning, your macro definition or macro invocation is not correct.
    If you are intended to use such macro, it is not desirable at all.
    The only MODE == STANDARD MCPP issues this warning.  MCPP does not
    issues this warning, however, if invoked with -c option.  In MODE ==
    PRE_STANDARD, the same situation may arise but no warning is issued.
    MODE == POST_STANDARD MCPP never issues this warning because
    rescanning does not involve the text succeeding the replacement list.
    (A macro may be expanded quite differently or causes an
    "unterminated macro call" error.)

Less than necessary N argument (s) in macro call "macro( a)"
    An insufficient number of arguments of a macro invocation.  Normally,
    this causes an error, but in case of missing only one argument of a
    macro that takes a variable number of arguments, MCPP issues a
    warning.  This is to decrease migration problems of variable
    argument macros between GNU C and C99.

5.5.6   Line Number Related Warnings

This section covers line number related warnings.

Line number "32768" is out of range of [1,32767]
    In C90 and C++, the first argument of a #line must be within the
    range of 1 to 32767.  0 is also out of range.  With __STDC_VERSION__
    >= 199901L or __cplusplus >= 199901L, the valid range is 1 to
    2147483647.  Therefore, in C90 or C++ mode, MCPP issues a warning,
    not an error, to the range of 32768 to 2147483647.  MODE >= STANDARD
    MCPP issues this warning.

In C90, when you use #line to specify a value slightly below 32767, you
won't receive an error, but sooner or later, the line number will exceed
32767, in which case, MCPP continues to increase the line number while
issuing a warning.  Some compiler-proper may not accept this large line
number.  It is not desirable to specify a large number with #line.

Line number 32768 got beyond range
    The source line number has reached 32768, at which a warning is
    issued one time.

Line number 32769 is out of range
    When the __LINE__ macro is expanded, the lime number has exceeded
    32767.

5.5.7   #pragma __warning, #warning

#warning
    A #pragma __warning (#warning) directive has been executed.
    Following the above message, the #pragma __warning line is displayed.
    (If an argument of #pragma __warning has a token error, such as
    unterminated string, #pragma __warning is not executed.)  Although
    this directive appears in the Warning Level 1 section, this warning
    is issued at every warning level.  MODE >= STANDARD MCPP has #pragma
    __warning, while MODE == PRE_STANDARD MCPP has #warning.


5.6     Warnings (Class 2)

This section covers warnings to code that does not contains a bug but
causes a portability problem.

Only the MCPP compiled with MODE >= STANDARD issues the following four
warnings:

Parsed "//" as comment
    A text from // to the end of the line is interpreted as a comment.
    This is a legal notation in C99 and C++.  In C90 mode MCPP treats it
    as a comment after issuing a warning.

Variable argument macro is defined
    Although it is the C99 Standard that stipulates variable argument
    macros, a variable argument macro has been defined in C90, C95 or C+
    + mode.

Empty argument in macro call "MACRO( a, ,"
    A macro invocation has an empty argument, in which case, MCPP
    regards the argument as zero number of pp-token sequences and treats
    it as such.  The empty argument is legal in C99, while it is
    undefined in C90, thus causing a lack of portability.  (MCPP regards
    an empty argument even without a ',' not as an empty argument, but
    as a missing argument, thus issuing an error.  Since zero number of
    arguments and one empty argument is syntactically indistinguishable,
    MCPP does not make both an error.)  Writing an empty argument in
    source code is not generally preferable.  I recommend that you
    should code:
        #define EMPTY
    , if possible, and then write EMPTY where an empty argument is
    written.

Skipped the #pragma line
    GNU C V.3 provides several #pragma directives in the form of #pragma
    GCC <args>.  Its preprocessor processes some of them, but MCPP does
    not support them.  This warning is issued to a #pragma directive
    compiler-specific preprocessors process but MCCP does not.

Not a valid preprocessing token "+12"
    Concatenating two pp-tokens with the ## operator results in an
    invalid token "+12", which normally causes an error.  However, MCPP
    invoked with the -lang-asm  (-x    assembler-with-cpp, -a) option
    does not regard it as an error.

The following warning is issued only by MODE == POST_STANDARD MCPP:

Header-name enclosed by <, > is an obsolescent feature <stdio.h>
    The header name in the form of <stdio.h> is one of the
    specifications I want to abolish.  I recommend to use "stdio.h".

    The following two warnings are issued only in some compiler systems.
    Of course, the coding in question is valid in those particular
    systems, but it lacks of portability, so a warning is issued to
    remind users of it.

#include_next is not allowed by Standard
#warning is not allowed by Standard
    These directives are valid in GNU C but not Standard C-conforming
    and lack of portability.

Converted  to /
    A #include directive contains \ in the header name.  MCPP converts \
    into /.  "" is a valid path delimiter in OSs, such as DOS and
    Windows, but undefined in Standard C.  It is safe to use /.  The
    only MCPP ported to DOS/Windows issues this warning only once. (MCPP
    does not regard
    " preceded by \\ as a delimiter of a string literal, raising an "
    unterminated string literal" error.)

'$' in identifier "THIS$AND$THAT"
    An identifier has a '$'.  The only MCPP compiled with DOLLAR_IN_NAME
    set to TRUE issues this warning only once because '$' lacks of
    portability although it is valid in this MCPP.  '$' being regarded
    as a pp-token, other MCPP parses THIS$AND$THAT into five components
    THIS,  $,  AND, $ and THAT, resulting in a compiler error.


5.7     Warnings (Class 4)

Standard C guarantees some minimum translation limits.  It is desirable
that a preprocessor imposes translation limits that exceed these values,
but source programs that uses cpp' own translation limits will restrict
portability.  MCPP provides some macros in "system.H" that allows you to
set translation limits to any values you like.  MODE >= STANDARD MCPP
issues a warning to source code that exceeds a Standard C guaranteed
limit.  However, these messages are excluded from Class 1 and 2 because
they may be issued frequently, depending on standard headers of compiler
systems or source programs.

Logical source line longer than 509 bytes
    The length of a logical source line has exceeded 509 bytes.

Quotation longer than 509 bytes "very_very_long_string"
    The length of a string literal, character constant or header name
    has exceeded 509 bytes.

More than 8 nesting of #include
    The depth of nested #includes has exceeded 8.  This warning is
    issued only when it reaches 9.

More than 8 nesting of #if (#ifdef) sections
    The depth of nested #ifs, #ifdefs, or #ifndefs has exceeded 8.  This
    warning is issued only when it reaches 9.

More than 1024 macros defined
    The number of defined macros has reached 1024.  This number includes
    both of pre-defined macros and those defined in header files.

String literal longer than 509 bytes "very_very_long_string"
    Expansion of a macro using the # operator has generated a string
    literal longer than 509 bytes.

The following warnings are not issued in a skipped #if group.

More than 32 nesting of parens in #if expression
    The depth of nested parentheses in a #if expression has exceeded 32.
    This warning is issued only when it reaches 33.

More than 31 parameters
    The number of parameters of a macro definition has exceeded 31.

Identifier longer than 31 bytes "very_very_long_name"
    The length of an identifier has exceeded 31 bytes.

With __STDC_VERSION__ >= 199901L, the Standard specified translation
limits are as follows:

    Length of logical source line:                  4095 bytes
    Length of string literal, character constant, or header name:
                                                    4095 bytes
    Identifier length:                              63 characters
    Depth of nested #includes:                      15
    Depth of nested #ifs, #ifdefs, or #ifndefs:     63
    Depth of nested parentheses in #if expression:  63
    Number of macro parameters:                     127
    Number of definable macros:                     4095

Note that the length of a UCN or multi-byte-character as an identifier
is expressed as the number of characters, not bytes.

When MCPP is invoked with the -+ option to specify C++ preprocessing,
the Standard specified translation limits are as follows:

    Length of logical source line:                  65536 bytes
    Length of string literal, character constant, or header name:
                                                    65536 bytes
    Identifier length:                              1024 characters
    Depth of nested #includes:                      256
    Depth of nested #ifs, #ifdefs, or #ifndefs:     256
    Depth of nested parentheses in #if expression:  256
    Number of macro parameters:                     256
    Number of definable macros:                     65536

Note that MCPP allows the maximum number of macro parameters of 255.  So,
when it reaches 256, MCPP issues an error.

The following warnings are excluded from class 1 and 2 because they are
issued too frequently.

Converted 0x0c to a space
    [FF], [VT], [CR] (other than '\n'), and [LF] (other than '\n') in
    source code as token separators are converted into a space character.
    How to deal with these token separators located on a directive line
    is undefined in Standard C.  If they are located in comments, string
    literals, or character constants, MCPP does not convert them. (Of
    course, MCPP can do so, but I do not want MCPP to impose a greater
    restriction on a character set used since it essentially depends on
    the compiler-proper.)  On the other hand, [TAB] as a token separator
    is converted into a space character, but no warning is issued
    because it does not affect compilation at all. ([TAB] means nothing
    but a space to both of preprocessor and compiler-proper.)

Undefined symbol
    The identifier "name" does not have a macro value defined.  It is
    evaluated to zero.  This is not an error at all, but may be a
    program bug.  No warning is issued to an argument of a #if defined.
    This warning can be avoided by writing #if defined name && (name ..),
    instead of #if name .., or by invoking MCPP with the -D name=0
    option.  C++ gives "true" and "false" special treatment and
    evaluates to 1 and 0, respectively, without a warning.

Multi-character wide character constant L'ab' isn't portable
    A wide character constant value varies even among compiler systems
    using the same character set because the encoding scheme of wide
    character constants and how to evaluate multi-characters depend on
    compiler systems.  Therefore, #if expressions using them do not
    provide portability.  The only MODE ==STANDARD MCPP issues this
    warning.

Multi-character or multi-byte character constant '' isn't portable
    Since how to evaluate the value of a multi-character or multi-byte
    character constant depends on compiler systems,  #if expressions
    using them do not provide portability. The only MODE !=
    POST_STANDARD MCPP issues this warning.

The following two warnings are issued only by MODE >= STANDARD MCPP:

Macro with mixing of ## and # operators isn't portable
    A function-like macro has a token sequence of "## #" in the
    replacement list.  These two operators lack of portability because
    their priority is unspecified in Standard C.  MCPP takes precedence
    # over ##.  Note that if a function-like macro has a token sequence
    in the reverse order "# ##", MCPP regards it as an error because the
    operand of the # operator must be a parameter.

Macro with multiple ## operators isn't portable
    A macro definition has only one token or parameter inserted between
    ## operators in the replacement list.  This macro may lack of
    portability because the evaluation order of ## operators is
    unspecified in Standard C.  MCPP applies the ## operator from left
    to right.


5.8     Warnings (Class 8)

There is little chance that the indicated source code contains a bug,
but these messages are issued to call attention to it.  MCPP invoked
with the -W8 option issues these warnings.

In a skipped #if group, whether preprocessing directives, such as #ifdef,
#ifndef, #elif, #else, and #endif, are balanced or not is checked.
However, MCPP invoked with the -W8 option also checks illegal or unknown
directives.  MODE >= STANDARD MCPP issues a warning when the depth of
nested #ifs exceeds 8.

Illegal #directive "123" (in skipped block)
Unknown #directive "pseudo-directive" (in skipped block)
More than 8 nesting of #if (#ifdef) sections (in skipped block)
#include_next is not allowed by Standard
#warning is not allowed by Standard

The following warnings are related to #if expression.  Given an
expression of #if a || b, for example, if "a" is true, "b" is not
evaluated.  However, MCPP invoked with -W8 issues a warning to non-
evaluated sub-expressions, in which case, the note saying
"in non-evaluated sub-expression" is appended.  In case of an overflow,
a wrap-rounded value is used in subsequent operations.  In case of a
division by 0, the maximum integer is assumed as a division result and
is used in subsequent operations.

Constant "123456789012" is out of range
Integer character constant 'abcde' is out of range
Wide character constant L'abc' is out of range
CHARBIT bits can't represent escape sequence '\x123'
CHARBIT*2 bits can't represent escape sequence L'\x12345'
Division by zero
Undefined symbol "name", evaluated to 0
sizeof: Unknown type "type"
sizeof: Illegal type combination with "type"
Multi-character wide character constant L'ab' isn't portable
Multi-character or multi-byte character constant '' isn't portable
Undefined escape sequence '\x'
UCN cannot specify the value "0000007f"
Negative number "-1" is converted to positive "4294967295"
Result of "op" is out of range
Illegal shift count "-1"
"op" of negative number isn't portable

sizeof is disallowed in C Standard
    The purpose of this warning is to remind users of the fact that
    Standard C does not allow for #if sizeof although MCPP compiled with
    MODE == PRE_STANDARD and OK_SIZE == TRUE implements it.

"MACRO
    An undefined name is specified with #undef.  Standard C does not
    regard it as an error.

Macro "macro" needs arguments
    A token with the same name as a macro with arguments appears in a
    standalone manner.  MCPP does not expand it and leave it as it is.
    The only MODE == PRE_STANDARD MCPP issues this warning. (MODE >=
    STANDARD MCPP does not issue a warning since such a token does not
    cause any problem.)

String literals "str1" "str2" are concatenated
    Adjacent string literals "str1" and "str2" have been concatenated
    into "str1str2".  The purpose of this warning is to make sure that
    the programmer really intended that, instead of "str1", "str2".
    Adjacency of a wide character string literal and a character string
    literal is undefined in C90 and C++ Standards, however, in C99, they
    are concatenated to be a wide character string literal.  MCPP
    concatenates them into a wide character string literal.


5.9     Warnings (Class 16)

Trigraphs and digraphs are not used at all in an environment where they
are not need to.  If they are found in such an environment, attention
needs to be paid.  The purpose of the -W16 option is to find such
trigraphs and digraphs.  On the other hand, these warnings are very
bothersome in an environment where trigraphs or digraphs are used on a
regular basis because they are issued very frequently.  For this reason,
I set up a separate class for these warnings.  Anyway, MCPP issues these
messages only in the state where the trigraphs or digraphs are enabled.

2 trigraph(s) converted
    Two trigraph sequences in this physical line have been converted.
    Does the programmer really intend to code trigraph?  MCPPs compiled
    with MODE == STANDARD issues this warning.

2 digraph(s) converted
    Two digraph sequences in this line have been converted.  Does the
    programmer really intend to code digraphs?  MODE >= STANDARD MCPP
    issues this warning.  MCPP compiled with MODE == STANDARD and
    HAVE_DIGRAPHS == FALSE converts a digraph into a regular token in
    the following manner after preprocessing:

        <% -> {      <: -> [      %:    -> #
        %> -> }      :> -> ]      %:%:  -> ##

    Therefore, the compiler-proper is not necessary to be able to handle
    digraphs.  However, MODE == POST_STANDARD MCPP converts a digraph
    into a regular pp-token during the translation phase 1.  The
    difference of this behavior between the modes appears when a #
    operator converts a digraph into a string; MODE == STANDARD MCPP
    directly converts a digraph sequence into a string, while MODE ==
    POST_STANDARD MCPP converts it into a regular pp-token, and then
    into a string.  In addition, if a string literal contains a
    character sequence which is equivalent to a digraph sequence, MODE =
    = STANDARD MCPP does not convert it, while MODE == POST_STANDARD
    MCPP converts it into a character sequence of the corresponding pp-
    tokens.

MODE == STANDARD MCPP does not issue a warning to a diagraph that
appears on a preprocessing-directive line and disappears in a due course
because this warning is issued only to converted diagraphs.


5.10    Diagnostic Messages Index

    Diagnostic Messages             Fatal   Error   Warning class
                                    error           1   2   4   8  16



                    6.  Reporting on Bugs and Others

6.1     MCPP's Bug?

I have developed the Validation Suite to verify conformance of
preprocessing to Standard C.  I will release it along with MCPP source.
The Validation Suite is intended to allow you to verify all the Standard
C preprocessing specifications.  Of course, I used the Validation Suite
to check MCPP.  And what is more, I have compiled MCPP in many compiler
systems to verify compliance.  Therefore, I am confident that MCPP is
now almost flawless, free of bugs and misinterpretation of
specifications, however, I cannot deny the possibility that it still
contains some bugs.

If you find a strange behavior, do not hesitate to let me know.  If you
receive a diagnostic message saying "Bug: ...", it is undoubtedly a bug
of MCPP or a compiler system. (Probably, it's MCPP's.)  How illegal a
user program may be, should MCPP lose control, it is MCPP that is to be
blamed for it.


6.2     malloc() Related Bugs

I have written a library called kmmalloc that contains malloc() and
other memory handling routines. (For details, see 4.extra of porting.txt.)

If such MCPPs exit with an error number from 2120 to 2124 (or 120 to 124),
MCPP or a compiler system undoubtedly contains a bug. (Probably, it is
library function's).

If you write the following directive near the end of a sample source for
testing:

    #pragma __debug_cpp __memory

the heap memory information is output both at the directive and at the
end of the preprocessing.  If a message saying "Heap error: ... " is
output, it undoubtedly indicates that MCPP or a compiler system contains
a bug.

If you find a bug, try to narrow down the problematic area by
sandwiching a portion of the sample source with #if 0 and #endif and
testing it.  Repeat this process until you spot the code with the bug.
I also recommend that you should link MCPP with my malloc(), during the
debugging process.


6.3     How to Report Bugs

When you report a bug, please be sure to provide the following
information:

    Compiler system MCPP is ported to
    A sample source (shorter is better) that allows reproduction of what
        looks like a bug.
    Preprocessing results


6.4     Give Us Your Feedback

Other than bugs, I would appreciate if you give me feedback on MCPP
usage, diagnostic messages or this manual.

For your feedback or information, please post to newsgroups comp.std.c,
fj.comp.lang.c, or send via e-mail.


                                                                   [eof]
