address@hidden (fencepost's file diff daemon)] fencepost:/home/gd/gnuorg/standards.texi -- recent changes

[Akim Demaille]

Hm...  What do you people think about this?

After all, the core code of projects is not supposed to contain any
HAVE_.  Rather, AC_REPLACE_FUNCS etc. should be used, system.h for
other issues, so it seems to me that this is a better solution than
having #if HAVE_ in the main code, but worse than having no such HAVE_
at all.

They should promote the latter IMHO.  OTOH, is it really the purpose
of the GCS to include such information?  That pertains to
autoconf.texi IMHO.

> --- Begin Message ---
>
>
>
> Subject: 
>
> fencepost:/home/gd/gnuorg/standards.texi -- recent changes
>
>
>
>
> Date: 
>
> Wed, 15 Aug 2001 06:10:05 -0400
>
>
>
>
> This is an automated report from fencepost.
> Recent changes to /home/gd/gnuorg/standards.texi:
>
>  124 -r--r--r--    1 bkuhn    user       121517 Aug 14 13:39 
> /home/gd/gnuorg/standards.texi
>
> diff  -u -b  
> /gd3/diffmon/diffmon/old_file_dir/fencepost/fencepost:!home!gd!gnuorg!standards.texi.gz
>  /home/gd/gnuorg/standards.texi
> --- /tmp/diffmon11658   Wed Aug 15 06:10:05 2001
> +++ /home/gd/gnuorg/standards.texi      Tue Aug 14 13:39:15 2001
> @@ -3,7 +3,7 @@
>  @setfilename standards.info
>  @settitle GNU Coding Standards
>  @c This date is automagically updated when you save this file:
> address@hidden lastupdate March 23, 2001
> address@hidden lastupdate August 14, 2001
>  @c %**end of header
>  
>  @ifinfo
> @@ -60,7 +60,7 @@
>  
>  @titlepage
>  @title GNU Coding Standards
> address@hidden Richard Stallman
> address@hidden Richard Stallman, et al.
>  @author last updated @value{lastupdate}
>  @page
>  
> @@ -99,6 +99,7 @@
>  * Managing Releases::       The Release Process
>  * References::              References to Non-Free Software or Documentation
>  * Index::
> +
>  @end menu
>  
>  @node Preface
> @@ -272,6 +273,7 @@
>  * Compatibility::               Compatibility with other implementations
>  * Using Extensions::            Using non-standard features
>  * Standard C::                  Using Standard C features
> +* Conditional Compilation::     Compiling Code Only If A Conditional is True
>  @end menu
>  
>  @node Source Language
> @@ -462,6 +464,57 @@
>  #define P_(proto) ()
>  #endif
>  @end example
> +
> address@hidden Conditional Compilation
> address@hidden Conditional Compilation
> +
> +When supporting configuration options already known when building your
> +program we prefer using @code{if (... )} over conditional compilation,
> +as in the former case the compiler is able to perform more extensive
> +checking of all possible code paths.
> +
> +For example, please write
> +
> address@hidden
> +  if (HAS_FOO)
> +    ...
> +  else
> +    ...
> address@hidden smallexample
> +
> +instead of:
> +
> address@hidden
> +  #ifdef HAS_FOO
> +    ...
> +  #else
> +    ...
> +  #endif
> address@hidden smallexample
> +
> +A modern compiler such as GCC will generate exactly the same code in
> +both cases, and we have been using similar techniques with good success
> +in several projects.
> +
> +While this is not a silver bullet solving all portability problems,
> +following this policy would have saved the GCC project alone many person
> +hours if not days per year.
> +
> +In the case of function-like macros like @code{REVERSIBLE_CC_MODE} in
> +GCC which cannot be simply used in @code{if( ...)} statements, there is
> +an easy workaround.  Simply introduce another macro
> address@hidden as in the following example:
> +
> address@hidden
> +  #ifdef REVERSIBLE_CC_MODE
> +  #define HAS_REVERSIBLE_CC_MODE 1
> +  #else
> +  #define HAS_REVERSIBLE_CC_MODE 0
> +  #endif
> address@hidden smallexample
> +
> address@hidden Standard C and Pre-Standard C
> address@hidden @sc{ansi} C standard
>  
>  @node Program Behavior
>  @chapter Program Behavior for All Programs
>
>
> --- End Message ---