[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Autoconf] Re: What compiler was it?

From: Ian Lance Taylor
Subject: [Autoconf] Re: What compiler was it?
Date: 19 Sep 2000 09:09:06 -0700

   Date: Mon, 11 Sep 2000 17:42:38 -0400 (EDT)
   From: Pavel Roskin <address@hidden>

Sorry for the slow reply.  I was on vacation.

   You made the following change in Autoconf:

   Thu May 28 18:37:36 1998  Ian Lance Taylor  <address@hidden>
           (AC_TRY_CPP): Remove lines from stderr which contain only the name
           of the file.

   Do you remember what compiler produced such lines? I'm especially
   interested whether it's a pure C compiler or C++ is affected as well.

Here is the entire ChangeLog entry that I see in my copy of the

        Add support for Visual C++:
        * acgeneral.m4 (ac_exeext, ac_objext): New variables.
        (ac_link): Use ac_exeext.
        (AC_TRY_CPP): Remove lines from stderr which contain only the name
        of the file.
        (AC_REPLACE_FUNCS): Use ac_objext.
        * acspecific.m4 (AC_PROG_CC): Check whether compiler supports -g
        even if it is not gcc.
        (AC_PROG_CXX): Likewise.
        (AC_PROG_CPP): Try running the compiler with the -nologo option.
        (AC_FUNC_ALLOCA): Check for _MSC_VER.  Use ac_objext.
        (AC_FUNC_MEMCMP): Use ac_objext.
        (AC_STRUCT_ST_BLOCKS): Likewise.
        (AC_OBJEXT): New macro.
        (AC_CYGWIN32, AC_MINGW32, AC_EXEEXT): New macros.
        * configure: Rebuild.

As you can see from the first line of the ChangeLog entry, the
compiler in question was Visual C++.  This is a Microsoft product.  It
handles both C and C++, so C++ is indeed affected.

   Also a comment needs to be added to AC_TRY_CPP to avoid such questions in
   the future.

   Unfortunately, the mailing list archives start with October 1998, so I
   cannot find the answer there.

   For everybody else, it's a good idea to comment what breakage is being
   worked around to avoid questions two years later :-)

I could have added a comment to AC_TRY_CPP.  I'm undecided as to
whether that would be a good idea or not.  I think the ChangeLog entry
does describe the breakage adequately.  But then, I am not of the
school of thought in which every action requires an extensive


reply via email to

[Prev in Thread] Current Thread [Next in Thread]