autoconf-patches
[Top][All Lists]
Advanced

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

RE: Fix EXEEXT mechanism


From: Tim Van Holder
Subject: RE: Fix EXEEXT mechanism
Date: Mon, 22 Jan 2001 19:28:13 +0100

> Good, but then do that for all the languages please.
My bad - ran across this when configuring a C package; didn't occur
to me to check for other languages.

> |  - use AS_IF(AC_TRY_EVAL(ac_link)) instead of AC_LINK_IFELSE
> |    in AC_EXEEXT_O, as the latter also requires a correct ac_exeext
> It's your change that makes this needed, right?  I'm not feeling good
Actually, no. You can't use AC_TRY_LINK here for the same reason you
can't use it in EXEEXT_DEFAULT: we're trying to find the correct
ac_exeext, and AC_TRY_LINK uses test -s conftest$ac_exeext as part of
its result. So unless EXEEXT_DEFAULT found .exe (and it doesn't for me),
this test fails, and EXEEXT_O won't even try to look for conftest.exe,
instead assuming the link failed (while it didn't).

> about weakening the test.  I'd much more prefer having the macros
> become more independent from each other, for instance, can't we write
> AC_LANG_COMPILER_WORKS so that it does not depend upon the extensions?
> For instance it could work with `-o' and use the default output?
As long as it doesn't use AC_TRY_LINK and friends, but instead checks
whether $ac_link is successful, this could work. But since we want a
good ac_exeext to be found, I don't see why this is so important (after
all, the exeext tests are only run once, aren't they?)

> Hm, I don't know.  This patch seems good to me, but then I would like
> to ask you to include a few AC_BEFORE to make sure we never flip these
> guys again.  *And* please, adjust the other languages too.
Not sure what you mean by the AC_BEFOREs (but the info file will help
there, I expect); I'll resubmit the patch with all languages corrected
(probably later today; otherwise, tomorrow).




reply via email to

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