Hi Roland,
Cced to you, because according to the ChangeLog you are the one who
worked most on this macro. Tracing the history of the macro is
difficult, but once you understood it was named AC_MINUS_C_MINUS_O,
you see it has this double checking since the beginning, `initial
revision'.
See
ht
IRIX 6.5, autoconf 2.13
The tclodc packages uses the following construct in configure.in:
#
# Determines the correct binary file extension (.o, .obj, .exe etc.)
#---
The current CVS documentation mixes in the text body the documentation
of recent and obsolete macros. I'm in favor of moving the obsolete
macros documentation into the chapter `Old Names', renamed as
`Obsolete Macros'. Plain lexicographic order seems fine there, no
need to introduce new section
Autoconf CVS has solved this by explicitly including *.C.
case $ac_file in
*.c | *.C | *.o | *.obj | *.xcoff) ;;
*) ac_cv_exeext=`echo $ac_file | sed -e s/conftest//` ;;
Using *.ac_ext sounds good, but it means we have to AC_REQUIRE it
beforehand.
I have no strong opinion o
I have taken a look at the AC_PROG_CC_STDC macro in CVS autoconf and,
while I think it's a step ahead, the solution doesn't strike me as very
far-sighted.
What I would like is a general solution that tells me what compiler I'm
using so I can add custom options. For example, the project I contribu
From: Akim Demaille <[EMAIL PROTECTED]>
Date: 04 Apr 2000 14:05:01 +0200
The macro AC_PROG_CC_C_O (appended below for convenience), according
to the documentation, is expected to check whether the C compiler
supports -c and -o together. But if you read the code, you can see
tha
> "Ian" == Ian Lance Taylor <[EMAIL PROTECTED]> writes:
Ian> That's because the macro was intended for use with GNU make, to
Ian> decide what the default compilation rule should be.
Ah, OK. Thanks Ian!
Ian> On the other hand, changing it breaks backward compatibility for
Ian> no real ad
Date: Tue, 4 Apr 2000 16:34:04 +0200 (MET DST)
From: Peter Eisentraut <[EMAIL PROTECTED]>
What I would like is a general solution that tells me what compiler I'm
using so I can add custom options. For example, the project I contribute
to needs the option '-Olimit 2000' on OSF/1 for
On Apr 4, 2000, Paul Eggert <[EMAIL PROTECTED]> wrote:
> For example, if a compiler needs the option "-Olimit 2000" for some
> reason, then you should write a test case that exposes the need for
> this option, have configure try to compile the test case, and have
> it append the "-Olimit 2000" o
On Apr 4, 2000, Akim Demaille <[EMAIL PROTECTED]> wrote:
> I'm in favor of moving the obsolete macros documentation into the
> chapter `Old Names', renamed as `Obsolete Macros'.
Sounds good to me.
> I'll leave a few @ref where needed.
Even better :-)
> I'm ready to do that.
And it keeps get
On Apr 3, 2000, Russ Allbery <[EMAIL PROTECTED]> wrote:
> Akim Demaille <[EMAIL PROTECTED]> writes:
>> Would doing something like
>> tmp=/tmp/ac.${RANDOM-$$}
>> be good?
> Sure, that would definitely be an improvement on any host with a
> shell that has $RANDOM support, and shouldn't be any w
> On Apr 4, 2000, Paul Eggert <[EMAIL PROTECTED]> wrote:
>
> > For example, if a compiler needs the option "-Olimit 2000" for some
> > reason, then you should write a test case that exposes the need for
> > this option, have configure try to compile the test case, and have
> > it append the
>> On Apr 4, 2000, Paul Eggert <[EMAIL PROTECTED]> wrote:
>>
>> > For example, if a compiler needs the option "-Olimit 2000" for some
>> > reason, then you should write a test case that exposes the need for
>> > this option, have configure try to compile the test case, and have
>> > it appe
13 matches
Mail list logo
