u feel better? We'd rather hash out any bugs
before the release :-) That's why I suggested you might want to try
the CVS tree...
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat
On Jan 25, 2001, Stephen Torri <[EMAIL PROTECTED]> wrote:
> I want to check for the existence of a library not just a function of the
> library as AC_CHECK_HEADERS does. Is there a way to do it?
How about AC_CHECK_LIB or AC_HAVE_LIBRARY?
--
Alexandre Oliva Enjoy Guaran
On Jan 25, 2001, Akim Demaille <[EMAIL PROTECTED]> wrote:
> Should we (i) make sure not to use config.site in the test suite, or
> (ii) have this test grep out this message?
(ii)
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red
> LIBS="$LIBS -l$1 $5"
Nope. First, you look for libraries that don't depend on any others.
Then, you test for libraries that depend on them. Order matters, and
autoconf is doing The Right Thing (TM).
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~o
ter you arrange for INSTALL to be preserved too :-)
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist
(Check out config-ml.in and see how it is
Ralf> applied with automake in newlib.)
> Where is it?
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/config-ml.in
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.c
ortant one, IMO.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist*Please* write to mailing lists, not to me
On Jan 22, 2001, Peter Eisentraut <[EMAIL PROTECTED]> wrote:
> Alexandre Oliva writes:
>> On Jan 22, 2001, Peter Eisentraut <[EMAIL PROTECTED]> wrote:
>>
>> > No message, just continue. Or maybe some message, but continue anyway.
>>
>> At the ex
On Jan 22, 2001, Peter Eisentraut <[EMAIL PROTECTED]> wrote:
> No message, just continue. Or maybe some message, but continue anyway.
At the expense of having all execution tests fail for the wrong
reason?
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliv
> explicit declarations.
> If Alexandre thinks it is OK that stdlib.h be *always* included in
> *all the tests*, then it's fine with me.
My reasoning is that, if it can be #included twice without error, it
can't hurt. I can't think of any problem with it, can you? If not,
I
st belongs in.
This is in the application-specific domain :-)
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Eva
On Jan 21, 2001, Matthew Schalit <[EMAIL PROTECTED]> wrote:
> Is this because I have no CXX or F77 at this point?
Quite likely. Would you give CVS autoconf a spin and see how it goes?
It should probably handle the lack of these programs more gracefully.
--
Alexandre Oliva Enjo
other reason.
It should work for C and older versions of g++, so it should remain
first. A new entry with `#include ' should be added second.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, re
hink of
config.h.in, when the modified configure.in hasn't introduced nor
removed any AC_DEFINEs). In this case, it won't be checked in (unless
someone pushes for it). Thus, even if CVS were smart enough to retain
relative timestamps globally, it would still get this case wrong.
-
On Jan 19, 2001, "Lars J. Aas" <[EMAIL PROTECTED]> wrote:
> On Fri, Jan 19, 2001 at 06:30:48PM -0200, Alexandre Oliva wrote:
> : I'd rather avoid this, if I could help it. Didn't `#include '
> : work as an ac_declaration of exit?
> Sure.
And couldn&
n't `#include '
work as an ac_declaration of exit?
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Eva
include
This would test whether it is safe to #include it twice. The
assumption being that any compiler that is picky enough to require a
declaration of exit() will have a stdlib.h that can be #included
twice.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Ha
es for templates (name + .in).
> Agreed, _because_ of the last argument.
Ditto. .sh and .pl should be scripts, not files that have to be
preprocessed to become scripts. These should be .shin and .plin
Of course, on DOS, these will become m4i, shi and pli, but this is no
big deal.
--
Ale
On Jan 18, 2001, Akim Demaille <[EMAIL PROTECTED]> wrote:
> I'd be OK with making this a warning, at least that's much more
> conventional and understandable. And WARNINGS=error restores the old
> behavior.
Agreed.
> -W forbidden-tokens
> -W forbidden-tokens-i
On Jan 18, 2001, "Lars J. Aas" <[EMAIL PROTECTED]> wrote:
> I've attached a config.log that shows more problems with exit().
We should probably move the '' case first. This will probably work
with VC++, and will be fine for most existing C++ compilers.
-
thin comments? It would be disabled by default, but we'd use
it in our own testsuite.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.uni
x27;s no such macro in autoconf. But you can easily build one using
AC_CHECK_HEADER.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.unicamp.br
ke 1.4a ourselves, but it's not ok
if our package doesn't work with a released version of automake.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampol
On Dec 29, 2000, Harlan Stenn <[EMAIL PROTECTED]> wrote:
> Still doesn't work, and should the second problem be forked into a thread
> for autoconf while this one stays on automake?
Yep, the latter is certainly a problem in autoconf.
What if you take the quotes off "$
supports the `--run' switch relies upon `:'.
Does missing work if, instead of:
"$@" && exit 0
the missing script did:
prog=$1
shift
"$prog" ${1+"$@"} && exit 0
?
--
Alexandre Oliva Enjoy Guarana', see http://www
lude these files in
the distribution, as long as there is a Makefile.am in the directory
containing them. Is there such a Makefile.am? Or did you add the
`config' directory to EXTRA_DIST in the top-level Makefile.am?
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~ol
luded in the distribution
tarball?
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist*Please* write to mail
. We've just release 2.49b, there's the CVS
version, and there won't be that many people in activity during the
holidays, so I see no reason why 2.49c couldn't wait for mid-January
or even February.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/
our use of `4' instead of `a'
is just a technical detail to make it slightly different from the
trademark Automate, that I've held for a while ;-)
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.
OS :-(
> Otherwise we should check what declaration if preferred for the given
> compiler and use it.
It seems to me that the only safe way to go is to test first :-(
Fortunately, we don't have to worry about std::exit(int), since
stdlib.h, unlike cstdlib, must bring std::exit(int) into the gl
On Dec 18, 2000, Pavel Roskin <[EMAIL PROTECTED]> wrote:
> #ifdef __cplusplus
> #undef exit
> #define exit(x) return (x)
> #endif
But what if a hand-coded test calls exit() from some function other
than main()?
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.uni
configure.in'', not ``build
%.sh out of %, with the additional dependency on configure.in''. I
think :-)
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicamp
her way round, in case both exist. Or
should we just print an error and abort in this case?
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.un
On Dec 16, 2000, Akim Demaille <[EMAIL PROTECTED]> wrote:
>>>>>> "Earnie" == Earnie Boyd <[EMAIL PROTECTED]> writes:
Earnie> --- Akim Demaille <[EMAIL PROTECTED]> wrote:
>>> >>>>> "Alexandre" == Alexandre Oli
rectory at the end. Better rm -rf it
up-front, unconditionally.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist*Please* write to mailing lists, not to me
eave that as an exercise for the reader :-)
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist*Please* write to mailing lists, not to me
ning configure, it will overwrite some files in the build
tree anyway (config.h, Makefile, etc), and, if you have a.exe in the
build tree, built off your Makefile, it's probably easy enough to
rebuild it.
But saving and restoring doesn't sound bad either...
--
Alexandre Oliva Enjoy Gu
On Dec 15, 2000, Assar Westerlund <[EMAIL PROTECTED]> wrote:
> And is using basename a reasonably solution in this case?
basename is not portable. You should use sed instead.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat
On Dec 14, 2000, "Lars J. Aas" <[EMAIL PROTECTED]> wrote:
> On Thu, Dec 14, 2000 at 03:26:51PM -0200, Alexandre Oliva wrote:
> : Maybe we should just rule the user shouldn't have a program named `a'
> : and be done with it?
> I want to have a program call
.
Doesn't it even report the fact that it doesn't know how to build
foo.o?
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.unicamp.b
ead of a.out. The problem
is that a.out and a.exe are outside the file name space of autoconf.
The user might well have an a.exe program in their tree that they
wouldn't want configure to overwrite/remove.
Maybe we should just rule the user shouldn't have a program named `a'
and be d
let them receive our signal: we are
> setting a standard they have to follow :)
This doesn't fit in very well with autoconf's standards :-)
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, re
a patch for GCC that fixed the extension conventions, but it
can't be applied because it breaks the build because of
autoconf-related issues. Catch 22 :-(
I guess we'll have to retain the host checking of Cygwin, Mingw and
OS2 and explicitly set EXEEXT to .exe if host is any of these
macro. Since it's supplied only for
backware-compatibility issues, and its use is strongly discouraged, I
think we can go with that.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com
On Dec 12, 2000, Earnie Boyd <[EMAIL PROTECTED]> wrote:
> --- Alexandre Oliva <[EMAIL PROTECTED]> wrote:
>> On Dec 12, 2000, Mo DeJong <[EMAIL PROTECTED]> wrote:
>>
>> > I guess the question is, should we try to compile something and the
>> &g
x27;t it tell cygwin from mingw?
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist*Please* write to mailing lists, not to me
ut or a.exe, which
is out of autoconf's file namespace.
> There remains just one question: what shall we do out of AC_CYGWIN etc.
> simply AU define them as [...]
Yep.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer ao
ns used by the compiler. Fine.
Yep.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist*Please* write to mailing lists, not to me
oss-compilers, so this point is moot.
> Why isn't `install' which takes care of this on the *host* machine?
Because install isn't part of a cross-build environment, so it knows
only about the build machine, not the host machine?
--
Alexandre Oliva Enjoy Guarana',
On Dec 12, 2000, Peter Eisentraut <[EMAIL PROTECTED]> wrote:
> Akim Demaille writes:
>> * host or build characteristic?
> It's neither, nor should it matter. It's a feature of the compiler.
I wholeheartedly agree.
--
Alexandre Oliva Enjoy Guarana', see htt
t; That's what I call a build characteristic.
Wrong. The compiler is supposed to follow conventions of the host,
not of the build machine.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS
7; on GNU/Linux, at least for the
architecture. WRT object file format and C library version, is there
any reason why testing for __ELF__ and __GLIBC__ isn't enough?
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cyg
On Nov 30, 2000, Akim Demaille <[EMAIL PROTECTED]> wrote:
> | +./configure: 24794 Alarm call
> This is why it failed: the shell complains of the signal it received.
> I don't know how to address this issue.
`grep -v "Alarm call"' before comparing the o
On Nov 29, 2000, Akim Demaille <[EMAIL PROTECTED]> wrote:
> * aclang.m4 (AC_LANG_INT_SAVE): Don't add any trailing new line,
> and close the file.
Ok
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer
ir}/../..
fi`)
(note that this is similar to the sample snippet in config-ml.in about
how to add it to configure.in, but there's one additional `..' in each
set, because libsupc++ is a sub-directory of libstdc++-v3)
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~o
ION, 2.49b)
>> >
>> > gmake[1]: *** [autoconf.m4f] Error 1
>>
>> I could not reproduce the problem on RedHat Linux 6.2. Give us more
>> information about the OS you are using and the commands you are running.
> My bad - sorry. Solaris 2.8.
Which version
hing alike in the autoconf macro
archive. http://research.cys.de/autoconf-archive/
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.unicamp.br, gnu.org}
F
On Nov 23, 2000, Johan Rydberg <[EMAIL PROTECTED]> wrote:
> Is it posible to add an generic argument (not an --enable-* or
> --with-* argument) to the configure script?
Nope.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red
On Nov 20, 2000, Ralf Corsepius <[EMAIL PROTECTED]> wrote:
>> I suppose a compiler that targets Cygwin will create .obj object files
>> regardless of its host platform.
> Hmm, I don't think so.
I stand corrected. Anyone can tell I keep off MS-Windows as much as
p
gardless of its host platform. I.e., when using a Cygwin compiler
(such that --host=cygwin), you'd get OBJEXT=.obj, regardless of the
build platform, on which the compiler runs.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer
f char charchar[2]'?
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist*Please* write to mailing lists, not to me
On Nov 17, 2000, Mo DeJong <[EMAIL PROTECTED]> wrote:
> I always assumed that EXEEXT and OBJEXT were based on the
> host. After all, they should depend on the output of the compiler,
> right?
Agreed.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br
On Nov 16, 2000, Akim Demaille <[EMAIL PROTECTED]> wrote:
> Does anybody know whether using fopen (foo, "wb") is portable?
I recall having read about it in ``The C Primer'', when I first
learned C back in 198x (for a small value of x).
--
Alexandre Oli
On Nov 14, 2000, Akim Demaille <[EMAIL PROTECTED]> wrote:
>>>>>> "Alexandre" == Alexandre Oliva <[EMAIL PROTECTED]> writes:
Alexandre> On Nov 10, 2000, Akim Demaille <[EMAIL PROTECTED]> wrote:
>>> My bet is that IFS splitting is not p
27;s no difference between the quoted
and unquoted versions.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Eva
length. I'm inclined to think we should support prefixes as
long as the user wants, unless we have a very good reason not to do
so. I mean, it's not like it would be too costly to add a sed-based
fallback here. Or would it?
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unic
On Nov 9, 2000, Pavel Roskin <[EMAIL PROTECTED]> wrote:
> Hello, Peter!
>> Exactly where do the GNU standards say anything about long option
>> abbreviation?
> Well, whenever I don't check something I'm punished for that :-)
Oops. Ditto :-)
--
Alexandr
On Nov 8, 2000, Benoit Hudson <[EMAIL PROTECTED]> wrote:
> Where in autoconf would I look to do this consistently?
In the autoconf CVS tree. I believe both issues have already been
addressed :-)
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat
> additional tedious maintenance nightmare.
Agreed. But beware that this decision goes against the GNU Coding
Standards.
I guess we should have an m4sh (or m4sugar?) macro to expand:
m4_case_longopt([foo])
to
--foo | --fo | --f
and
m4_case_longoptarg([foo])
to
--foo=* | --fo=* | --f=*
--
g_gxx, ac_cv_prog_g77):
>> Be AU_DEFUN'd, not AU_ALIAS'd.
> I'm afraid you are fixing the symptoms, not the cause.
I agree. When $# is 0, we shouldn't add parentheses.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer
On Nov 6, 2000, Earnie Boyd <[EMAIL PROTECTED]> wrote:
> Why not just prefix the reserved autoconf variables with a `_'
> character?
autoconf used to use just AC_/ac_. Now, it's claiming ownership of
all of A?_. That looks a little bit exagerated to me.
--
Alexandre
ffi, available at sources.redhat.com
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist*Please* write
On Nov 3, 2000, Akim Demaille <[EMAIL PROTECTED]> wrote:
>>> What should we do?
Pavel> Return to A[CHMSTU]
Agreed.
> I'd rather take almost the full range and except AR if we need to.
I'd prefer that the maintainers of autoconf weren't so greedy abo
ing for pthreads.h (search in
configure.in, acinclude.m4 and aclocal.m4) and add an `echo $CC'
around that, to tell whether -pthread has been added to CC.
> So the point:
[end of message]
I see. Problem solved :-) :-)
--
Alexandre Oliva Enjoy Guarana', see http://www.i
27;, I get
the expected output.
> default="a b c"
> : list1=${list1=$default}
> : list2=${list2="$default"}
> echo "$list1" | cat -v
> echo "$list2" | cat -v
This works on Ultrix both with and without `:'.
--
Alexandre Oliva
I remove the special case for '.' then it
> still works with no modification to the subpackages configuration
> scripts.
Try `cd build; ../configure' after removing the `.' special case and
note that it won't work. You're just barking up the wrong tree.
--
Al
On Oct 26, 2000, Mo DeJong <[EMAIL PROTECTED]> wrote:
> They will want to run:
> ./configure ; make install
> and expect it to work, and why shouldn't they?
Works for me. When doesn't it work for you? Did you read my other
message in this thread?
--
Alexandre Ol
{Two}
+ echo {lines}
{lines}
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist*Please* write to m
;ve been unable to log into it yet.
I'm looking into it.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.unicamp.br, gnu.org}
Free Software
On Oct 26, 2000, Akim Demaille <[EMAIL PROTECTED]> wrote:
> Yep, I did. I think we are going to drop the `expr' branch.
Yep. We've had too much trouble with `expr' so far.
Maybe we could test whether `expr' works for us, and set `expr=false'
if it doesn&#
level directory, you'd have
to adjust every sub-package so as to ``know'' it's been packaged
differently.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC
inglequote upfront, and use
it in error messages instead?
> Maybe we should go to using `read'
What do you mean? In what sense would this help?
> Is xargs portable?
Nope.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer
> See the module `config' on subversions.gnu.org
... or get the latest version from ftp://ftp.gnu.org/pub/gnu/config/
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-U
le> I had no problem until now. I hadn't compiled textutils until
Daniele> 20 minutes ago, but now I did and it all went smoothly.
> That's a clear sign we have to maintain your architecture alive :)
Yup
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.b
ising, since it's just a set of shell-scripts and m4
macros.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.unicamp.br, gnu.org}
Free Sof
g other sed's
Good point.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist*Please* write to mailing lists, not to me
On Oct 23, 2000, David Andrew Michael Noelle <[EMAIL PROTECTED]> wrote:
> Can anyone tell me what might cause this?
A config.cache or config.site that incorrectly sets cache variables.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat
On Oct 20, 2000, David Morgan <[EMAIL PROTECTED]> wrote:
>> test s${CONFIG_FILES+et} = set || CONFIG_FILES=$config_files
>>
> This works
How about:
CONFIG_FILES=${CONFIG_FILES-$config_files}
?
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unic
On Oct 18, 2000, Akim Demaille <[EMAIL PROTECTED]> wrote:
> * acspecific.m4 (AC_PROG_LN_S): If neither `ln -s' nor `ln' work,
> fall back to `cp'.
Ok
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer
e is precisely to detect this kind of problem.
Might I ask you to please investigate why it gets in an infinite loop?
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at I
ou think it would be to allow for AC_REQUIRE to accept
arguments for the macro to be called?
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.
[AC_TRY_COMPILE([#include ], [toto()],
TOTO=true, TOTO=false)])
^ ^
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD stud
ght be a macro able to do what you want. The problem is that
you'd have to supply the right type for the variable, and any headers
required to declare the types (and maybe the variable itself), so it's
not too much different from using AC_TRY_LINK directly.
--
Alexandre Oliva Enjoy Gua
r this problem is to have both
VAR_FOR_HOST and VAR_FOR_BUILD, instead of just VAR. When
not cross-compiling, both would default from just VAR, if VAR is set,
but for programs that support cross-compilation, this distinction is
important, and AC_PATH_PROG just isn't enough :-(
--
Alexandre Ol
CK_FILE to get rid of the warning, without understanding the
real consequences of doing it.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.unic
ompiler happens to just use the stdlib.h offered by the OS?
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist*Please* write to mailing lists, not to me
extern "C" void exit (int);
#endif
#include
The first program we'd attempt to compile would determine wehther
CXX_NEEDS_EXIT_DECL, and #define it in confdefs.h, so that further
compilations would be able to tell whether to declare it or not.
--
Alexandre Oliva Enjoy Guarana',
run configure again.
Thanks to David Taylor <[EMAIL PROTECTED]> for reporting the problem!
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist*Please* write to mailing lists, not to me
-9]*($pattern)[_A-Za-z0-9]*)/)
^^^
I don't think we can start macro names with digits. How about:
- match (\$0, /($pattern)[_A-Za-z0-9]*/)
+ match (\$0, /([_A-Za-z][_A-Za-z0-9]*($pattern)[_A-Za-z0-9]*)/)
?
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unica
for configure.gnu.
* doc/autoconf.texi: Document checking for configure.gnu when
recursing subdirectories.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Uni
101 - 200 of 531 matches
Mail list logo
