>>> "Tom" == Tom Tromey <[EMAIL PROTECTED]> writes:
[...]
adl> - $_am_dirpart can fooled when $install_sh is defined
adl> using $am_missing_run
Tom> When does this happen?
When install-sh is not found, see the bottom of
http://sources.redhat.com/ml/automake/2001-02/msg00335.html
adl> - $
> From: Tom Tromey [mailto:[EMAIL PROTECTED]]
> > "Arjan" == Arjan J Molenaar <[EMAIL PROTECTED]> writes:
>
> Arjan> Imagine this: I want to build the gnome-libs from
> Arjan> source. Therefore I need GLib, GTK+, Pango and a few other
> Arjan> packages already build and installed. It would
>>> "Pavel" == Pavel Roskin <[EMAIL PROTECTED]> writes:
[...]
Pavel> I understand that your test case was to demonstrate the
Pavel> first problem. But then "make distcheck" is irrelevant -
Pavel> the problem can be detected already after "make distdir"
Yes!
[...]
Pavel> However, please co
Tom Tromey <[EMAIL PROTECTED]> writes:
> I agree -- please keep them at the declaration and remove them from
> initialize_per_input. Keeping them updating in parallel won't work.
OK, I will.
Tom Tromey <[EMAIL PROTECTED]> writes:
> Akim> +?MAINTAINER-MODE?$(srcdir)/configure:
>@MAINTAINER_MODE_TRUE@$(srcdir)/%CONFIGURE-AC% $(ACLOCAL_M4) $(CONFIGURE_DEPENDENCIES)
> Akim> +?!MAINTAINER-MODE?$(srcdir)/configure: $(srcdir)/%CONFIGURE-AC% $(ACLOCAL_M4)
>$(CONFIGURE_DEPENDENCIES)
>
> Wh
Tom Tromey <[EMAIL PROTECTED]> writes:
> I'm not convinced this really makes things cleaner.
> It is just shifting the ugliness around.
:) :) :)
Tom Tromey <[EMAIL PROTECTED]> writes:
> I don't really want to propagate its use. It is good to make it work,
> and you can make uninstall-info if you really want. However I don't
> see a real need for it. If we do add it I'd prefer we not document
> it.
OK. I will include it as is since cu
Tom Tromey <[EMAIL PROTECTED]> writes:
> > "Akim" == Akim Demaille <[EMAIL PROTECTED]> writes:
>
> Akim> Question: why do we grep . after grep -v? It is not guaranteed that
> Akim> grep -v will exit positive when it output something?
>
> I don't know. cvs annotate claims you added those l
Hello!
Automake from CVS doesn't work with perl5.6.0 today:
$ perl5.6.0 -c automake
Global symbol "%require_file_found" requires explicit package name at
automake line 3291.
BEGIN not safe after errors--compilation aborted at automake line 3830.
Obviously, %require_file_found should be declared
Tom Tromey <[EMAIL PROTECTED]> writes:
> With this patch we'll be doing this all the time instead of only in a
> certain obscure situation. Bummer.
I agree...
Another issue is that of @FOO@ variable in bin_PROGRAMS etc.
> Akim> -test "`fgrep '.c.o' Makefile.in | wc -l`" -eq 1
> Akim> +grep '
Tom Tromey <[EMAIL PROTECTED]> writes:
> I don't want to unify the .am readers for 1.5. My experience is that
> automake users have all kinds of ugly stuff in their Makefile.am. We
> definitely don't want to support this stuff long-term, but I think
> changing it now would be ill-advised.
I ag
Pavel Roskin <[EMAIL PROTECTED]> writes:
> Hello!
>
> Automake from CVS doesn't work with perl5.6.0 today:
>
> $ perl5.6.0 -c automake
> Global symbol "%require_file_found" requires explicit package name at
> automake line 3291.
> BEGIN not safe after errors--compilation aborted at automake lin
Hello, Akim!
> > Obviously, %require_file_found should be declared by "use vars", but I
> > would prefer if the author does it. In fact, it may be a legitimate case
> > for a variable to be declared local.
>
> You tried in the middle of a batch of patches. It was fixed before
> your message arri
Hello, Akim!
configure.am is a wonderful name (how long before we see Makefile.ac?),
but you forgot to add in to Makefile.am
Rebuilding makefiles discovers an interesting side effect - OBJEXT and
EXEEXT are now defined in Makefile.in as constants. I had to commit them,
but please make sure that
Pavel Roskin <[EMAIL PROTECTED]> writes:
> Great! I wish your answer about the next problem would be the same!
>
> Only one test is failing now - installsh.test. Perl reports the same
> mesages many times, so I ran the output through sort and uniq.
Yep, I am aware of this. This is a side effec
I don't feel like feeling this issue, I already have much to fight
with, but FYI, there are two tests like this which have distcheck
stuck. Until you press enter. It turns out that...
automake-1.4e/=build/tests % make check VERBOSE=yes TESTS=insh.test
make check-TESTS
make[1]: Entering direct
Akim> Ahem, sorry...
Akim> I did this:
Akim> set -e
I don't see that in the patch.
Tom
Tom Tromey <[EMAIL PROTECTED]> writes:
> I don't think it is actually important to support non-GNU make
> implementations here. We've never supported Makefile rebuilding with
> vendor makes in the past. No other GNU Makefile has. We've also
> never had a complaint about it. Since it sounds li
Akim> Another issue is that of @FOO@ variable in bin_PROGRAMS etc.
This has always been a problem with EXEEXT. The rule is that if you
use @FOO@ in _PROGRAMS then you have to remember to add EXEEXT
yourself in configure.in.
Akim> I did this:
Akim> set -e
Or maybe you meant you changed your pat
Tom Tromey <[EMAIL PROTECTED]> writes:
> Akim> Ahem, sorry...
> Akim> I did this:
>
> Akim> set -e
>
> I don't see that in the patch.
Sorry I was unclear. I meant that I updated the patch, and applied
what I included in that message.
Tom Tromey <[EMAIL PROTECTED]> writes:
> Akim> Another issue is that of @FOO@ variable in bin_PROGRAMS etc.
>
> This has always been a problem with EXEEXT. The rule is that if you
> use @FOO@ in _PROGRAMS then you have to remember to add EXEEXT
> yourself in configure.in.
Time for another set
I preferred to have transform _required_ a defined value for each key,
to make sure there are no accidents. So I fixed the sites where
%transform receives an uninitialized value.
There are many other places to fix, most typically $2 etc. which do
not exist, such as in the example below.
Index:
Tom, I'd like to know what you'd prefer: fixing the uses of $2 etc.,
or the definition of the groups in the regexps? For instance, I have
this:
Index: automake.in
===
RCS file: /cvs/automake/automake/automake.in,v
retrieving revisio
Akim> I preferred to have transform _required_ a defined value for
Akim> each key, to make sure there are no accidents. So I fixed the
Akim> sites where %transform receives an uninitialized value.
Sounds great.
What remains in order to enable `use strict'?
Tom
> "Akim" == Akim Demaille <[EMAIL PROTECTED]> writes:
>> Also, `all' isn't the only time that Makefiles can be rebuilt.
Akim> I very much agree, and was puzzled by this. I guessed we meant
Akim> to provide some support for the most common case.
Could be. Maybe the ChangeLog provides a clu
Akim> Time for another set of explanations :)
Akim> I seem to have understood that we don't support
Akim> if foo
Akim> bin_PROGRAMS += foo
Akim> fi
Akim> (or am I wrong?). If so, why?
I think we ought to support this but we currently do not. We don't
support it because, previously, making it
Tom Tromey <[EMAIL PROTECTED]> writes:
> Akim> I preferred to have transform _required_ a defined value for
> Akim> each key, to make sure there are no accidents. So I fixed the
> Akim> sites where %transform receives an uninitialized value.
>
> Sounds great.
>
> What remains in order to enabl
Tom Tromey <[EMAIL PROTECTED]> writes:
> Akim> Time for another set of explanations :)
> Akim> I seem to have understood that we don't support
>
> Akim> if foo
> Akim> bin_PROGRAMS += foo
> Akim> fi
>
> Akim> (or am I wrong?). If so, why?
>
>
> I think we ought to support this but we current
Hello!
CVS Automake attempts to install noinst_SCRIPTS. The following test fails:
___
#! /bin/sh
# Test to make sure that noinst_SCRIPTS are not installed.
# From Pavel Roskin.
. $srcdir/defs || exit 1
cat > Makefile.am << 'END'
noinst_SCRIPTS = foo
END
$AUTOMAKE |
> "Pavel" == Pavel Roskin <[EMAIL PROTECTED]> writes:
Pavel> CVS Automake attempts to install noinst_SCRIPTS.
I'm guessing this is related to Akim's latest patches.
Pavel> cat > Makefile.am << 'END'
Pavel> noinst_SCRIPTS = foo
Pavel> END
Please check this in.
Bonus points if you also updat
Pavel Roskin <[EMAIL PROTECTED]> writes:
> Only one test is failing now - installsh.test. Perl reports the same
> mesages many times, so I ran the output through sort and uniq.
> $ make check TESTS=installsh.test VERBOSE=1 2>&1 | sort | uniq
> 1 of 1 tests failed
> === Running test ./installsh.t
Hello!
`make distcheck' is working again. No more questions from `rm'.
Regards,
Pavel Roskin
__
Index: ChangeLog
===
RCS file: /cvs/automake/automake/ChangeLog,v
retrieving revision 1.1149
retrieving
Hello, Tom!
> I'm guessing this is related to Akim's latest patches.
Akim does a great job, but he needs a Mozilla-style tinderbox
(http://tinderbox.mozilla.org/) to alert him when something goes wrong.
> Please check this in.
> Bonus points if you also update it to check for `checkdir', which
Hi,
I am a newbie to the autotools, and
needed some help with a project I am working on.
I want to be able to generate a
Makefile that takes some command line params and accordingly sets or does not
set the "-g", "-O" etc flags. I don't want them to be parameters to the
configure s
Index: ChangeLog
from Akim Demaille <[EMAIL PROTECTED]>
* automake.in (&handle_merge_targets): Ventilate its non `all'
related content into...
(&generate_makefile): here.
(&handle_merge_targets): Rename as...
(&handle_all): this.
Remove a useless `
I *love* query-replace patches, there are not enough of this sort :)
Index: ChangeLog
from Akim Demaille <[EMAIL PROTECTED]>
* automake.in: Backquote is inert in double quotes.
Index: Makefile.am
--- Makefile.am Mon, 05 Mar 2001 22:33:22 +0100 akim (am/f/46_Makefile.a 1.15 644)
+++ Mak
Index: ChangeLog
from Akim Demaille <[EMAIL PROTECTED]>
* automake.in (&push_dist_common): Simplify.
Index: automake.in
--- automake.in Sun, 04 Mar 2001 21:51:47 +0100 akim (am/f/39_automake.i 1.113 755)
+++ automake.in Sun, 04 Mar 2001 22:39:31 +0100 akim (am/f/39_automake.i 1.113 755)
FYI, here is the result on the fileutils:
~/src/fu % makesort -u Makefile.in.old Makefile.in1:26 remo
--- Makefile.in.old.sortTue Mar 6 01:27:04 2001
+++ Makefile.in.sortTue Mar 6 01:27:04 2001
@@ -307,8 +307,8 @@
END { for (i in files) print i; }'
The goal is to use -w *and* to have the test suite fail when there are
-w-warnings.
Note that I could catch the `$_ needs be saved' issue because I could
see the errors from sinclude.test, i.e., before that I realized that
--Wno-error was needed by this test, I had to observe the many
complaints
On Mar 5, 2001, Alexandre Duret-Lutz <[EMAIL PROTECTED]> wrote:
"Tom" == Tom Tromey <[EMAIL PROTECTED]> writes:
Tom> What if we make install-strip unconditionally use install-sh? Do
Tom> we really care about the small performance impact on a
Tom> rarely-used target?
> I don't. Maybe Ale
On Mar 5, 2001, Tom Tromey <[EMAIL PROTECTED]> wrote:
> We use a rule that rebuilds Makefile if Makefile.in has changed. GNU
> make will try to build `Makefile', if such a target exists, before
> running any other target, including `distclean'.
Maybe we could convince the maintainer(s?) of GNU
41 matches
Mail list logo
