> "Chris" == Chris Cowan <[EMAIL PROTECTED]> writes:
Chris> If possible, I would like to edit the suffix rules in one
Chris> place. I realize that I can use include processing. I'm just
Chris> wondering if there is a way I can add the rules to the toplevel
Chris> Makefile.am or configure.in
> ">" == K Viswanathan <[EMAIL PROTECTED]> writes:
>> I am facing a problem with texinfo.tex. When I invoke
>> automake from the top-level directory, I am getting an
>> error message like,
>> docs/Makefile.am: `texinfo.tex' missing @setfilename
I think this could only happen if you put `tex
> "F.J." == F J Franklin <[EMAIL PROTECTED]> writes:
F.J.> Hi, I notice automake still assumes that makefiles are called
F.J.> "Makefile" (see depout.m4, lines 27-28) and not "GNUmakefile".
F.J.> I posted a patch a while back but got ignored. Please could
F.J.> someone add the two lines whic
> "Srinivas" == Srinivas Cheruku <[EMAIL PROTECTED]> writes:
Srinivas> runtest not found, check skipped
Srinivas> It says check skipped. can you tell me why it has skipped
Srinivas> the check.
"runtest not found".
`runtest' is the name of the dejagnu driver program.
You must not have dejagn
> "Patrick" == Dahiroc, Patrick <[EMAIL PROTECTED]> writes:
Patrick> lnklib: mylib.a
Patrick>ln -s ./mylib.a $(top_builddir)/lib
Patrick> i dont get the right link when i run 'make lnkdir'.
I think you're invoking `ln -s' incorrectly.
You probably want:
ln -s `pwd`/mylib.a $(to
Automake 1.5b is available. This is a prerelease.
You can get it here:
ftp://alpha.gnu.org/gnu/automake/automake-1.5b.tar.bz2
ftp://alpha.gnu.org/gnu/automake/automake-1.5b.tar.gz
>From the NEWS file:
* Autoconf 2.52 is required.
* `dist' generates all the archive flavors, as did `dist-al
> "Ted" == Ted Irons <[EMAIL PROTECTED]> writes:
Ted> It is my understanding (from reading the info pages)
Ted> that the install-data-hook target is done after the
Ted> install-data target.
Yes.
Ted> install-data-local:
Ted> cmd1
Ted> cmd2
Here you're using `install-dat
> "Alexandre" == Alexandre Duret-Lutz <[EMAIL PROTECTED]> writes:
>> SUFFIXES = .k
Alexandre> (This is superfluous, Automake will infer it from the .k.o
Alexandre> rule below)
Is it still documented as being required?
Tom
> "Lars" == Lars J Aas <[EMAIL PROTECTED]> writes:
Lars> However, since the problem has always been there and may prove
Lars> difficult to fix, you could just go ahead and get Automake 1.5
Lars> out the door and plan this rewrite for 1.5.1 or something? 1.5
Lars> is long overdue as it is.
T
> "David" == Lawrence, David (STEI) <[EMAIL PROTECTED]> writes:
David> I am already building using an rpm tree under by home
David> directory. But it would appear that RPM wants me to place the
David> files into their install location before the binary RPM is
David> built. Is this a hard re
> ">" == Motoyuki Kasahara <[EMAIL PROTECTED]> writes:
>> Makefile.in defines the variables `_am_include' and `_am_quote',
>> but make on that OS cannot handle variable name beginning with
>> `_'. Would you rename the variables?
Thanks. Another annoying vendor make bug strikes!
I'll defin
> "Lars" == Lars J Aas <[EMAIL PROTECTED]> writes:
Tom> How could we fix it? I mean, conceptually.
First off, let me just say that I was considering a larger problem. I
was thinking about how to solve the exponential explosion problem for
any random variable whose contents we cared about.
> "David" == Lawrence, David (STEI) <[EMAIL PROTECTED]> writes:
David> I did not mean to offend by offering the example, I was only
David> trying to outline my desire within the terms I understood best.
Don't worry, I don't think you offended anybody.
David> If possible, I would like to sup
> ">" == OKUJI Yoshinori <[EMAIL PROTECTED]> writes:
>> I found one bug about the `mostlyclean' target. The bug occurs when,
>> in a Texinfo manual, you combine multiple indices into one index, but
>> you don't use the command for this index at all.
Thanks. I think your patch is fine and I'
> "Lars" == Lars J Aas <[EMAIL PROTECTED]> writes:
Lars> I don't think the exponential nature of how AM_CONDITIONALs are
Lars> handled is fixed yet, which *I* would consider a showstopper.
Lars> Haven't checked latest CVS though, so excuse me if it's already
Lars> fixed.
I don't think it has
> "Robert" == Robert Boehne <[EMAIL PROTECTED]> writes:
Robert> I'm having problems with my Libtool/Automake project when
Robert> using dependency tracking. It seem that the variable
Robert> DEPDIR isn't set anywhere in any of the Makefiles.
Robert> This may be because the top level Makefile
I've released automake 1.4h. Get it here:
ftp://sources.redhat.com/pub/automake/automake-1.4h.tar.gz
This version fixes a bunch of bugs that were in 1.4f. Right now
automake is in feature freeze. I believe there is one bug I want to
fix (handling of a computed AC_AUX_DIR), and of course t
> "Pavel" == Pavel Roskin <[EMAIL PROTECTED]> writes:
Pavel> I suggest that if a failing test is submitted, it should be
Pavel> added to XFAIL_TESTS in the same commit. It should be removed
Pavel> from XFAIL_TESTS in the same commit that fixes it.
I agree. Sorry about that. I checked in th
> "Dean" == Dean Povey <[EMAIL PROTECTED]> writes:
Dean> I didn't actually know about the AM_ macro when I wrote this
Dean> code though.
Make sure you look at the automake macro. For instance this code is
required to work around a GNU make feature:
# We grep out `Entering directory' an
> "Clark" == Clark Rawlins <[EMAIL PROTECTED]> writes:
I never saw an answer to this.
Clark> I got automake and autoconf out of cvs and did
Clark> [ ... ]
Clark> on both projects autoconf first then automake
Clark> and now I get some errors when I run aclocal and automake.
Clark> %aclocal
C
> "Lars" == Lars J Aas <[EMAIL PROTECTED]> writes:
Tom> You can use ACLOCAL_AMFLAGS (bad name, I know) in Makefile.am.
Lars> That doesn't help if you invoke aclocal manually though - only when
Lars> make invokes aclocal (or have I misunderstood something?).
Nope, you're right.
We really hav
> "Bob" == Bob Proulx <[EMAIL PROTECTED]> writes:
Bob> Therefore shouldn't I run the following sequence to bootstrap the
Bob> project from nothing? This order comes from page 70 of the goat book
Bob> which I know describes an older version of the programs.
Bob> aclocal -I config \
Bob>
> "Nathan" == Nathan Neulinger <[EMAIL PROTECTED]> writes:
Nathan> I dunno, is it really "mere aggregation" if that 'missing'
Nathan> script is required as part of the build process for a non-gpl
Nathan> app?
I decided to make it clear by adding the exception to `missing'.
Tom
> "Alexandre" == Alexandre Oliva <[EMAIL PROTECTED]> writes:
Alexandre> I'd appreciate a comment before the test explaining why it
Alexandre> is safe to do the test there even though missing isn't
Alexandre> always guaranteed to run in the directory containing
Alexandre> configure.{in,ac}.
I
> "Clark" == Clark Rawlins <[EMAIL PROTECTED]> writes:
Clark> As Gary suggested I made a new debian package by applying the
Clark> diff file from the automake source deb packages to the
Clark> automake-1.4f distribution.
What automake patches are in Debian?
Tom
> "Johannes" == Kremp, Johannes (Extern) <[EMAIL PROTECTED]> writes:
Johannes> and in my 'Makefile.am':
Johannes> if HELLO
Johannes> include hellofiles
Johannes> bin_PROGRAMS = hello
Johannes> hello_SOURCES = $(FILES) #from file 'hellofiles'
Johannes> endif
Tha
> "Akim" == Akim Demaille <[EMAIL PROTECTED]> writes:
Akim> I've read into more details your proposal, and it looks very good,
Akim> much more robust than the current approach in Automake IIRC.
How so?
Akim> BTW, does Automake *use* that feature?
Yes. It is critical for dependency trackin
> "Ralf" == Ralf Corsepius <[EMAIL PROTECTED]> writes:
Ralf> The actual reason for my report is that I am experiencing more
Ralf> severe bugs related to depcomp handling:
Ralf> - Sometimes depcomp does not get installed with automake -a.
I think that currently automake will install depcomp
> "Ralf" == Ralf Corsepius <[EMAIL PROTECTED]> writes:
Ralf> Something very fishy seems to be going on with automake-cvs and
Ralf> autoconf-cvs (both updated yesterday):
Ralf> /opt/dev/share/automake/am/depend2.am: AMDEP does not appear in
Ralf> AM_CONDITIONAL
I tried your example but was u
> "Richard" == Richard Boulton <[EMAIL PROTECTED]> writes:
Richard>2001-06-06 Richard Boulton <[EMAIL PROTECTED]>
Richard>
Richard>* (conditionals_true_when): Pass first parameters by reference,
Richard>avoiding bug which put all parameters in @CONDS instead of @WHENS.
Richa
> "Bruce" == Bruce Korb <[EMAIL PROTECTED]> writes:
Bruce> AC_DEFUN(AG_PROG_BYACC,
Bruce> [missing_dir=ifelse([$1],,`cd $ac_aux_dir && pwd`, $1)
Bruce> AC_CHECK_PROGS(YACC, byacc yacc 'bison -y', "$missing_dir/missing yacc")])
This ought to work fine. In particular $AC_CHECK_PATTERN (look d
> "Bruce" == Bruce Korb <[EMAIL PROTECTED]> writes:
Bruce> Why is this now necessary? It did not used to be. Is this
Bruce> intended? Is there an "upgrading README" file kicking around?
It is probably a newly introduced bug.
You don't mention what version you are using though.
Could you s
> "Bruce" == Bruce Korb <[EMAIL PROTECTED]> writes:
Bruce> What is this? Maybe not everyone wants to use your version of
Bruce> AC_PROG_YACC. I use my own: AG_PROG_BYACC because I have a
Bruce> strong preference for BYACC. But, your latest automake is
Bruce> making some sort of attempt to
Last night I wrote a history of the automatic dependency schemes used
in Automake. I've checked it in to the Automake web pages. Read it
here, if you're interested:
http://sources.redhat.com/automake/dependencies.html
Teaser: it turns out that every dependency tracking scheme (not only
inc
Today I wrote a dumb little script to estimate how often automake is
used. The script searches a set of SRPM files for `configure.in' and
`Makefile.am' files, and generates a report showing how many packages
use autoconf and how many use automake.
Then I ran this script on all the SRPMs in Red H
Richard --
I got notification today that your copyright assignment has come in.
So I hope to look at this patch pretty soon.
If I haven't replied about it in the next week or so, please feel free
to remind me.
Thanks,
Tom
> "Clark" == Clark Rawlins <[EMAIL PROTECTED]> writes:
Clark> The subject says it all.
Automake 1.4f should work with autoconf 2.13 or 2.50.
Tom
> "Johannes" == Kremp, Johannes (Extern) <[EMAIL PROTECTED]> writes:
Johannes> sorry for this question, but where can i get cvs automake?
Start at the home page.
http://sources.redhat.com/automake/
Tom
> "Alexandre" == Alexandre Oliva <[EMAIL PROTECTED]> writes:
Richard> With separate Makefile.am's in each directory,
Richard> automake should be able to figure the bar/foo out from
Richard> the directory paths. The user shouldn't have to worry
Richard> about what the path to the top-level is
> "adl" == Alexandre Duret-Lutz <[EMAIL PROTECTED]> writes:
Tom> The idea would be to read all the Makefile.am
Tom> at once and then generate a single large Makefile.in.
adl> There is something nice about having one Makefile.am in each
adl> subdirectory, it's that it helps to make selfcontai
Lately I've been thinking that it was a mistake to introduce
_LTLIBRARIES as a parallel primary to _LIBRARIES. I think we could
just have _LIBRARIES and then base decisions on the extension. This
would also let us use the same mechanism for Java jar and zip files.
This is of course a post-1.5 t
> "Johannes" == Kremp, Johannes (Extern) <[EMAIL PROTECTED]> writes:
What version of automake are you using?
Please include version info in every bug report.
Johannes> and in my 'Makefile.am': if HELLO
Johannes> ...
Johannes>
> "Johannes" == Kremp, Johannes (Extern) <[EMAIL PROTECTED]> writes:
Johannes> hi, is there a patch which make it possible to make
Johannes> dependency-tracking with other compiler than 'gcc'?
This is one of the major new features in the cvs automake.
Tom
> "Emil" == Emil Ong <[EMAIL PROTECTED]> writes:
Emil> I'm sorry, I'm using the latest CVS. I just updated this
Emil> morning. I'm still getting the same error (File format not
Emil> recognized) with install-strip.
Ok, there's definitely something wrong here.
But without a test case I don'
> "Alexandre" == Alexandre Oliva <[EMAIL PROTECTED]> writes:
>> I was thinking I could create a new `patch' category in the automake
>> Gnats and ask people to put such patches there as attachments. What
>> do people think of this?
Alexandre> You can't reliably extract patches from Gnats.
> "Paolo" == Paolo Bonzini <[EMAIL PROTECTED]> writes:
Paolo> If not, I have done the necessary work and can
Paolo> prepare a patch (I'm not including it because I
Paolo> had the great idea of working directly on the
Paolo> installed scripts and data files, so preparing the
Paolo> patch might
> "Emil" == Emil Ong <[EMAIL PROTECTED]> writes:
Emil> In a project that I'm working on, I have both scripts and
Emil> executables. The scripts are Bourne shell and awk and I tell
Emil> automake about them with bin_SCRIPTS. When I do a make
Emil> install-strip, the Makefile also tries to st
> "David" == David Petrou <[EMAIL PROTECTED]> writes:
David> Am I doing the right thing by specifying `LDADD =
David> ../src/.libs/libpdl.a'.
Nope. Doesn't it work to use `../src/libpdl.la'?
David> 2. I've noticed that there's no target to make an html version
David> of a texinfo file. I
> "Sreekant" == Sreedharan, Sreekant <[EMAIL PROTECTED]> writes:
Sreekant>\os
Sreekant>\os\portable
Sreekant>\os\unix
Sreekant> How do I do it? Automake does not allow source code to
Sreekant> reside in a different directory other than the one in which
Sreekant> you h
> "Peter" == Peter Eisentraut <[EMAIL PROTECTED]> writes:
Peter> make install INSTALL_PROGRAM='install -m 000' INSTALL_DATA='install -m 123' #
etc.
This is probably the only sort of thing we'll support in automake.
I imagine you can set these variables in your config.site. Then any
config
> "Stephen" == Stephen Cameron <[EMAIL PROTECTED]> writes:
Tom> I want to know if this happens because argmatch.o already exists,
Tom> or if it is because the SCO `cc' doesn't like `-c' and `-o'
Tom> together.
Stephen> It appears to be the latter:
We should already have the mechanisms we ne
> "Kevin" == Kevin Dalley <[EMAIL PROTECTED]> writes:
Kevin> Here's a slightly revised test. I removed the Makefile.am
Kevin> modification. libtool is still built by configure, but it is
Kevin> never removed. For a real life application, the libraries
Kevin> might appear in a subdirectory,
> "Kevin" == Kevin Dalley <[EMAIL PROTECTED]> writes:
Kevin> In a Makefile which is created using automake-1.4f, libtool is not
Kevin> removed by distclean-am, in fact, it doesn't seem to be removed at
Kevin> all. In automake-1.4-p4, libtool is removed by distclean-am. These
Kevin> are the
> "Derek" == Derek R Price <[EMAIL PROTECTED]> writes:
Derek> Sounds like this could be fixed in depcomp?
Maybe...
Stephen> I'm Trying to update an old SCO unixware 2 box...Currently I
Stephen> have CVS 1.10.8 there. (Yeah, shame on me)
Stephen> The cc's -o option refuses to overwrite exis
> "Lars" == Lars J Aas <[EMAIL PROTECTED]> writes:
Lars> It would be great if one could add -I options to aclocal by just placing
Lars> macros in configure.ac, like:
Lars> AC_CONFIG_MACRO_DIR(conf/m4macros)
That might be nice.
I forget though what Akim has in mind for this for autoconf.
The
> "Steve" == Steve M Robbins <[EMAIL PROTECTED]> writes:
Steve> The upshot is that "foo.c" is compiled using AM_CFLAGS but NOT
Steve> foo_CFLAGS, while "foo2.c" is compiled NOT using AM_CFLAGS but using
Steve> foo_CFLAGS.
I'm checking in a fix for this.
Tom
We're starting to get patches which can't be checked in until
post-1.5. I'm not even really interested in reviewing them until
then.
I was thinking I could create a new `patch' category in the automake
Gnats and ask people to put such patches there as attachments. What
do people think of this?
> "Akim" == Akim Demaille <[EMAIL PROTECTED]> writes:
> "Nathan" == Nathan Neulinger writes:
Nathan> Would y'all consider extending the license exception to
Nathan> include the 'missing' script as well? (I'm referring to the
Nathan> exception that allows distributing autoconf support fil
> "Kevin" == Kevin Dalley <[EMAIL PROTECTED]> writes:
Kevin> If AC_CONFIG_AUX_DIR is not used, and version.texi is in a doc
Kevin> directory, than "automake -a" installs mdate-sh in top_srcdir,
Kevin> but the Makefile.in in the doc directory tries to use mdate-sh
Kevin> in srcdir, which mdate
> "Rusty" == Rusty Ballinger <[EMAIL PROTECTED]> writes:
Rusty> Is there a simpler way to do this? It seems like you'd wind up
Rusty> doing a lot of work which automake already does.
Another method is to use a macro (available at the macro archive) to
expand $prefix or whatever at configure
> "Richard" == Richard Boulton <[EMAIL PROTECTED]> writes:
Richard> i) variable_conditions_reduce() currently checks conditions
Richard> for redundancy against only those conditions which it has
Richard> already checked, rather than against all the conditions it
Richard> hasn't yet discarded.
> "Steve" == Steve M Robbins <[EMAIL PROTECTED]> writes:
Steve> Given the following Makefile.am
Steve> bin_PROGRAMS = foo
Steve> foo_SOURCES = foo.c sub/foo2.c
Steve> it turns out that "foo.c" is compiled using $(COMPILE), and hence
Steve> includes AM_CFLAGS. However, "foo2.c" is compiled
> "Paul" == Paul D Smith <[EMAIL PROTECTED]> writes:
Paul> Without GPATH, new files checked out will be placed in the target
Paul> directory; with it they will be updated in the source directory.
We could add that, but we couldn't rely on it.
Well, maybe we could, since really we only suppor
> "Tim" == Tim Van Holder <[EMAIL PROTECTED]> writes:
Tim> BTW, what is the correct way (if there is any) to do this in
Tim> stock 1.4? It's just that this is supposed to work in a
Tim> distributed package, so relying on CVS automake is probably not a
Tim> good idea.
Well, you could fight wi
> "Tim" == Tim Van Holder <[EMAIL PROTECTED]> writes:
Tim> I need to conditionally (based on a --with configure option)
Tim> add a fairly large number (~50) of sources to foo_SOURCES.
What version of automake are you using?
Tim> OPTIONAL=
Tim> if MYCONDITIONAL
Tim> OPTIONAL = lotsasource.c
> ">" == marq <[EMAIL PROTECTED]> writes:
>> What the attached patch does is simply to register new languages:
>> fortran 90, fortran 95, and their preprocessed versions. The main
>> rationale behind this is:
Thanks!
This won't go into automake 1.5. We're trying to get ready to
release.
> "Stéphane" == Stéphane Genaud <[EMAIL PROTECTED]> writes:
Stéphane> modnam.def : ttimes.f
Stéphane> echo " data modnam/'$(datadir)'" > modnam.def
Stéphane> and i need this file to be written before the ttimes.f compiles.
You have a few choices.
One is this:
BUILT_SOURCES
> "Hair" == Raja R Harinath <[EMAIL PROTECTED]> writes:
Clark> test_SOURCES = d1/s1.cc d2/s2.cc
Hair> Try using
Hair> AUTOMAKE_OPTIONS = subdir-objects
subdir-objects will cause Makefile to build d1/s1.o and d2/s2.o.
Without it, the above test_SOURCES will still work but the .o files
will
> "Clark" == Clark Rawlins <[EMAIL PROTECTED]> writes:
Clark> Is there any kind of concensus on what would be the 'correct'
Clark> way to do this in automake? Is there a consensus that automake
Clark> should continue to support these old compilers?
Yes, automake should support these old com
> "Kevin" == Kevin Dalley <[EMAIL PROTECTED]> writes:
Kevin> In automake-1.4f, in a directory using dejagnu, I ended up with a
Kevin> if %
Kevin> in the middle of Makefile.in
Thanks. I checked in your patch.
Tom
> "Richard" == Richard Boulton <[EMAIL PROTECTED]> writes:
Richard> You would lose the ability to build only sources in a
Richard> directory and its subdirs while ignoring files they depended
Richard> upon elsewhere in the tree which are out of date, but I'd
Richard> consider that a feature r
I think right now automake should have enough machinery that you could
write a single Makefile.am for an entire project. This ought to have
some performance benefits for the build. Has anybody tried this in a
serious way? I'm curious to hear about experiences.
Last night I gave some thought t
> "Richard" == Richard Boulton <[EMAIL PROTECTED]> writes:
Richard> i) Is there any documentation of this yet? I couldn't find
Richard> any, and simply had to guess what to do from error messages
Richard> and sources.
There isn't. I'm working on documentation. Thanks for mentioning
thi
> "Gary" == Gary V Vaughan <[EMAIL PROTECTED]> writes:
Gary> Don't forget that 1.4-p2 is simply a branch of the 2 year old
Gary> 1.4 with a small number of commonly reported bugs fixed. If the
Gary> fault you are seeing was fixed in mainline CVS after 1.4 was
Gary> released, you could backpo
> "Tim" == Tim Van Holder <[EMAIL PROTECTED]> writes:
Tim> Then again, you could simply mention this problem with missing in
Tim> the docs and tell developers to configure using MAKEINFO=makeinfo
Tim> if they intend to change the docs; that way, missing wouldn't be
Tim> run and the problem wo
> "Harlan" == Harlan Stenn <[EMAIL PROTECTED]> writes:
Harlan> I have an automake/autoconf application where the auto*
Harlan> generated stuff has to co-exist with a different "make"
Harlan> application, and cannot use [Mm]akefile for the names of the
Harlan> generated targets.
Harlan> I fig
> "Steven" == Steven G Johnson <[EMAIL PROTECTED]> writes:
>> You can even arrange to have the objects kept in subdirectories using:
>> AUTOMAKE_OPTIONS = subdir-objects
>> This option isn't really documented yet, but it is on my to-do list.
Steven> I assume you mean that bar/baz.c is normal
> "Steven" == Steven G Johnson <[EMAIL PROTECTED]> writes:
Steven> In src/Makefile.am, putting lib_foo_la_SOURCES = bar/baz.c,
Steven> etcetera, does not seem to be supported. Is there, or will
Steven> there be at some point, a proper way of going about this sort
Steven> of thing? Or at lea
> "Alexandre" == Alexandre Oliva <[EMAIL PROTECTED]> writes:
>> I've noticed that if you try to build a texi file that has an error
>> you get a warning from the `missing' program:
Alexandre> Presumably missing should check, when the execution of a
Alexandre> program fails, whether the progr
I've noticed that if you try to build a texi file that has an error
you get a warning from the `missing' program:
cd ../automake \
&& /bin/sh /x2/egcs-stuff/automake/automake/lib/missing --run makeinfo `echo
../automake/automake.texi | sed 's,.*/,,'`
./automake.texi:1675: Next field of node `A
> "Kevin" == Kevin Ryde <[EMAIL PROTECTED]> writes:
Kevin> EXTRA_HEADERS = foo.h
Kevin> don't get into a "make dist". Is this meant to work?
I don't think this was ever done. It is a bit confusing, I'll admit.
Kevin> It did in recent cvs versions but I guess noinst_HEADERS might
K
> "Kevin" == Kevin Ryde <[EMAIL PROTECTED]> writes:
Kevin> I think it'd be good if automake checked its support files like
Kevin> "missing" and "ansi2knr.c" are up-to-date, not just that they exist.
I agree, but I don't think we should put this in 1.5.
Could you submit a PR for this?
Kevin>
> "Attila" == Attila Csosz <[EMAIL PROTECTED]> writes:
Attila> I'm successfully generated system dependent dependency info in
Attila> the ".dep" directorys with the following commands:
Attila> 'automake;configure;make', after extracting the packages. How
Attila> could I get 'Makefile.in' file
> "Maciej" == Maciej W Rozycki <[EMAIL PROTECTED]> writes:
Maciej> This fixes the configure-time error, indeed, but is it
Maciej> correct? I believe dependencies should only be controlled by
Maciej> the "no-dependencies" setting in AUTOMAKE_OPTIONS and thus my
Maciej> fix is correct and your
> "Maciej" == Maciej W Rozycki <[EMAIL PROTECTED]> writes:
Maciej> INSTALL_PROGRAM='${INSTALL} -s' ./configure
Maciej> make
Maciej> make install
Maciej> which I actually use for plain autoconf packages, which have
Maciej> no idea of install-strip. Still, it's more like a workaround
Maciej>
> "Alexandre" == Alexandre Oliva <[EMAIL PROTECTED]> writes:
>> Note that I'm writing of a performance. Install-sh is a serious
>> performance hit for non-trivial installs.
Alexandre> How about only use install-sh for install-strip on cross builds?
I'm ok with this idea. Can it be easily
> "Axel" == Axel Thimm <[EMAIL PROTECTED]> writes:
Axel> The request was for experience in making the
Axel> autoconf/automake/libtool set work with Borland C++.
Doing this would be fine from the automake perspective.
If the compiler is too different from a Unix compiler it might be
easiest t
> "Derek" == Derek R Price <[EMAIL PROTECTED]> writes:
Derek> I would guess that a similar exception for mdate-sh would be
Derek> acceptable if others thought it a good idea. I don't know much
Derek> about it, but glancing through the code seems to confirm that
Derek> an exception wouldn't h
> "Harlan" == Harlan Stenn <[EMAIL PROTECTED]> writes:
Akim> Could the fact that we are using a macro for the dependency be
Akim> responsible? Harlan, what happens if you replace the macro
Akim> invocation with its value?
Harlan> This has something to do with it.
I'm satisfied that I under
> "Lars" == Lars J Aas <[EMAIL PROTECTED]> writes:
Lars> The "comment following trailing backslash" warning ought not to
Lars> be given if the backslash is already inside a comment. Just my
Lars> .02 cents.
Alexandre Duret-Lutz supplied a patch for this, and I just checked it
in.
Tom
only when DEP_FILES is
Maciej> non-empty.
I think this patch fixed the problem:
2001-05-17 Tom Tromey <[EMAIL PROTECTED]>
* m4/depend.m4 (AM_DEPENDENCIES): If depcomp doesn't exist, revert
to no dependency tracking.
I'm guessing you forgot to run aclocal.
Tom
> "Maciej" == Maciej W Rozycki <[EMAIL PROTECTED]> writes:
Maciej> I'm doing cross-compiles regularly and I've never had a
Maciej> problem with `make install-strip' using install from
Maciej> fileutils. The point is cross-tools, strip included, get
Maciej> installed in ${prefix}/${target_al
> "Kevin" == Kevin Ryde <[EMAIL PROTECTED]> writes:
Kevin> I think it should create and distribute the info files
Kevin> unconditionally, but only install (or uninstall) them under
Kevin> WANT_FOO.
>> I'd prefer not to add a special case like this.
Kevin> Oh, is distributing them unconditio
> "Pavel" == Pavel Roskin <[EMAIL PROTECTED]> writes:
Pavel> I don't know whether Autoconf should be more robust or Automake
Pavel> should take less (or more?) hackerish approach.
Probably the former.
Pavel> Since Autoconf-2.50 has been released, it would be fair to drop
Pavel> support for
> "Peter" == Peter Eisentraut <[EMAIL PROTECTED]> writes:
Peter> Make config.status put all the configuration information into a
Peter> single makefile and have all the other makefiles include that
Peter> one. It's saved me many boring waits.
That's an interesting idea. We could even imple
Well, here's why the ansi2knr stuff didn't work.
Curious that I didn't remember this. Or run across it before.
I think it must be because we've moved a lot more of the variables
into the *.am files. In the past we generated a lot of variables in
automake itself. Automake put all the variables
> "Kevin" == Kevin Ryde <[EMAIL PROTECTED]> writes:
Kevin> Would it be easy to let info_TEXINFOS be conditionally defined?
Kevin> I think it should create and distribute the info files
Kevin> unconditionally, but only install (or uninstall) them under
Kevin> WANT_FOO.
I'd prefer not to add
> "Rasmus" == Rasmus Tamstorf <[EMAIL PROTECTED]> writes:
Rasmus> I think the issue is that with autoconf you setup your
Rasmus> compiler flags (debug / optimized etc) for the entire build
Rasmus> tree when you run ./configure. However, while developing you
Rasmus> may want 90% to be optimize
> "Harlan" == Harlan Stenn <[EMAIL PROTECTED]> writes:
Harlan> While (IMO) it "did better" than autoconf for quite a few
Harlan> years, I bailed on it several years ago because it was not
Harlan> suitable (at that time) for srcdir != builddir, and that was
Harlan> something I needed Badly. T
> "Guido" == Guido Draheim <[EMAIL PROTECTED]> writes:
Guido> alternatives? I mean free, documented, mature, easy
Guido> alternatives to be seen?
I'm unaware of any. There is Metaconf (the Perl thing), but in my
experience it isn't deeply better than autoconf. It is different, but
less wel
301 - 400 of 1294 matches
Mail list logo