Re: [PATCH] Update gnulib/Makefile.in:aclocal_m4_deps

[Sergio Durigan Junior <sergiodj@redhat.com>]

On Friday, August 31 2018, Pedro Alves wrote:

> On 08/30/2018 09:00 PM, Sergio Durigan Junior wrote:
>> On Thursday, August 30 2018, Simon Marchi wrote:
>> 
>>> On 2018-08-30 11:57, Sergio Durigan Junior wrote:
>
>>>>  $(srcdir)/aclocal.m4: @MAINTAINER_MODE_TRUE@ $(aclocal_m4_deps)
>>>>  	cd $(srcdir) && $(ACLOCAL) $(ACLOCAL_AMFLAGS)
>>>
>>> This looks good to me, the set of m4 files listed here matches the
>>> list of actual m4 files in import/m4.  Follow-up question, could we
>>> use $(wildcard ...) instead of listing them by hand?
>> 
>> Yeah, I think that could work.  That's basically what I did to generate
>> this list: "ls *.m4".  If you want, I can edit the patch and make it use
>> $(wildcard) before pushing it.
>> 
>
> Really not sure that's a good idea.  We don't use $wildcard for listing .c files
> either, for example, and I think for good reason.  It makes the set of files to build
> dependent of what you happen to have or not have locally, instead of determined
> statically.  That in general affects development, changing git branches, etc. 
> Consider that GDB even links successfully if you miss including/linking some .c file
> in the build, given the _initialize_foo registration mechanism.

Hm, alright, that's a good point.

> Maybe not so much an issue with the m4 files, but I'd think a more
> principle approach to automate this would be to make the
> update-gnulib.sh script generate/update a Makefile fragment file that
> contained the aclocal_m4_deps m4 files list, check that file into the
> tree, and then gdb/gnulib/Makefile.in would source/include that
> fragment file.

Yeah, that works as well.  I'll work on a patch and submit it soon.

Thanks,

-- 
Sergio
GPG key ID: 237A 54B1 0287 28BF 00EF  31F4 D0EB 7628 65FC 5E36
Please send encrypted e-mail if possible
http://sergiodj.net/