[Mesa-dev] [PATCH] mapi: fix out-of-tree build dependencies

Nicolai Hähnle nhaehnle at gmail.com
Tue Oct 11 16:13:45 UTC 2016


On 11.10.2016 16:26, Emil Velikov wrote:
> On 11 October 2016 at 14:46, Nicolai Hähnle <nhaehnle at gmail.com> wrote:
>> From: Nicolai Hähnle <nicolai.haehnle at amd.com>
>>
>> Pick up changes to API XML files also when Mesa is built outside the source
>> tree.
>> ---
>>  src/mapi/Makefile.am | 4 ++--
>>  1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/src/mapi/Makefile.am b/src/mapi/Makefile.am
>> index d6bf5d8..46afe3b 100644
>> --- a/src/mapi/Makefile.am
>> +++ b/src/mapi/Makefile.am
>> @@ -49,22 +49,22 @@ AM_CPPFLAGS =                                                       \
>>         -I$(top_srcdir)/src/mapi                                \
>>         -I$(top_builddir)/src/mapi
>>
>>  include Makefile.sources
>>
>>  MKDIR_GEN = $(AM_V_at)$(MKDIR_P) $(@D)
>>  PYTHON_GEN = $(AM_V_GEN)$(PYTHON2) $(PYTHON_FLAGS)
>>
>>  glapi_gen_mapi_deps := \
>>         mapi_abi.py \
>> -       $(wildcard glapi/gen/*.xml) \
>> -       $(wildcard glapi/gen/*.py)
>> +       $(wildcard $(top_srcdir)/src/mapi/glapi/gen/*.xml) \
>> +       $(wildcard $(top_srcdir)/src/mapi/glapi/gen/*.py)
>>
> Hmm are you sure about this ? AFAICT by default you get $srcdir which
> combined with glapi/gen/foo should work just fine... or is it the
> wildcard that's causing the mayhem ? Please add a word or two about it
> in the commit message.

Well, I'm really not an expert on make intricacies of this level, but 
`make -p` confirms that the wildcards didn't match anything previously 
(glapi_gen_mapi_deps is set to only mapi_abi.py and nothing else), and 
with the explicit $(top_srcdir) they match as they should. Perhaps this 
should be considered a bug in make, but it is what it is. I can add a 
comment about this to the commit.

Cheers,
Nicolai

>
> In the long run we _really_ want to drop our dispatch and use the
> GLVND one. The GLVND maintainer isn't buying it so... we might have to
> keep it forever. Sigh...
>
> -Emil
>


More information about the mesa-dev mailing list