You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Max Bowsher <ma...@ukf.net> on 2005/10/04 11:49:04 UTC
Concern about current swig includes dependency info generation
The current swig build uses a number of swig include files which are
generated, not shipped.
An unfortunate consequence of this is that the generated build-outputs.mk
contains *different* dependencies depending on whether gen-make.py has been
run before or after a "make swig-somelanguage".
If those dependencies are truly needed, they need to be generated before a
swig build has occurred, so that they make it into the build-outputs.mk
shipped in our release tarballs.
If those dependencies are not truly needed, they should not be output at
all, to avoid creating unnecessary differences between developers' working
copies and release builds.
Max.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Concern about current swig includes dependency info generation
Posted by Max Bowsher <ma...@ukf.net>.
David James wrote:
> On 10/4/05, Max Bowsher <ma...@ukf.net> wrote:
>> The current swig build uses a number of swig include files which are
>> generated, not shipped.
>>
>> An unfortunate consequence of this is that the generated build-outputs.mk
>> contains *different* dependencies depending on whether gen-make.py has
>> been
>> run before or after a "make swig-somelanguage".
>>
>> If those dependencies are truly needed, they need to be generated before
>> a
>> swig build has occurred, so that they make it into the build-outputs.mk
>> shipped in our release tarballs.
>>
>> If those dependencies are not truly needed, they should not be output at
>> all, to avoid creating unnecessary differences between developers'
>> working
>> copies and release builds.
> Nice catch, Max! This problem should be fixed in r16446. I very much
> appreciate your careful work on the SWIG bindings -- it makes a big
> difference!
This is a step in the right direction, I think, but IIUC, there is still a
flaw - the list of *_h.swg files is still constructed by globbing - but when
the glob is evaluated, the files do not exist yet.
Max.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Concern about current swig includes dependency info generation
Posted by David James <ja...@gmail.com>.
On 10/4/05, Max Bowsher <ma...@ukf.net> wrote:
> The current swig build uses a number of swig include files which are
> generated, not shipped.
>
> An unfortunate consequence of this is that the generated build-outputs.mk
> contains *different* dependencies depending on whether gen-make.py has been
> run before or after a "make swig-somelanguage".
>
> If those dependencies are truly needed, they need to be generated before a
> swig build has occurred, so that they make it into the build-outputs.mk
> shipped in our release tarballs.
>
> If those dependencies are not truly needed, they should not be output at
> all, to avoid creating unnecessary differences between developers' working
> copies and release builds.
Nice catch, Max! This problem should be fixed in r16446. I very much
appreciate your careful work on the SWIG bindings -- it makes a big
difference!
Cheers,
David
--
David James -- http://www.cs.toronto.edu/~james
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org