Hi Lennart: Could you have some comments on this patch Zhang Xin(Wing) Intel(SSG/OTC) ShangHai China >-----Original Message----- >From: pulseaudio-discuss-bounces at mail.0pointer.de [mailto:pulseaudio- >discuss-bounces at mail.0pointer.de] On Behalf Of Zhang, Xing Z >Sent: 2009?4?2? 18:15 >To: General PulseAudio Discussion >Subject: Re: [pulseaudio-discuss] Export pulsecore and pulsecommon header >file > >Hi Lennart: > We will certainly hand in our patches of module to upstream finally. >But the work is still in processing so I want to do this when it is >finalized. >Before that I would like to submit attached patch for optional exporting >internal headers. > I test it in last changeset(3294c89adba34d186591d98faa3a1f929b363e59) >it works fine. > >Zhang Xin(Wing) >Intel(SSG/OTC) ShangHai China > >>-----Original Message----- >>From: pulseaudio-discuss-bounces at mail.0pointer.de [mailto:pulseaudio- >>discuss-bounces at mail.0pointer.de] On Behalf Of Lennart Poettering >>Sent: 2009?3?26? 21:43 >>To: pulseaudio-discuss at mail.0pointer.de >>Subject: Re: [pulseaudio-discuss] Export pulsecore and pulsecommon header >>file >> >>On Wed, 25.03.09 18:43, Zhang, Xing Z (xing.z.zhang at intel.com) wrote: >> >>> Hi Lennart: >> >>> More and more people like me are working on writing module of PA. A >>> problem we encountered is current PA dev-pkg released by >>> distribution vendor only includes header files for application but >>> not for module. >>> >>> Did you think about export pulsecore/pulsecommon header files to >>> module programmers so that modules can be developed with dev pkg >>> such as pulseaudio-devel-0.9.15-10.1.fedaro.i586.rpm. I tried to do >>> this by attached patch but accomplished nothing due to config.h >>> produced by autoconf can not be exported. >>> >>> Do you have any suggestion to pass this issue? It will really >>> benefit module programmers much if the process of writing a PA >>> module is as easy as writing an application in distribution. >> >>At this point the PA module API is not declared stable, and I am not >>sure it will be anytime soon. Thus maintaining modules outside of the >>tree is pain. If you develop your own modules the best way is probably >>keeping things in a git clone of the upstream repo and then rebasing >>from time to time. Also, I encourage everyone to submit your patches >>upstream. If the code is good I am happy to merge even modules that >>are not of any direct use to myself or anyone else besides your >>company. If you submit things upstream this has the big advantage that >>I will look over them from time to time and might even update them >>when I change the module API in some way. >> >>I am sorry, but I think allowing the core headers to be installed >>would send the wrong signal to developers: it suggests the API would >>be stable -- which it is not. >> >>We could cut a compromise however: install those header files only if >>a specific configure switch >>--enable-core-headers-i-know-this-is-not-stable or so. To accomplish >>that we'd however need some non-trivial changes to fix the config.h >>situation. That could be done the same way how glib generates >>/usr/lib/glib-2.0/include/glibconfig.h. Patches welcome. >> >>Lennart >> >>-- >>Lennart Poettering Red Hat, Inc. >>lennart [at] poettering [dot] net ICQ# 11060553 >>http://0pointer.net/lennart/ GnuPG 0x1A015CC4 >>_______________________________________________ >>pulseaudio-discuss mailing list >>pulseaudio-discuss at mail.0pointer.de >>https://tango.0pointer.de/mailman/listinfo/pulseaudio-discuss
