Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module
Hi, Please review the attached patch. Change Description: 1. Pulse Audio Remix Plug-in added as audio output device. 2. Support for Multisense initialization/shutdown only when multisense feature is used. 3. Enhancement to _edje_multisense_shutdown to fix memory leak. 4. Multisense example enhancement, correction of macros(ENABLE_MULTISENSE HAVE_REMIX) used and unnecessary statements are removed in edje_multisense.c. 5. Memory Cleanup and pipe closing is done in worker thread and thread end/cancel callbacks are removed. 6. Interface multisense_factory_shutdown implementation, Macro WRN implementation using eina_log. 7. Fix for buffer underrun in ALSA plugin and REMIX flags are used in src/bin/Makefile.am. 8. Multisense support enabled by default. Signed-Off-By: PRINCE KUMAR DUBEY prince.du...@samsung.com Thank you, Prince -- From: Flávio Ceolin flavio.ceo...@profusion.mobi Sent: Tuesday, July 10, 2012 3:56 AM To: Enlightenment developer list enlightenment-devel@lists.sourceforge.net Cc: PRINCE KUMAR DUBEY prince.du...@samsung.com Subject: Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module Hi Prince, On Mon, Jul 9, 2012 at 3:50 AM, Carsten Haitzler ras...@rasterman.com wrote: On Fri, 06 Jul 2012 14:59:52 + (GMT) PRINCE KUMAR DUBEY prince.du...@samsung.com said: Hi, -- From: Carsten Haitzler (The Rasterman) [1]ras...@rasterman.com Sent: Friday, July 06, 2012 1:38 PM To: Enlightenment developer list [2]enlightenment-devel@lists.sourceforge.net Cc: PRINCE KUMAR DUBEY [3]prince.du...@samsung.com Subject: Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module On Mon, 02 Jul 2012 07:07:19 + (GMT) PRINCE KUMAR DUBEY [4]prince.du...@samsung.com said: these are really suspicious. you're messing with starting/stopping mainloops here + ecore_main_loop_quit(); +//again main loop begins to process thread cancellation +ecore_main_loop_begin(); :( not good. imagine u shut down edje while still in the mainloop. :) (not relevant for elementary but for things using raw edje+ecore+evas/ etc. ...) Anyway in ecore_thread_cancel callback extra ecore_main_loop is being quit, So will it still be problem! Reason for using ecore_main_loop_begin is that ecore_thread_cancel/end callbacks are not getting called in _edje_multisense_shutdown (for memory cleanup purpose). So, begining the main loop again to cleanup the memory in ecore_thread_cancel. I have altrenate approach to free the memory, when control comes out from while loop in worker thread, is this approach is ok ? alternative approach definitely... not the ecore mainloop begin and quit thing :) -- - Codito, ergo sum - I code, therefore I am -- The Rasterman (Carsten Haitzler)ras...@rasterman.com -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel Besides the problem with the main_loop pointed out by raster, i've found some small issues in the patch. I've attached your patch with the review. Overall, the patch is working properly :) Regards, Flavio Ceolin edje_multisense_pulseaudio_plugin.patch Description: Binary data -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module
Hi Prince, On Tue, Jul 10, 2012 at 10:13 AM, Prince Kr Dubey prince.du...@samsung.com wrote: Hi, -- From: Flávio Ceolin flavio.ceo...@profusion.mobi Sent: Tuesday, July 10, 2012 3:56 AM To: Enlightenment developer list enlightenment-devel@lists.sourceforge.net Cc: PRINCE KUMAR DUBEY prince.du...@samsung.com Subject: Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module Hi Prince, On Mon, Jul 9, 2012 at 3:50 AM, Carsten Haitzler ras...@rasterman.com wrote: On Fri, 06 Jul 2012 14:59:52 + (GMT) PRINCE KUMAR DUBEY prince.du...@samsung.com said: Hi, -- From: Carsten Haitzler (The Rasterman) [1]ras...@rasterman.com Sent: Friday, July 06, 2012 1:38 PM To: Enlightenment developer list [2]enlightenment-devel@lists.sourceforge.net Cc: PRINCE KUMAR DUBEY [3]prince.du...@samsung.com Subject: Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module On Mon, 02 Jul 2012 07:07:19 + (GMT) PRINCE KUMAR DUBEY [4]prince.du...@samsung.com said: these are really suspicious. you're messing with starting/stopping mainloops here + ecore_main_loop_quit(); +//again main loop begins to process thread cancellation +ecore_main_loop_begin(); :( not good. imagine u shut down edje while still in the mainloop. :) (not relevant for elementary but for things using raw edje+ecore+evas/ etc. ...) Anyway in ecore_thread_cancel callback extra ecore_main_loop is being quit, So will it still be problem! Reason for using ecore_main_loop_begin is that ecore_thread_cancel/end callbacks are not getting called in _edje_multisense_shutdown (for memory cleanup purpose). So, begining the main loop again to cleanup the memory in ecore_thread_cancel. I have altrenate approach to free the memory, when control comes out from while loop in worker thread, is this approach is ok ? alternative approach definitely... not the ecore mainloop begin and quit thing :) -- - Codito, ergo sum - I code, therefore I am -- The Rasterman (Carsten Haitzler)ras...@rasterman.com -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel Besides the problem with the main_loop pointed out by raster, i've found some small issues in the patch. I've attached your patch with the review. Overall, the patch is working properly :) Regards, Flavio Ceolin I'll check your comments and update the patch. Thanks, Prince -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel Thanks for the update ! Regards, Flavio Ceolin -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module
Hi, -- From: Flávio Ceolin flavio.ceo...@profusion.mobi Sent: Tuesday, July 10, 2012 3:56 AM To: Enlightenment developer list enlightenment-devel@lists.sourceforge.net Cc: PRINCE KUMAR DUBEY prince.du...@samsung.com Subject: Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module Hi Prince, On Mon, Jul 9, 2012 at 3:50 AM, Carsten Haitzler ras...@rasterman.com wrote: On Fri, 06 Jul 2012 14:59:52 + (GMT) PRINCE KUMAR DUBEY prince.du...@samsung.com said: Hi, -- From: Carsten Haitzler (The Rasterman) [1]ras...@rasterman.com Sent: Friday, July 06, 2012 1:38 PM To: Enlightenment developer list [2]enlightenment-devel@lists.sourceforge.net Cc: PRINCE KUMAR DUBEY [3]prince.du...@samsung.com Subject: Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module On Mon, 02 Jul 2012 07:07:19 + (GMT) PRINCE KUMAR DUBEY [4]prince.du...@samsung.com said: these are really suspicious. you're messing with starting/stopping mainloops here + ecore_main_loop_quit(); +//again main loop begins to process thread cancellation +ecore_main_loop_begin(); :( not good. imagine u shut down edje while still in the mainloop. :) (not relevant for elementary but for things using raw edje+ecore+evas/ etc. ...) Anyway in ecore_thread_cancel callback extra ecore_main_loop is being quit, So will it still be problem! Reason for using ecore_main_loop_begin is that ecore_thread_cancel/end callbacks are not getting called in _edje_multisense_shutdown (for memory cleanup purpose). So, begining the main loop again to cleanup the memory in ecore_thread_cancel. I have altrenate approach to free the memory, when control comes out from while loop in worker thread, is this approach is ok ? alternative approach definitely... not the ecore mainloop begin and quit thing :) -- - Codito, ergo sum - I code, therefore I am -- The Rasterman (Carsten Haitzler)ras...@rasterman.com -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel Besides the problem with the main_loop pointed out by raster, i've found some small issues in the patch. I've attached your patch with the review. Overall, the patch is working properly :) Regards, Flavio Ceolin I'll check your comments and update the patch. Thanks, Prince -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module
On Fri, 06 Jul 2012 14:59:52 + (GMT) PRINCE KUMAR DUBEY prince.du...@samsung.com said: Hi, -- From: Carsten Haitzler (The Rasterman) [1]ras...@rasterman.com Sent: Friday, July 06, 2012 1:38 PM To: Enlightenment developer list [2]enlightenment-devel@lists.sourceforge.net Cc: PRINCE KUMAR DUBEY [3]prince.du...@samsung.com Subject: Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module On Mon, 02 Jul 2012 07:07:19 + (GMT) PRINCE KUMAR DUBEY [4]prince.du...@samsung.com said: these are really suspicious. you're messing with starting/stopping mainloops here + ecore_main_loop_quit(); +//again main loop begins to process thread cancellation +ecore_main_loop_begin(); :( not good. imagine u shut down edje while still in the mainloop. :) (not relevant for elementary but for things using raw edje+ecore+evas/ etc. ...) Anyway in ecore_thread_cancel callback extra ecore_main_loop is being quit, So will it still be problem! Reason for using ecore_main_loop_begin is that ecore_thread_cancel/end callbacks are not getting called in _edje_multisense_shutdown (for memory cleanup purpose). So, begining the main loop again to cleanup the memory in ecore_thread_cancel. I have altrenate approach to free the memory, when control comes out from while loop in worker thread, is this approach is ok ? alternative approach definitely... not the ecore mainloop begin and quit thing :) -- - Codito, ergo sum - I code, therefore I am -- The Rasterman (Carsten Haitzler)ras...@rasterman.com -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module
Hi Prince, On Mon, Jul 9, 2012 at 3:50 AM, Carsten Haitzler ras...@rasterman.com wrote: On Fri, 06 Jul 2012 14:59:52 + (GMT) PRINCE KUMAR DUBEY prince.du...@samsung.com said: Hi, -- From: Carsten Haitzler (The Rasterman) [1]ras...@rasterman.com Sent: Friday, July 06, 2012 1:38 PM To: Enlightenment developer list [2]enlightenment-devel@lists.sourceforge.net Cc: PRINCE KUMAR DUBEY [3]prince.du...@samsung.com Subject: Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module On Mon, 02 Jul 2012 07:07:19 + (GMT) PRINCE KUMAR DUBEY [4]prince.du...@samsung.com said: these are really suspicious. you're messing with starting/stopping mainloops here + ecore_main_loop_quit(); +//again main loop begins to process thread cancellation +ecore_main_loop_begin(); :( not good. imagine u shut down edje while still in the mainloop. :) (not relevant for elementary but for things using raw edje+ecore+evas/ etc. ...) Anyway in ecore_thread_cancel callback extra ecore_main_loop is being quit, So will it still be problem! Reason for using ecore_main_loop_begin is that ecore_thread_cancel/end callbacks are not getting called in _edje_multisense_shutdown (for memory cleanup purpose). So, begining the main loop again to cleanup the memory in ecore_thread_cancel. I have altrenate approach to free the memory, when control comes out from while loop in worker thread, is this approach is ok ? alternative approach definitely... not the ecore mainloop begin and quit thing :) -- - Codito, ergo sum - I code, therefore I am -- The Rasterman (Carsten Haitzler)ras...@rasterman.com -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel Besides the problem with the main_loop pointed out by raster, i've found some small issues in the patch. I've attached your patch with the review. Overall, the patch is working properly :) Regards, Flavio Ceolin edje_multisense_pulseaudio_plugin.patch.review Description: Binary data -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module
On Mon, 02 Jul 2012 07:07:19 + (GMT) PRINCE KUMAR DUBEY prince.du...@samsung.com said: these are really suspicious. you're messing with starting/stopping mainloops here + ecore_main_loop_quit(); +//again main loop begins to process thread cancellation +ecore_main_loop_begin(); :( not good. imagine u shut down edje while still in the mainloop. :) (not relevant for elementary but for things using raw edje+ecore+evas/ etc. ...) Hi, Please review the attached patch after fixing review comments. Change Description: 1. Support for Multisense initialization/shutdown only when multisense feature is used. 2. Enhancement to _edje_multisense_shutdown to fix memory leak. 3. Multisense example enhancement and correction of macros(ENABLE_MULTISENSE HAVE_REMIX) use in edje_multisense.c. 4. Interface multisense_factory_shutdown implementation, Macro WRN implementation using eina_log. 5. Fix for buffer underrun in ALSA plugin and REMIX flags are used in src/bin/Makefile.am. 6. Multisense support enabled by default. 7.Pulse Audio Remix Plug-in added as audio output device. Signed-Off-By: Prince Kr Dubeyprince.du...@samsung.com Thanks Regards, Prince -- From: Flavio Ceolin flavio.ceo...@profusion.mobi Sent: Tuesday, June 26, 2012 11:19 PM To: Enlightenment developer list enlightenment-devel@lists.sourceforge.net Subject: Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module Hi Prince, On Tue, Jun 26, 2012 at 11:47 AM, Prince Kr Dubey prince.du...@samsung.com wrote: Hi Flavio, I'll create another patch with the modification suggested by you and Gustavo. And will post the same to community ASAP. Thanks Regards, Prince On 06/25/2012 14:20:20 , Flavio Ceolin flavio.ceo...@profusion.mobi wrote: Hi Prince, Gustavo Sverzut Barbieri barbi...@profusion.mobi writes: On Thu, Jun 21, 2012 at 10:32 AM, PRINCE KUMAR DUBEY prince.du...@samsung.com wrote: Hi, Please find my query/comments inlined below. Ok, and could you use some email client that does proper reply? It's a mess and unusual to read your mails without the proper quote blocks. +EAPI extern Eina_Bool _on_edjecc; It seems to be not necessary. [Prince] this variable is required, in order to provide check to unnecessary multisense framework initialization during EDC compilation/decompilation mode. So, during EDC compilation, _on_edjecc=EINA_TRUE in edje_cc.c and hence, _edje_multisense_init() is being ignored in edje_main.c. No, never do this. One should have a global setter and maybe a getter to enable or disable features, or even better just load multisense if it's required. [Prince] Multisense load is not needed during compilation/decompilation, because it inculdes remix and its plug-in(i.e. eetsndreader, ALSA) load, which are specificaly required during runtime only. As per your suggestion above, should we add interfaces for setter/getter to enable this feature in edje ? Please share your opinion. I dislike having to manually enable or disable them. Just use it as last resource. Ideally the multisense would be disabled and whenever a group uses multisense, it would be enabled. You do not need to further disable the setup later, it's fine to keep it on. It's like a lazy-loaded singleton. As edje_cc will never edje_object_file_set() it will never hit the path to load multisense. Then we're fine for it and every other software that does not use it. Are you intending to do the modifications suggested ? Your patch with those modifications looks pretty good. Best Regards, Flavio Ceolin -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel Thanks !! Best Regards, Flavio Ceolin -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists
Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module
Hi Flavio, I'll create another patch with the modification suggested by you and Gustavo. And will post the same to community ASAP. Thanks Regards, Prince On 06/25/2012 14:20:20 , Flavio Ceolin flavio.ceo...@profusion.mobi wrote: Hi Prince, Gustavo Sverzut Barbieri barbi...@profusion.mobi writes: On Thu, Jun 21, 2012 at 10:32 AM, PRINCE KUMAR DUBEY prince.du...@samsung.com wrote: Hi, Please find my query/comments inlined below. Ok, and could you use some email client that does proper reply? It's a mess and unusual to read your mails without the proper quote blocks. +EAPI extern Eina_Bool _on_edjecc; It seems to be not necessary. [Prince] this variable is required, in order to provide check to unnecessary multisense framework initialization during EDC compilation/decompilation mode. So, during EDC compilation, _on_edjecc=EINA_TRUE in edje_cc.c and hence, _edje_multisense_init() is being ignored in edje_main.c. No, never do this. One should have a global setter and maybe a getter to enable or disable features, or even better just load multisense if it's required. [Prince] Multisense load is not needed during compilation/decompilation, because it inculdes remix and its plug-in(i.e. eetsndreader, ALSA) load, which are specificaly required during runtime only. As per your suggestion above, should we add interfaces for setter/getter to enable this feature in edje ? Please share your opinion. I dislike having to manually enable or disable them. Just use it as last resource. Ideally the multisense would be disabled and whenever a group uses multisense, it would be enabled. You do not need to further disable the setup later, it's fine to keep it on. It's like a lazy-loaded singleton. As edje_cc will never edje_object_file_set() it will never hit the path to load multisense. Then we're fine for it and every other software that does not use it. Are you intending to do the modifications suggested ? Your patch with those modifications looks pretty good. Best Regards, Flavio Ceolin -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module
Hi Prince, On Tue, Jun 26, 2012 at 11:47 AM, Prince Kr Dubey prince.du...@samsung.com wrote: Hi Flavio, I'll create another patch with the modification suggested by you and Gustavo. And will post the same to community ASAP. Thanks Regards, Prince On 06/25/2012 14:20:20 , Flavio Ceolin flavio.ceo...@profusion.mobi wrote: Hi Prince, Gustavo Sverzut Barbieri barbi...@profusion.mobi writes: On Thu, Jun 21, 2012 at 10:32 AM, PRINCE KUMAR DUBEY prince.du...@samsung.com wrote: Hi, Please find my query/comments inlined below. Ok, and could you use some email client that does proper reply? It's a mess and unusual to read your mails without the proper quote blocks. +EAPI extern Eina_Bool _on_edjecc; It seems to be not necessary. [Prince] this variable is required, in order to provide check to unnecessary multisense framework initialization during EDC compilation/decompilation mode. So, during EDC compilation, _on_edjecc=EINA_TRUE in edje_cc.c and hence, _edje_multisense_init() is being ignored in edje_main.c. No, never do this. One should have a global setter and maybe a getter to enable or disable features, or even better just load multisense if it's required. [Prince] Multisense load is not needed during compilation/decompilation, because it inculdes remix and its plug-in(i.e. eetsndreader, ALSA) load, which are specificaly required during runtime only. As per your suggestion above, should we add interfaces for setter/getter to enable this feature in edje ? Please share your opinion. I dislike having to manually enable or disable them. Just use it as last resource. Ideally the multisense would be disabled and whenever a group uses multisense, it would be enabled. You do not need to further disable the setup later, it's fine to keep it on. It's like a lazy-loaded singleton. As edje_cc will never edje_object_file_set() it will never hit the path to load multisense. Then we're fine for it and every other software that does not use it. Are you intending to do the modifications suggested ? Your patch with those modifications looks pretty good. Best Regards, Flavio Ceolin -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel Thanks !! Best Regards, Flavio Ceolin -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module
Hi Prince, Gustavo Sverzut Barbieri barbi...@profusion.mobi writes: On Thu, Jun 21, 2012 at 10:32 AM, PRINCE KUMAR DUBEY prince.du...@samsung.com wrote: Hi, Please find my query/comments inlined below. Ok, and could you use some email client that does proper reply? It's a mess and unusual to read your mails without the proper quote blocks. +EAPI extern Eina_Bool _on_edjecc; It seems to be not necessary. [Prince] this variable is required, in order to provide check to unnecessary multisense framework initialization during EDC compilation/decompilation mode. So, during EDC compilation, _on_edjecc=EINA_TRUE in edje_cc.c and hence, _edje_multisense_init() is being ignored in edje_main.c. No, never do this. One should have a global setter and maybe a getter to enable or disable features, or even better just load multisense if it's required. [Prince] Multisense load is not needed during compilation/decompilation, because it inculdes remix and its plug-in(i.e. eetsndreader, ALSA) load, which are specificaly required during runtime only. As per your suggestion above, should we add interfaces for setter/getter to enable this feature in edje ? Please share your opinion. I dislike having to manually enable or disable them. Just use it as last resource. Ideally the multisense would be disabled and whenever a group uses multisense, it would be enabled. You do not need to further disable the setup later, it's fine to keep it on. It's like a lazy-loaded singleton. As edje_cc will never edje_object_file_set() it will never hit the path to load multisense. Then we're fine for it and every other software that does not use it. Are you intending to do the modifications suggested ? Your patch with those modifications looks pretty good. Best Regards, Flavio Ceolin -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module
Hi Flavio, Please find my reply inline below. Thanks Regards, Prince --- Original Message --- Date: Wed, 20 Jun 2012 15:40:14 -0300 From: Flavio Ceolin Subject: Re: [E-devel] enlightenment-devel Digest, Vol 74, Issue 78 To: Enlightenment developer list Message-ID: 877gv13jzl@voyager.profusion.mobi Content-Type: text/plain PRINCE KUMAR DUBEY writes: Hi Prince, Hi Flavio, I am still not able to reproduce the seg-fault. Is seg-fault happening always ? Yep, it's happening always i use the edje_player with the multisense example. The problem is: + length = count * sizeof(RemixCount); it makes length be greater than PA_PLAYER_BUFFERLEN then: +*(player-playbuffer + i) = (PLAYER_PCM) value; will access invalid memory. I rewrote this function and worked as expected. Please check if it makes sense. static void pa_player_playbuffer(RemixEnv *env __UNUSED__, PA_Player_Data *player, RemixPCM *data, RemixCount count) { int ret; RemixCount i, j; RemixPCM value; size_t length, total_written; length = count * sizeof(RemixPCM); total_written = 0; while (total_written length) { j = length - total_written; j = (j PA_PLAYER_BUFFERLEN) ? PA_PLAYER_BUFFERLEN : j; for (i = 0; i j; i++) { value = *data++ * (player-max_value); *(player-playbuffer + i) = (PLAYER_PCM) value; } ret = pa_simple_write(player-server, player-playbuffer, j, player-error); if (ret 0) { WRN(pa_simple_write() failed: (%s), pa_strerror(player-error)); return; } total_written += j; } } [Prince] make sense :), thanks. Please find my reply to your queries inline below. Thanks Regards, Prince --- Original Message --- Date: Mon, 18 Jun 2012 14:09:59 -0300 From: Fl?vio Ceolin Subject: Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module To: Enlightenment developer list Message-ID: Content-Type: text/plain; charset=ISO-8859-1 Hi Prince, On Mon, Jun 18, 2012 at 9:56 AM, PRINCE KUMAR DUBEY wrote: Hi Flavio, I tested edje_player multisense.edje to reproduce seg fault as mentioned below, but couldn't succeed with memcheck as well. My svn rev is 72374. Please let me know the scenario to reproduce it. And regarding the point 5, you are right, multisense will not build without remix as per current configure file. But, multisense is not all about only sound, it can have vibration etc. So, if we make remix optional for multisense feature, edje must compile e.g. ENABLE_MULTISENSE=1 and HAVE_REMIX=0. Regards, Prince I've tested in rev 72117 and now 72400 in both the problem happened. I built with these options: /configure --enable-tests --enable-build-examples --enable-multisense CFLAGS=-Wall -g -O0 As for the patch, I took a look in it and I would like to understand some points. +static void +pa_player_playbuffer(RemixEnv *env __UNUSED__, PA_Player_Data *player, RemixPCM *data, RemixCount count) +{ + int ret; + RemixCount i; + RemixPCM value; + size_t length; + + length = count * sizeof(RemixCount); Shouldn't it be length = count * sizeof(RemixPCM); [Prince] you are right, it must be the size of same data type. + + for (i = 0; i length; i++) Shouldn't check if length is greater than the size of the buffer ? [Prince] I agree. we can put the check so that written length must never exceeds the buffer size (PA_PLAYER_BUFFERLEN). + { +value = *data++ * (player-max_value); +*(player-playbuffer + i) = (PLAYER_PCM) value; With this cast we are discarding part of the data, is it ok ? [Prince] Type casting is redundant, not required at all. Just to be clear, PLAYER_PCM is short while the data(RemixPCM) is a float, so in this cast you are losing some data, I just don't know if it's ok. + } + + ret = pa_simple_write(player-server, player-playbuffer, length, player-error); + + if (ret 0) WRN(pa_simple_write() failed: (%s), pa_strerror(player-error)); + + return; +} Best regards, Flavio Ceolin --- -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel Prince I forgot to ask you if there is a reason to export this symbol: +EAPI extern Eina_Bool _on_edjecc; It seems to be not necessary. [Prince] this variable is required, in order
Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module
On Thursday, June 21, 2012, PRINCE KUMAR DUBEY wrote: Hi Flavio, Please find my reply inline below. Thanks Regards, Prince --- Original Message --- Date: Wed, 20 Jun 2012 15:40:14 -0300 From: Flavio Ceolin Subject: Re: [E-devel] enlightenment-devel Digest, Vol 74, Issue 78 To: Enlightenment developer list Message-ID: 877gv13jzl@voyager.profusion.mobi javascript:; Content-Type: text/plain PRINCE KUMAR DUBEY writes: Hi Prince, Hi Flavio, I am still not able to reproduce the seg-fault. Is seg-fault happening always ? Yep, it's happening always i use the edje_player with the multisense example. The problem is: + length = count * sizeof(RemixCount); it makes length be greater than PA_PLAYER_BUFFERLEN then: +*(player-playbuffer + i) = (PLAYER_PCM) value; will access invalid memory. I rewrote this function and worked as expected. Please check if it makes sense. static void pa_player_playbuffer(RemixEnv *env __UNUSED__, PA_Player_Data *player, RemixPCM *data, RemixCount count) { int ret; RemixCount i, j; RemixPCM value; size_t length, total_written; length = count * sizeof(RemixPCM); total_written = 0; while (total_written length) { j = length - total_written; j = (j PA_PLAYER_BUFFERLEN) ? PA_PLAYER_BUFFERLEN : j; for (i = 0; i j; i++) { value = *data++ * (player-max_value); *(player-playbuffer + i) = (PLAYER_PCM) value; } ret = pa_simple_write(player-server, player-playbuffer, j, player-error); if (ret 0) { WRN(pa_simple_write() failed: (%s), pa_strerror(player-error)); return; } total_written += j; } } [Prince] make sense :), thanks. Please find my reply to your queries inline below. Thanks Regards, Prince --- Original Message --- Date: Mon, 18 Jun 2012 14:09:59 -0300 From: Fl?vio Ceolin Subject: Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module To: Enlightenment developer list Message-ID: Content-Type: text/plain; charset=ISO-8859-1 Hi Prince, On Mon, Jun 18, 2012 at 9:56 AM, PRINCE KUMAR DUBEY wrote: Hi Flavio, I tested edje_player multisense.edje to reproduce seg fault as mentioned below, but couldn't succeed with memcheck as well. My svn rev is 72374. Please let me know the scenario to reproduce it. And regarding the point 5, you are right, multisense will not build without remix as per current configure file. But, multisense is not all about only sound, it can have vibration etc. So, if we make remix optional for multisense feature, edje must compile e.g. ENABLE_MULTISENSE=1 and HAVE_REMIX=0. Regards, Prince I've tested in rev 72117 and now 72400 in both the problem happened. I built with these options: /configure --enable-tests --enable-build-examples --enable-multisense CFLAGS=-Wall -g -O0 As for the patch, I took a look in it and I would like to understand some points. +static void +pa_player_playbuffer(RemixEnv *env __UNUSED__, PA_Player_Data *player, RemixPCM *data, RemixCount count) +{ + int ret; + RemixCount i; + RemixPCM value; + size_t length; + + length = count * sizeof(RemixCount); Shouldn't it be length = count * sizeof(RemixPCM); [Prince] you are right, it must be the size of same data type. + + for (i = 0; i length; i++) Shouldn't check if length is greater than the size of the buffer ? [Prince] I agree. we can put the check so that written length must never exceeds the buffer size (PA_PLAYER_BUFFERLEN). + { +value = *data++ * (player-max_value); +*(player-playbuffer + i) = (PLAYER_PCM) value; With this cast we are discarding part of the data, is it ok ? [Prince] Type casting is redundant, not required at all. Just to be clear, PLAYER_PCM is short while the data(RemixPCM) is a float, so in this cast you are losing some data, I just don't know if it's ok. + } + + ret = pa_simple_write(player-server, player-playbuffer, length, player-error); + + if (ret 0) WRN(pa_simple_write() failed: (%s), pa_strerror(player-error)); + + return; +} Best regards, Flavio Ceolin --- -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net javascript:; https://lists.sourceforge.net/lists/listinfo
Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module
On Thu, Jun 21, 2012 at 10:32 AM, PRINCE KUMAR DUBEY prince.du...@samsung.com wrote: Hi, Please find my query/comments inlined below. Ok, and could you use some email client that does proper reply? It's a mess and unusual to read your mails without the proper quote blocks. +EAPI extern Eina_Bool _on_edjecc; It seems to be not necessary. [Prince] this variable is required, in order to provide check to unnecessary multisense framework initialization during EDC compilation/decompilation mode. So, during EDC compilation, _on_edjecc=EINA_TRUE in edje_cc.c and hence, _edje_multisense_init() is being ignored in edje_main.c. No, never do this. One should have a global setter and maybe a getter to enable or disable features, or even better just load multisense if it's required. [Prince] Multisense load is not needed during compilation/decompilation, because it inculdes remix and its plug-in(i.e. eetsndreader, ALSA) load, which are specificaly required during runtime only. As per your suggestion above, should we add interfaces for setter/getter to enable this feature in edje ? Please share your opinion. I dislike having to manually enable or disable them. Just use it as last resource. Ideally the multisense would be disabled and whenever a group uses multisense, it would be enabled. You do not need to further disable the setup later, it's fine to keep it on. It's like a lazy-loaded singleton. As edje_cc will never edje_object_file_set() it will never hit the path to load multisense. Then we're fine for it and every other software that does not use it. -- Gustavo Sverzut Barbieri http://profusion.mobi embedded systems -- MSN: barbi...@gmail.com Skype: gsbarbieri Mobile: +55 (19) 9225-2202 -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module
Hi Flavio, I tested edje_player multisense.edje to reproduce seg fault as mentioned below, but couldn't succeed with memcheck as well. My svn rev is 72374. Please let me know the scenario to reproduce it. And regarding the point 5, you are right, multisense will not build without remix as per current configure file. But, multisense is not all about only sound, it can have vibration etc. So, if we make remix optional for multisense feature, edje must compile e.g. ENABLE_MULTISENSE=1 and HAVE_REMIX=0. Regards, Prince Date: Thu, 14 Jun 2012 17:31:43 -0300 From: Fl?vio Ceolin flavio.ceo...@profusion.mobi Subject: Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module To: Enlightenment developer list enlightenment-devel@lists.sourceforge.net Cc: csm...@samsung.com Message-ID: cakxyohkiyzzfxra7ff8jvsks-xqughyupeegnlidk_q4leo...@mail.gmail.com Content-Type: text/plain; charset=iso-8859-1 Hello Prince, On Thu, Jun 14, 2012 at 10:14 AM, PRINCE KUMAR DUBEY prince.du...@samsung.com wrote: Hi, ? Please review the attached patch for pulse audio remix pulg-in added as output device for multisense feature with other enhancements, details are given below. Change Description: 1. Multisense support enabled by default, Pulse Audio option added, flac, alsa and ogg/vorbis are made optional for multisense during configuration/compilation. 2. Pulse Audio Remix Plug-in added as audio output device. 3. Disabling of Multisense module initialization/shutdown in EDC compilation mode. 4. Enhancement to _edje_multisense_shutdown to fix memory leak. 5. Multisense example enhancement and modification in edje_multisense.c so, edje can compile without availability of remix. Signed-Off-By: Prince Kr Dubeyprince.du...@samsung.com Thank you, Regards, Prince Kr Dubey -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel I' ve found a segfault when running edje_player (multisense example, multisense.edj), you could check it in the attached file. It's valid you know that this problem is not happening when running the edje-multisense's example. Also, I didn't understand the point 5 (edje can compile without availability of remix). As far as I could see there is no way to build multisense without libremix right ? Regards, Flavio Ceolin -- next part -- ==11191== Memcheck, a memory error detector ==11191== Copyright (C) 2002-2011, and GNU GPL'd, by Julian Seward et al. ==11191== Using Valgrind-3.7.0 and LibVEX; rerun with -h for copyright info ==11191== Command: edje_player multisense.edj ==11191== Parent PID: 8099 ==11191== ==11191== Syscall param write(buf) points to uninitialised byte(s) ==11191==at 0x355100E04D: ??? (in /usr/lib64/libpthread-2.15.so) ==11191==by 0x4C7251C: _edje_multisense_internal_sound_sample_play (edje_multisense.c:344) ==11191==by 0x4C73EFF: _edje_program_run (edje_program.c:798) ==11191==by 0x4C74238: _edje_program_run (edje_program.c:595) ==11191==by 0x4C76007: _edje_emit_handle (edje_program.c:1228) ==11191==by 0x4C7174E: _edje_message_queue_process (edje_message_queue.c:764) ==11191==by 0x4C718B6: _edje_job (edje_message_queue.c:160) ==11191==by 0x69CF234: _ecore_job_event_handler (ecore_job.c:115) ==11191==by 0x69C95DC: _ecore_call_handler_cb (ecore_private.h:319) ==11191==by 0x69CA57B: _ecore_event_call (ecore_events.c:559) ==11191==by 0x69D141D: _ecore_main_loop_iterate_internal (ecore_main.c:1814) ==11191==by 0x69CFCBF: ecore_main_loop_begin (ecore_main.c:931) ==11191== Address 0x7fefff97c is on thread 1's stack ==11191== ==11191== Thread 3: ==11191== Invalid write of size 2 ==11191==at 0xE4538E5: pa_player_playbuffer.isra.0 (pa_simple_snd_player.c:211) ==11191==by 0xE4539F4: pa_player_process (pa_simple_snd_player.c:264) ==11191==by 0x56DFA9C: remix_process_fast (remix_base.c:434) ==11191==by 0x56DFB29: remix_process (remix_base.c:450) ==11191==by 0x56E4787: remix_sound_process (remix_sound.c:400) ==11191==by 0x56DFA9C: remix_process_fast (remix_base.c:434) ==11191==by 0x56DFB29: remix_process (remix_base.c:450) ==11191==by 0x56E3714: remix_layer_process (remix_layer.c:498) ==11191==by 0x56DFA9C: remix_process_fast (remix_base.c:434) ==11191==by 0x56DFB29: remix_process (remix_base.c:450) ==11191
Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module
Hi Prince, On Mon, Jun 18, 2012 at 9:56 AM, PRINCE KUMAR DUBEY prince.du...@samsung.com wrote: Hi Flavio, I tested edje_player multisense.edje to reproduce seg fault as mentioned below, but couldn't succeed with memcheck as well. My svn rev is 72374. Please let me know the scenario to reproduce it. And regarding the point 5, you are right, multisense will not build without remix as per current configure file. But, multisense is not all about only sound, it can have vibration etc. So, if we make remix optional for multisense feature, edje must compile e.g. ENABLE_MULTISENSE=1 and HAVE_REMIX=0. Regards, Prince I've tested in rev 72117 and now 72400 in both the problem happened. I built with these options: /configure --enable-tests --enable-build-examples --enable-multisense CFLAGS=-Wall -g -O0 As for the patch, I took a look in it and I would like to understand some points. +static void +pa_player_playbuffer(RemixEnv *env __UNUSED__, PA_Player_Data *player, RemixPCM *data, RemixCount count) +{ + int ret; + RemixCount i; + RemixPCM value; + size_t length; + + length = count * sizeof(RemixCount); Shouldn't it be length = count * sizeof(RemixPCM); + + for (i = 0; i length; i++) Shouldn't check if length is greater than the size of the buffer ? + { +value = *data++ * (player-max_value); +*(player-playbuffer + i) = (PLAYER_PCM) value; With this cast we are discarding part of the data, is it ok ? + } + + ret = pa_simple_write(player-server, player-playbuffer, length, player-error); + + if (ret 0) WRN(pa_simple_write() failed: (%s), pa_strerror(player-error)); + + return; +} Best regards, Flavio Ceolin -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
[E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module
Hi, Please review the attached patch for pulse audio remix pulg-in added as output device for multisense feature with other enhancements, details are given below. Change Description: 1. Multisense support enabled by default, Pulse Audio option added, flac, alsa and ogg/vorbis are made optional for multisense during configuration/compilation. 2. Pulse Audio Remix Plug-in added as audio output device. 3. Disabling of Multisense module initialization/shutdown in EDC compilation mode. 4. Enhancement to _edje_multisense_shutdown to fix memory leak. 5. Multisense example enhancement and modification in edje_multisense.c so, edje can compile without availability of remix. Signed-Off-By: Prince Kr Dubeyprince.du...@samsung.com Thank you, Regards, Prince Kr Dubey edje_multisense_pulseaudio_plugin.patch Description: Binary data -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] [Patch] [Edje] Added: Pulseaudio remix plug-in for edje multisense module
Hello Prince, On Thu, Jun 14, 2012 at 10:14 AM, PRINCE KUMAR DUBEY prince.du...@samsung.com wrote: Hi, Please review the attached patch for pulse audio remix pulg-in added as output device for multisense feature with other enhancements, details are given below. Change Description: 1. Multisense support enabled by default, Pulse Audio option added, flac, alsa and ogg/vorbis are made optional for multisense during configuration/compilation. 2. Pulse Audio Remix Plug-in added as audio output device. 3. Disabling of Multisense module initialization/shutdown in EDC compilation mode. 4. Enhancement to _edje_multisense_shutdown to fix memory leak. 5. Multisense example enhancement and modification in edje_multisense.c so, edje can compile without availability of remix. Signed-Off-By: Prince Kr Dubeyprince.du...@samsung.com Thank you, Regards, Prince Kr Dubey -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel I' ve found a segfault when running edje_player (multisense example, multisense.edj), you could check it in the attached file. It's valid you know that this problem is not happening when running the edje-multisense's example. Also, I didn't understand the point 5 (edje can compile without availability of remix). As far as I could see there is no way to build multisense without libremix right ? Regards, Flavio Ceolin ==11191== Memcheck, a memory error detector ==11191== Copyright (C) 2002-2011, and GNU GPL'd, by Julian Seward et al. ==11191== Using Valgrind-3.7.0 and LibVEX; rerun with -h for copyright info ==11191== Command: edje_player multisense.edj ==11191== Parent PID: 8099 ==11191== ==11191== Syscall param write(buf) points to uninitialised byte(s) ==11191==at 0x355100E04D: ??? (in /usr/lib64/libpthread-2.15.so) ==11191==by 0x4C7251C: _edje_multisense_internal_sound_sample_play (edje_multisense.c:344) ==11191==by 0x4C73EFF: _edje_program_run (edje_program.c:798) ==11191==by 0x4C74238: _edje_program_run (edje_program.c:595) ==11191==by 0x4C76007: _edje_emit_handle (edje_program.c:1228) ==11191==by 0x4C7174E: _edje_message_queue_process (edje_message_queue.c:764) ==11191==by 0x4C718B6: _edje_job (edje_message_queue.c:160) ==11191==by 0x69CF234: _ecore_job_event_handler (ecore_job.c:115) ==11191==by 0x69C95DC: _ecore_call_handler_cb (ecore_private.h:319) ==11191==by 0x69CA57B: _ecore_event_call (ecore_events.c:559) ==11191==by 0x69D141D: _ecore_main_loop_iterate_internal (ecore_main.c:1814) ==11191==by 0x69CFCBF: ecore_main_loop_begin (ecore_main.c:931) ==11191== Address 0x7fefff97c is on thread 1's stack ==11191== ==11191== Thread 3: ==11191== Invalid write of size 2 ==11191==at 0xE4538E5: pa_player_playbuffer.isra.0 (pa_simple_snd_player.c:211) ==11191==by 0xE4539F4: pa_player_process (pa_simple_snd_player.c:264) ==11191==by 0x56DFA9C: remix_process_fast (remix_base.c:434) ==11191==by 0x56DFB29: remix_process (remix_base.c:450) ==11191==by 0x56E4787: remix_sound_process (remix_sound.c:400) ==11191==by 0x56DFA9C: remix_process_fast (remix_base.c:434) ==11191==by 0x56DFB29: remix_process (remix_base.c:450) ==11191==by 0x56E3714: remix_layer_process (remix_layer.c:498) ==11191==by 0x56DFA9C: remix_process_fast (remix_base.c:434) ==11191==by 0x56DFB29: remix_process (remix_base.c:450) ==11191==by 0x56E6B5A: remix_track_twolayer_process (remix_track.c:350) ==11191==by 0x56DFA9C: remix_process_fast (remix_base.c:434) ==11191== Address 0x77a20f0 is 0 bytes after a block of size 8,192 alloc'd ==11191==at 0x4A06F18: calloc (vg_replace_malloc.c:566) ==11191==by 0xE453524: pa_player_reset_device (pa_simple_snd_player.c:81) ==11191==by 0xE4536B5: pa_player_init (pa_simple_snd_player.c:130) ==11191==by 0x56DF729: remix_new (remix_base.c:55) ==11191==by 0x4C727B5: _edje_multisense_init (edje_multisense.c:109) ==11191==by 0x4C6EF24: edje_init (edje_main.c:73) ==11191==by 0x4022F3: main (edje_player.c:662) ==11191== ==11191== Invalid read of size 4 ==11191==at 0xE4538ED: pa_player_playbuffer.isra.0 (pa_simple_snd_player.c:208) ==11191==by 0xE4539F4: pa_player_process (pa_simple_snd_player.c:264) ==11191==by 0x56DFA9C: remix_process_fast (remix_base.c:434) ==11191==by 0x56DFB29: remix_process (remix_base.c:450) ==11191==by 0x56E4787: remix_sound_process (remix_sound.c:400) ==11191==by 0x56DFA9C: