Re: [FFmpeg-devel] [FFmpeg-cvslog] Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08'
I'd like to use this opportunity to remind of http://ffmpeg.org/pipermail/ffmpeg-devel/2017-February/207513.html 2017-03-20 14:29 GMT+03:00 Michael Niedermayer : > On Mon, Mar 20, 2017 at 10:14:21AM +0100, Hendrik Leppkes wrote: > > On Mon, Mar 20, 2017 at 10:11 AM, Carl Eugen Hoyos > wrote: > > > 2017-03-20 10:03 GMT+01:00 Hendrik Leppkes : > > >> On Mon, Mar 20, 2017 at 9:59 AM, Carl Eugen Hoyos > wrote: > > >>> 2017-03-20 9:42 GMT+01:00 Clément Bœsch : > > On Mon, Mar 20, 2017 at 09:38:15AM +0100, Carl Eugen Hoyos wrote: > > > 2017-03-20 9:35 GMT+01:00 Clément Bœsch : > > > > On Mon, Mar 20, 2017 at 09:17:36AM +0100, Carl Eugen Hoyos wrote: > > > >> 2017-03-20 8:57 GMT+01:00 Clément Bœsch : > > > >> > ffmpeg | branch: master | Clément Bœsch | Mon Mar > 20 08:52:07 2017 +0100| [3835283293bfd38ba69203f4618f0f0f21377bcc] | > committer: Clément Bœsch > > > >> > > > > >> > Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08' > > > >> > > > > >> > * commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08': > > > >> > Drop memalign hack > > > >> > > > > >> > Merged, as this may indeed be uneeded since > > > >> > 46e3936fb04d06550151e667357065e3f646da1a. > > > >> > > > >> How are we supposed to test the next crash because > > > >> of missing alignment? > > > >> > > > > > > > > I thought it was only about compatibility with platforms not > having > > > > an aligned allocator. > > > > > > If you think such platforms exist, one more reason not to merge... > > > > There was apparently MinGW but it seemed to have been fixed > > (see the hash reference in the merge description). > > >>> > > >>> I can only add that the simple idea of removing support for > > >>> platforms without any cause fills me with shame. > > >>> > > >> > > >> If you can name a platfom that needs this in current git master, then > > >> we can move forward. > > > > > > Old MingW comes to mind... > > > > > > > As was pointed out above, old mingw was fixed in another commit right > > before this one. > > If you can find an actual reproducable case on current git master, > > please do provide said information. > > Is the mixture of memalign() and free() spec compliant ? > > IIRC you said in IRC memalign and free would be ok in posix but as i > checked now i cannot find memalign in posix only posix_memalign() > but maybe i looked at the wrong place > > > [...] > -- > Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB > > The worst form of inequality is to try to make unequal things equal. > -- Aristotle > > ___ > ffmpeg-devel mailing list > ffmpeg-devel@ffmpeg.org > http://ffmpeg.org/mailman/listinfo/ffmpeg-devel > > ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Re: [FFmpeg-devel] [FFmpeg-cvslog] Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08'
On Mon, Mar 20, 2017 at 10:14:21AM +0100, Hendrik Leppkes wrote: > On Mon, Mar 20, 2017 at 10:11 AM, Carl Eugen Hoyos wrote: > > 2017-03-20 10:03 GMT+01:00 Hendrik Leppkes : > >> On Mon, Mar 20, 2017 at 9:59 AM, Carl Eugen Hoyos > >> wrote: > >>> 2017-03-20 9:42 GMT+01:00 Clément Bœsch : > On Mon, Mar 20, 2017 at 09:38:15AM +0100, Carl Eugen Hoyos wrote: > > 2017-03-20 9:35 GMT+01:00 Clément Bœsch : > > > On Mon, Mar 20, 2017 at 09:17:36AM +0100, Carl Eugen Hoyos wrote: > > >> 2017-03-20 8:57 GMT+01:00 Clément Bœsch : > > >> > ffmpeg | branch: master | Clément Bœsch | Mon Mar 20 > > >> > 08:52:07 2017 +0100| [3835283293bfd38ba69203f4618f0f0f21377bcc] | > > >> > committer: Clément Bœsch > > >> > > > >> > Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08' > > >> > > > >> > * commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08': > > >> > Drop memalign hack > > >> > > > >> > Merged, as this may indeed be uneeded since > > >> > 46e3936fb04d06550151e667357065e3f646da1a. > > >> > > >> How are we supposed to test the next crash because > > >> of missing alignment? > > >> > > > > > > I thought it was only about compatibility with platforms not having > > > an aligned allocator. > > > > If you think such platforms exist, one more reason not to merge... > > There was apparently MinGW but it seemed to have been fixed > (see the hash reference in the merge description). > >>> > >>> I can only add that the simple idea of removing support for > >>> platforms without any cause fills me with shame. > >>> > >> > >> If you can name a platfom that needs this in current git master, then > >> we can move forward. > > > > Old MingW comes to mind... > > > > As was pointed out above, old mingw was fixed in another commit right > before this one. > If you can find an actual reproducable case on current git master, > please do provide said information. Is the mixture of memalign() and free() spec compliant ? IIRC you said in IRC memalign and free would be ok in posix but as i checked now i cannot find memalign in posix only posix_memalign() but maybe i looked at the wrong place [...] -- Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB The worst form of inequality is to try to make unequal things equal. -- Aristotle signature.asc Description: Digital signature ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Re: [FFmpeg-devel] [FFmpeg-cvslog] Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08'
On Mon, Mar 20, 2017 at 10:12:49AM +0100, Carl Eugen Hoyos wrote: > 2017-03-20 10:04 GMT+01:00 Clément Bœsch : > > >> Is there really no incentive that would make you work on subtitle > >> support in the filter chain? > > > > Yes, someone else working on the remaining 870+ commits to merge. > > You told me last year you don't care what happens with those > who must not be named - I wonder what changed your opinion. > I do not care what happens to the Libav project and that didn't change. But I do care about the FFmpeg project. Now, I'm fine reverting that merge, but I'm looking for an actual proof of something breaking. Regards, -- Clément B. signature.asc Description: PGP signature ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Re: [FFmpeg-devel] [FFmpeg-cvslog] Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08'
On Mon, Mar 20, 2017 at 10:11 AM, Carl Eugen Hoyos wrote: > 2017-03-20 10:03 GMT+01:00 Hendrik Leppkes : >> On Mon, Mar 20, 2017 at 9:59 AM, Carl Eugen Hoyos wrote: >>> 2017-03-20 9:42 GMT+01:00 Clément Bœsch : On Mon, Mar 20, 2017 at 09:38:15AM +0100, Carl Eugen Hoyos wrote: > 2017-03-20 9:35 GMT+01:00 Clément Bœsch : > > On Mon, Mar 20, 2017 at 09:17:36AM +0100, Carl Eugen Hoyos wrote: > >> 2017-03-20 8:57 GMT+01:00 Clément Bœsch : > >> > ffmpeg | branch: master | Clément Bœsch | Mon Mar 20 > >> > 08:52:07 2017 +0100| [3835283293bfd38ba69203f4618f0f0f21377bcc] | > >> > committer: Clément Bœsch > >> > > >> > Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08' > >> > > >> > * commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08': > >> > Drop memalign hack > >> > > >> > Merged, as this may indeed be uneeded since > >> > 46e3936fb04d06550151e667357065e3f646da1a. > >> > >> How are we supposed to test the next crash because > >> of missing alignment? > >> > > > > I thought it was only about compatibility with platforms not having > > an aligned allocator. > > If you think such platforms exist, one more reason not to merge... There was apparently MinGW but it seemed to have been fixed (see the hash reference in the merge description). >>> >>> I can only add that the simple idea of removing support for >>> platforms without any cause fills me with shame. >>> >> >> If you can name a platfom that needs this in current git master, then >> we can move forward. > > Old MingW comes to mind... > As was pointed out above, old mingw was fixed in another commit right before this one. If you can find an actual reproducable case on current git master, please do provide said information. - Hendrik ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Re: [FFmpeg-devel] [FFmpeg-cvslog] Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08'
2017-03-20 10:04 GMT+01:00 Clément Bœsch : >> Is there really no incentive that would make you work on subtitle >> support in the filter chain? > > Yes, someone else working on the remaining 870+ commits to merge. You told me last year you don't care what happens with those who must not be named - I wonder what changed your opinion. Carl Eugen ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Re: [FFmpeg-devel] [FFmpeg-cvslog] Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08'
2017-03-20 10:03 GMT+01:00 Hendrik Leppkes : > On Mon, Mar 20, 2017 at 9:59 AM, Carl Eugen Hoyos wrote: >> 2017-03-20 9:42 GMT+01:00 Clément Bœsch : >>> On Mon, Mar 20, 2017 at 09:38:15AM +0100, Carl Eugen Hoyos wrote: 2017-03-20 9:35 GMT+01:00 Clément Bœsch : > On Mon, Mar 20, 2017 at 09:17:36AM +0100, Carl Eugen Hoyos wrote: >> 2017-03-20 8:57 GMT+01:00 Clément Bœsch : >> > ffmpeg | branch: master | Clément Bœsch | Mon Mar 20 >> > 08:52:07 2017 +0100| [3835283293bfd38ba69203f4618f0f0f21377bcc] | >> > committer: Clément Bœsch >> > >> > Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08' >> > >> > * commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08': >> > Drop memalign hack >> > >> > Merged, as this may indeed be uneeded since >> > 46e3936fb04d06550151e667357065e3f646da1a. >> >> How are we supposed to test the next crash because >> of missing alignment? >> > > I thought it was only about compatibility with platforms not having > an aligned allocator. If you think such platforms exist, one more reason not to merge... >>> >>> There was apparently MinGW but it seemed to have been fixed >>> (see the hash reference in the merge description). >> >> I can only add that the simple idea of removing support for >> platforms without any cause fills me with shame. >> > > If you can name a platfom that needs this in current git master, then > we can move forward. Old MingW comes to mind... > Everything else is just trolling Thank you, it makes me always happy knowing I am the only one supposed to follow the new guidelines. Carl Eugen ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Re: [FFmpeg-devel] [FFmpeg-cvslog] Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08'
2017-03-20 10:07 GMT+01:00 wm4 : > You're invited to do something more productive as well. You mean more productive than helping users? Carl Eugen ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Re: [FFmpeg-devel] [FFmpeg-cvslog] Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08'
On Mon, Mar 20, 2017 at 9:59 AM, Carl Eugen Hoyos wrote: > 2017-03-20 9:42 GMT+01:00 Clément Bœsch : >> On Mon, Mar 20, 2017 at 09:38:15AM +0100, Carl Eugen Hoyos wrote: >>> 2017-03-20 9:35 GMT+01:00 Clément Bœsch : >>> > On Mon, Mar 20, 2017 at 09:17:36AM +0100, Carl Eugen Hoyos wrote: >>> >> 2017-03-20 8:57 GMT+01:00 Clément Bœsch : >>> >> > ffmpeg | branch: master | Clément Bœsch | Mon Mar 20 >>> >> > 08:52:07 2017 +0100| [3835283293bfd38ba69203f4618f0f0f21377bcc] | >>> >> > committer: Clément Bœsch >>> >> > >>> >> > Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08' >>> >> > >>> >> > * commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08': >>> >> > Drop memalign hack >>> >> > >>> >> > Merged, as this may indeed be uneeded since >>> >> > 46e3936fb04d06550151e667357065e3f646da1a. >>> >> >>> >> How are we supposed to test the next crash because >>> >> of missing alignment? >>> >> >>> > >>> > I thought it was only about compatibility with platforms not having >>> > an aligned allocator. >>> >>> If you think such platforms exist, one more reason not to merge... >> >> There was apparently MinGW but it seemed to have been fixed >> (see the hash reference in the merge description). > > I can only add that the simple idea of removing support for > platforms without any cause fills me with shame. > If you can name a platfom that needs this in current git master, then we can move forward. Everything else is just trolling - Hendrik ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Re: [FFmpeg-devel] [FFmpeg-cvslog] Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08'
On Mon, 20 Mar 2017 09:59:24 +0100 Carl Eugen Hoyos wrote: > 2017-03-20 9:42 GMT+01:00 Clément Bœsch : > > On Mon, Mar 20, 2017 at 09:38:15AM +0100, Carl Eugen Hoyos wrote: > >> 2017-03-20 9:35 GMT+01:00 Clément Bœsch : > >> > On Mon, Mar 20, 2017 at 09:17:36AM +0100, Carl Eugen Hoyos wrote: > >> >> 2017-03-20 8:57 GMT+01:00 Clément Bœsch : > >> >> > ffmpeg | branch: master | Clément Bœsch | Mon Mar 20 > >> >> > 08:52:07 2017 +0100| [3835283293bfd38ba69203f4618f0f0f21377bcc] | > >> >> > committer: Clément Bœsch > >> >> > > >> >> > Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08' > >> >> > > >> >> > * commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08': > >> >> > Drop memalign hack > >> >> > > >> >> > Merged, as this may indeed be uneeded since > >> >> > 46e3936fb04d06550151e667357065e3f646da1a. > >> >> > >> >> How are we supposed to test the next crash because > >> >> of missing alignment? > >> >> > >> > > >> > I thought it was only about compatibility with platforms not having > >> > an aligned allocator. > >> > >> If you think such platforms exist, one more reason not to merge... > > > > There was apparently MinGW but it seemed to have been fixed > > (see the hash reference in the merge description). > > I can only add that the simple idea of removing support for > platforms without any cause fills me with shame. All the code in mem.c fills me with shame. Besides, the memalign hack used to be required, and people still compile with it forcibly enabled due to cargo-cult. This is inefficient, so good enough reason to remove. Of course we shouldn't accidentally drop support for platforms. Can you tell us which platforms these are? > Is there really no incentive that would make you work on subtitle > support in the filter chain? You're invited to do something more productive as well. ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Re: [FFmpeg-devel] [FFmpeg-cvslog] Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08'
On Mon, Mar 20, 2017 at 09:59:24AM +0100, Carl Eugen Hoyos wrote: > 2017-03-20 9:42 GMT+01:00 Clément Bœsch : > > On Mon, Mar 20, 2017 at 09:38:15AM +0100, Carl Eugen Hoyos wrote: > >> 2017-03-20 9:35 GMT+01:00 Clément Bœsch : > >> > On Mon, Mar 20, 2017 at 09:17:36AM +0100, Carl Eugen Hoyos wrote: > >> >> 2017-03-20 8:57 GMT+01:00 Clément Bœsch : > >> >> > ffmpeg | branch: master | Clément Bœsch | Mon Mar 20 > >> >> > 08:52:07 2017 +0100| [3835283293bfd38ba69203f4618f0f0f21377bcc] | > >> >> > committer: Clément Bœsch > >> >> > > >> >> > Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08' > >> >> > > >> >> > * commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08': > >> >> > Drop memalign hack > >> >> > > >> >> > Merged, as this may indeed be uneeded since > >> >> > 46e3936fb04d06550151e667357065e3f646da1a. > >> >> > >> >> How are we supposed to test the next crash because > >> >> of missing alignment? > >> >> > >> > > >> > I thought it was only about compatibility with platforms not having > >> > an aligned allocator. > >> > >> If you think such platforms exist, one more reason not to merge... > > > > There was apparently MinGW but it seemed to have been fixed > > (see the hash reference in the merge description). > > I can only add that the simple idea of removing support for > platforms without any cause fills me with shame. > But I just said it's now fixed for the only known affected platform... > Is there really no incentive that would make you work on subtitle > support in the filter chain? Yes, someone else working on the remaining 870+ commits to merge. > Carl Eugen -- Clément B. signature.asc Description: PGP signature ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Re: [FFmpeg-devel] [FFmpeg-cvslog] Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08'
2017-03-20 9:42 GMT+01:00 Clément Bœsch : > On Mon, Mar 20, 2017 at 09:38:15AM +0100, Carl Eugen Hoyos wrote: >> 2017-03-20 9:35 GMT+01:00 Clément Bœsch : >> > On Mon, Mar 20, 2017 at 09:17:36AM +0100, Carl Eugen Hoyos wrote: >> >> 2017-03-20 8:57 GMT+01:00 Clément Bœsch : >> >> > ffmpeg | branch: master | Clément Bœsch | Mon Mar 20 >> >> > 08:52:07 2017 +0100| [3835283293bfd38ba69203f4618f0f0f21377bcc] | >> >> > committer: Clément Bœsch >> >> > >> >> > Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08' >> >> > >> >> > * commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08': >> >> > Drop memalign hack >> >> > >> >> > Merged, as this may indeed be uneeded since >> >> > 46e3936fb04d06550151e667357065e3f646da1a. >> >> >> >> How are we supposed to test the next crash because >> >> of missing alignment? >> >> >> > >> > I thought it was only about compatibility with platforms not having >> > an aligned allocator. >> >> If you think such platforms exist, one more reason not to merge... > > There was apparently MinGW but it seemed to have been fixed > (see the hash reference in the merge description). I can only add that the simple idea of removing support for platforms without any cause fills me with shame. Is there really no incentive that would make you work on subtitle support in the filter chain? Carl Eugen ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Re: [FFmpeg-devel] [FFmpeg-cvslog] Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08'
On Mon, 20 Mar 2017 09:38:15 +0100 Carl Eugen Hoyos wrote: > 2017-03-20 9:35 GMT+01:00 Clément Bœsch : > > On Mon, Mar 20, 2017 at 09:17:36AM +0100, Carl Eugen Hoyos wrote: > >> 2017-03-20 8:57 GMT+01:00 Clément Bœsch : > >> > ffmpeg | branch: master | Clément Bœsch | Mon Mar 20 > >> > 08:52:07 2017 +0100| [3835283293bfd38ba69203f4618f0f0f21377bcc] | > >> > committer: Clément Bœsch > >> > > >> > Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08' > >> > > >> > * commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08': > >> > Drop memalign hack > >> > > >> > Merged, as this may indeed be uneeded since > >> > 46e3936fb04d06550151e667357065e3f646da1a. > >> > >> How are we supposed to test the next crash because > >> of missing alignment? > >> > > > > I thought it was only about compatibility with platforms not having > > an aligned allocator. > > If you think such platforms exist, one more reason not to merge... Well we don't think such a platform supported by FFmpeg exists. > > Do you have a concrete example where it was actually used to > > debug such issue? > > The FreeBSD Firefox ticket? > > But it is actually every such ticket: How else can I exclude a compiler / > toolchain / libc issue? Just add an assert() verifying alignment before returning from av_malloc? ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Re: [FFmpeg-devel] [FFmpeg-cvslog] Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08'
On Mon, Mar 20, 2017 at 09:38:15AM +0100, Carl Eugen Hoyos wrote: > 2017-03-20 9:35 GMT+01:00 Clément Bœsch : > > On Mon, Mar 20, 2017 at 09:17:36AM +0100, Carl Eugen Hoyos wrote: > >> 2017-03-20 8:57 GMT+01:00 Clément Bœsch : > >> > ffmpeg | branch: master | Clément Bœsch | Mon Mar 20 > >> > 08:52:07 2017 +0100| [3835283293bfd38ba69203f4618f0f0f21377bcc] | > >> > committer: Clément Bœsch > >> > > >> > Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08' > >> > > >> > * commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08': > >> > Drop memalign hack > >> > > >> > Merged, as this may indeed be uneeded since > >> > 46e3936fb04d06550151e667357065e3f646da1a. > >> > >> How are we supposed to test the next crash because > >> of missing alignment? > >> > > > > I thought it was only about compatibility with platforms not having > > an aligned allocator. > > If you think such platforms exist, one more reason not to merge... > There was apparently MinGW but it seemed to have been fixed (see the hash reference in the merge description). > > Do you have a concrete example where it was actually used to > > debug such issue? > > The FreeBSD Firefox ticket? > I'm not familiar with all the tickets on our trac. Do you have a link? > But it is actually every such ticket: How else can I exclude a compiler / > toolchain / libc issue? By trying another toolchain/libc? -- Clément B. signature.asc Description: PGP signature ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Re: [FFmpeg-devel] [FFmpeg-cvslog] Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08'
2017-03-20 9:35 GMT+01:00 Clément Bœsch : > On Mon, Mar 20, 2017 at 09:17:36AM +0100, Carl Eugen Hoyos wrote: >> 2017-03-20 8:57 GMT+01:00 Clément Bœsch : >> > ffmpeg | branch: master | Clément Bœsch | Mon Mar 20 >> > 08:52:07 2017 +0100| [3835283293bfd38ba69203f4618f0f0f21377bcc] | >> > committer: Clément Bœsch >> > >> > Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08' >> > >> > * commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08': >> > Drop memalign hack >> > >> > Merged, as this may indeed be uneeded since >> > 46e3936fb04d06550151e667357065e3f646da1a. >> >> How are we supposed to test the next crash because >> of missing alignment? >> > > I thought it was only about compatibility with platforms not having > an aligned allocator. If you think such platforms exist, one more reason not to merge... > Do you have a concrete example where it was actually used to > debug such issue? The FreeBSD Firefox ticket? But it is actually every such ticket: How else can I exclude a compiler / toolchain / libc issue? Carl Eugen ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
Re: [FFmpeg-devel] [FFmpeg-cvslog] Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08'
On Mon, Mar 20, 2017 at 09:17:36AM +0100, Carl Eugen Hoyos wrote: > 2017-03-20 8:57 GMT+01:00 Clément Bœsch : > > ffmpeg | branch: master | Clément Bœsch | Mon Mar 20 08:52:07 > > 2017 +0100| [3835283293bfd38ba69203f4618f0f0f21377bcc] | committer: Clément > > Bœsch > > > > Merge commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08' > > > > * commit '4fb311c804098d78e5ce5f527f9a9c37536d3a08': > > Drop memalign hack > > > > Merged, as this may indeed be uneeded since > > 46e3936fb04d06550151e667357065e3f646da1a. > > How are we supposed to test the next crash because > of missing alignment? > I thought it was only about compatibility with platforms not having an aligned allocator. Do you have a concrete example where it was actually used to debug such issue? > Honestly, don't you realize why these patches were made? > What do you mean? > And isn't this breaking FreeBSD? > I have high doubts about FreeBSD libc not having aligned allocator. > Carl Eugen Note: I'm CC'ing ffmpeg-devel so this can be discussed more extensively. -- Clément B. signature.asc Description: PGP signature ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel