michaelni changed the topic of #ffmpeg-devel to: Welcome to the FFmpeg development channel | Questions about using FFmpeg or developing with libav* libs should be asked in #ffmpeg | This channel is publicly logged | FFmpeg 7.1.1 has been released! | Please read ffmpeg.org/developer.html#Code-of-conduct
minimal has quit [Quit: Leaving]
iive has quit [Quit: They came for me...]
averne has joined #ffmpeg-devel
TheVibeCoder has quit [Ping timeout: 260 seconds]
TheVibeCoder has joined #ffmpeg-devel
<BtbN> So, why exactly does schannel in dtls server mode want me to pass the client sockaddr? I don't see what it'd possibly do with that info.
<BtbN> Getting it would also involve a lot of patching of udp.c
<BtbN> So I'm wondering what the consequence of just lying would be
<BtbN> Though I need to get it anyway, since it'll eventually need to reply to that client... So might as well pass it
cone-420 has joined #ffmpeg-devel
<cone-420> ffmpeg Pavel Koshevoy master:59a666062567: avformat/demux: Fix segfault due to avcodec_open2 failure
<cone-420> ffmpeg Dawid Kozinski master:6e8bd5dd2588: avformat/apvdec: add framerate option
^Neo_ has joined #ffmpeg-devel
^Neo has quit [Ping timeout: 276 seconds]
Guest99 has joined #ffmpeg-devel
Guest99 has quit [Client Quit]
Kei_N_ has quit [Read error: Connection reset by peer]
mkver has joined #ffmpeg-devel
Martchus has joined #ffmpeg-devel
Martchus_ has quit [Ping timeout: 252 seconds]
<Lynne> TheVibeCoder: prores raw has a new mechanism to handle tiles with constant value
<Lynne> this fixed a few overreads
<Lynne> still figuring out where the hell its pulling its second half of DC values (tile width is increased, and the first half DCs are all fine, but not the second)
jamrial has quit []
mkver has quit [Quit: Leaving]
_whitelogger has joined #ffmpeg-devel
Richardcavell_ has joined #ffmpeg-devel
cone-420 has quit [Quit: transmission timeout]
rvalue has quit [Read error: Connection reset by peer]
rvalue has joined #ffmpeg-devel
^Neo has joined #ffmpeg-devel
^Neo_ has quit [Ping timeout: 244 seconds]
System_Error has quit [Remote host closed the connection]
Richardcavell_ has quit [Ping timeout: 245 seconds]
System_Error has joined #ffmpeg-devel
averne_ has joined #ffmpeg-devel
averne has quit [Ping timeout: 244 seconds]
averne_ is now known as averne
rvalue has quit [Ping timeout: 260 seconds]
MisterMinister has quit [Ping timeout: 272 seconds]
<fflogger> [newticket] momo-trip: Ticket #11651 ([undetermined] NULL Pointer Dereference in FFmpeg ffprobe) created https://trac.ffmpeg.org/ticket/11651
<Lynne> TheVibeCoder: found them, the tile width changed to 256, but there's a reset point and I'm pretty sure a codebook change for the DCs at the halfway block number in each tile
<Lynne> do you have the decompiled binary around, or really just the binary for the apple decoder?
<Lynne> I'd like to find if there's a second DC codebook
MetaNova has quit [Ping timeout: 276 seconds]
Richardcavell_ has joined #ffmpeg-devel
MetaNova has joined #ffmpeg-devel
ramiro has quit [Ping timeout: 276 seconds]
ramiro has joined #ffmpeg-devel
<TheVibeCoder> Lynne: https://0x0.st/8l5n.7z
<TheVibeCoder> dunno if it have support for version 1
IndecisiveTurtle has joined #ffmpeg-devel
Richardcavell_ has quit [Ping timeout: 268 seconds]
<wbs> BtbN: if you have time and have access to the fate server - can you check if some checkout of fateserver.git needs to be updated, to apply the change I just pushed - or is it automatically updated/in use just by pushing?
<kasper93> x86_64-archlinux-gcc-disableswscale slot seems to be affected so we can test on that, once deployed
<kasper93> wbs: if ($npass) seems iffy also, what if $npass is 0?
<kasper93> it would fall-through to "build only" green tests?
<kasper93> oh, but that also checks rep status
<kasper93> so probably is's le ok
<kasper93> also this ($npass? 'warn' : 'fail') is dead code
<kasper93> and this mirror is not syncing https://github.com/FFmpeg/fateserver
<fflogger> [newticket] sigdevel: Ticket #11652 ([ffmpeg] [security] libavcodec/hevc/hevcdec.c:2147:16 SEGV in hls_prediction_unit) created https://trac.ffmpeg.org/ticket/11652
Richardcavell_ has joined #ffmpeg-devel
rvalue has joined #ffmpeg-devel
<TheVibeCoder> compnn: upload speex sample to FATE, immediately, now!!!
<TheVibeCoder> or not, as speex is float decoder, would need also to upload .pcm decoded output
System_Error has quit [Ping timeout: 244 seconds]
System_Error has joined #ffmpeg-devel
ramiro has quit [Ping timeout: 276 seconds]
ramiro has joined #ffmpeg-devel
minimal has joined #ffmpeg-devel
Everything has joined #ffmpeg-devel
rvalue has quit [Read error: Connection reset by peer]
rvalue has joined #ffmpeg-devel
cone-375 has joined #ffmpeg-devel
<cone-375> ffmpeg Marvin Scholz master:7d384869751b: avcodec/vvc/refs: remove early return
bsFFFFFF has joined #ffmpeg-devel
bsFFFFFF has quit [Client Quit]
bsFFFFFF has joined #ffmpeg-devel
lemourin has quit [Quit: The Lounge - https://thelounge.chat]
<Lynne> TheVibeCoder: opencl? is this an official apple thing?
MisterMinister has joined #ffmpeg-devel
<TheVibeCoder> Lynne: apple official thing is not in .dll
<Lynne> who made the opencl version then?
<TheVibeCoder> this is only one i was able to get for free on web
<TheVibeCoder> search and you shall find
\\Mr_C\\ has joined #ffmpeg-devel
lemourin has joined #ffmpeg-devel
<Lynne> ah, adobe
<Lynne> someone sent me the official apple build, maybe that'll be better
<Lynne> or worse, if its just the hardware interop part
<Lynne> I wonder if non prores raw hq files are version 0...
<TheVibeCoder> hq is full 14 bit iirc
<TheVibeCoder> or 12bit
<TheVibeCoder> anything lower is 10bit
<TheVibeCoder> of actual range of decoded values
<TheVibeCoder> Lynne: and opencl binary is from atomos iirc
<Lynne> non-hq files are still version 1
<Lynne> I bet all implementations of prores raw are riddled with bugs
<Lynne> everyone universally hates it and avoids it like the plague for one simple reason - blackmagic refuse to support it in davinci
<TheVibeCoder> Lynne: what you mean by non-hq?
<Lynne> simply prores raw, not prores raw hq
<TheVibeCoder> and what is hq ?
<Lynne> its got half the bitrate of hq
kurosu has joined #ffmpeg-devel
<Lynne> your guess is as good as mine, I think they simply use a lower quantizer for hq
<Lynne> it looks the same bitstream-wise
<TheVibeCoder> Stream #0:1[0x2](eng): Video: proresraw (aprh / 0x68727061), bayer_rggb16le(unknown/unknown/linear, progressive), 4264x2408, 889923 kb/s, SAR 1:1 DAR 533:301, 23.98 fps, 23.98 tbr, 24k tbn (default)
<TheVibeCoder> this is version 0
<TheVibeCoder> with vendor atm0
<TheVibeCoder> the version 1 you gave and version 1 that i had already have all vendor peac
<TheVibeCoder> which is for Panasonic
Everything has quit [Quit: leaving]
<BtbN> wbs: I have access, but never touched that stuff. What exactly did you push where?
<BtbN> that's what's running at fate.ffmpeg.org, right?
<BtbN> I think I updated it? can you check?
<kasper93> looks like it works, --disable-swscale build is marked as warning as expected. But let's wait for wbs if that's what he wanted fixed
<kasper93> (I would make also test build errors a failure, but I guess some of them are passing, shrug)
<BtbN> I only ever look at fatebeta
<Lynne> TheVibeCoder: could you send me the sample?
<Lynne> I'd like to avoid breaking version 0
<TheVibeCoder> Lynne: https://0x0.st/8lFW.7z
<Lynne> cheers
<kasper93> seems the ffmpeg-fix_sub_duration_heartbeat is flakely, it sometimes get one more output line on this build for example https://fate.ffmpeg.org/history.cgi?slot=amd64-clang-cl-msvc
elvis_a_presley has quit [Quit: smoke-bomb ; grapple-hook]
elvis_a_presley has joined #ffmpeg-devel
cone-375 has quit [Quit: transmission timeout]
bsFFFFFF has quit [Ping timeout: 252 seconds]
MisterMinister has quit [Ping timeout: 260 seconds]
MisterMinister has joined #ffmpeg-devel
kurosu has quit [Quit: Connection closed for inactivity]
<BtbN> Hm, the receive logic in udp.c is quite something
<BtbN> why does it implement an entire ringbuffer on its own? Doesn't the OS take care of that?
Everything has joined #ffmpeg-devel
Everything has left #ffmpeg-devel [#ffmpeg-devel]
<nevcairiel> Not actually all that well
<nevcairiel> Before the thread and buffering existed udp was very unreliable
<BtbN> hm, that makes my undertaking a lot more complicated then
<BtbN> Though I guess I can just write the sockaddr into the buffer as well
<BtbN> and fetch it on udp_read
<BtbN> hm, the entirety of the udp.c code seems to be built around the assumption that the udp socket will only ever read or only ever write
<BtbN> Cause both udp_read and udp_write just check if (s->fifo), and then read from or write to it
ramiro has quit [Ping timeout: 244 seconds]
ramiro has joined #ffmpeg-devel
minimal has quit [Quit: Leaving]
<kierank> BtbN: what happens for mixed ipv4 and ipv6 packets
<kierank> you have hardcoded 4 bytes everywhere
<BtbN> how would a single packet ever be mixed?
<BtbN> The hardcoded 4 bytes are already there, and are the uint32 that contains the packet length.
<kierank> in the fifo they can be mixed
<kierank> if you listen on a port
<kierank> and don't bind()
<BtbN> sockaddr_storage can happily contain either. Thats its whole point.
<kierank> ok
<BtbN> I pretty much just shoved a sockaddr_storage between the packet lengths and its data in the fifo
Richardcavell_ has quit [Ping timeout: 265 seconds]
<wbs> kasper93: BtbN: looks indeed like it's deployed as expected and seems to work fine - thanks!
<TheVibeCoder> ffprobe -/version gives nice output
lemourin has quit [Quit: Ping timeout (120 seconds)]
lemourin has joined #ffmpeg-devel
IndecisiveTurtle has quit [Quit: IndecisiveTurtle]
iive has joined #ffmpeg-devel
Guest43 has joined #ffmpeg-devel
arch1t3cht1 has quit [Quit: The Lounge - https://thelounge.chat]
arch1t3cht1 has joined #ffmpeg-devel
Guest43 has quit [Quit: Client closed]
OC has joined #ffmpeg-devel
OC has quit [Quit: Client closed]
<kasper93> most secure tool
rvalue has quit [Read error: Connection reset by peer]
rvalue has joined #ffmpeg-devel
cone-186 has joined #ffmpeg-devel
<cone-186> ffmpeg Lidong Yan master:40a3d35da6ba: avcodec/vorbisenc: fix leak if av_mallocz failed
<cone-186> ffmpeg Nicolas Gaullier master:89ec66d61b2c: libavcodec/tests: fix gitignore hashtable
<cone-186> ffmpeg Michael Niedermayer master:cfd1f81e7d06: avcodec/speexdec: consider differing frame sizes in remaining space check
<cone-186> ffmpeg Lidong Yan master:a17596f55bca: avfilter/vf_lut3d: fix leak if allocate_3dlut failed
lemourin has quit [Quit: The Lounge - https://thelounge.chat]
<BtbN> schannel is making me go mad. It's failing. And it's failing because ERROR
\\Mr_C\\ has quit [Remote host closed the connection]
lemourin has joined #ffmpeg-devel