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 6.1.1 has been released! | Please read ffmpeg.org/developer.html#Code-of-conduct
devinheitmueller has quit [Quit: devinheitmueller]
Marth64 has joined #ffmpeg-devel
<Marth64>
hi good evening
<cone-774>
ffmpeg Andreas Rheinhardt master:08aa791d2038: avcodec/lpc: Don't use AAC defines directly
<cone-774>
ffmpeg Andreas Rheinhardt master:aaf26cffbac1: avcodec/lpc: Split inline functions into a header of their own
<cone-774>
ffmpeg Andreas Rheinhardt master:b9dea251d5ef: avcodec/aacdec: Move buffer to reduce padding
<cone-774>
ffmpeg Andreas Rheinhardt master:4ed0fea0fa61: avcodec/aactab: Deduplicate ltp_coef and tns_tmp2_map tables
justache has quit [Ping timeout: 255 seconds]
justache has joined #ffmpeg-devel
thilo has quit [Ping timeout: 246 seconds]
thilo has joined #ffmpeg-devel
Marth64 has quit [Ping timeout: 264 seconds]
Marth64 has joined #ffmpeg-devel
justache has quit [Ping timeout: 255 seconds]
justache has joined #ffmpeg-devel
Marth64 is now known as test
test is now known as Marth64
mkver has quit [Read error: Connection reset by peer]
mkver has joined #ffmpeg-devel
<Marth64>
i guess dvd needs seeking
<Marth64>
i will work on it. menu demuxing are done now but i will wait for other fixes merged before sending the patch
<Marth64>
but it seems like from usability perspective seeking and probing will be most handy
mkver has quit [Ping timeout: 256 seconds]
justache has quit [Ping timeout: 264 seconds]
Marth64 has quit [Ping timeout: 268 seconds]
Marth64 has joined #ffmpeg-devel
justache has joined #ffmpeg-devel
AbleBacon has quit [Read error: Connection reset by peer]
Marth64 has quit [Remote host closed the connection]
jamrial has quit []
Kei_N has quit [Read error: Connection reset by peer]
Kei_N_ has joined #ffmpeg-devel
justache has quit [Ping timeout: 264 seconds]
justache has joined #ffmpeg-devel
jarthur has quit [Quit: jarthur]
cone-774 has quit [Quit: transmission timeout]
Martchus_ has joined #ffmpeg-devel
Martchus has quit [Ping timeout: 256 seconds]
System_Error has quit [Remote host closed the connection]
System_Error has joined #ffmpeg-devel
rajivharlalka has joined #ffmpeg-devel
<rajivharlalka>
@JEEB I had been working on this bug (https://trac.ffmpeg.org/ticket/10692) and could understand the reason for assertion. Was able to log the nsamples and zeroes to find out that when the asendcmd starts atempo from 0.00 sec , the number of samples to be set to 0 gets equal to the number of the total samples in the buffer, which is unexpected. Though I couldn't understand why this was happening only when the filter
<rajivharlalka>
started from 0.00 sec.
System_Error has quit [Remote host closed the connection]
System_Error has joined #ffmpeg-devel
justache has quit [Ping timeout: 264 seconds]
justache has joined #ffmpeg-devel
<kierank>
(also as you can see we have active discussions here, the vote is IMHO
<kierank>
premature)
<kierank>
LOL
Kei_N_ has quit [Read error: Connection reset by peer]
Kei_N has joined #ffmpeg-devel
<elenril>
that CIA manual you linked to remains as relevant as ever
<JEEB>
wasn't it from while it was still called Office of Strategic Services?
<JEEB>
but yea, the manual's a good document. many things still hold.
<kierank>
It's interesting how some things don't need a long time to debate and some things conveniently do
System_Error has quit [Ping timeout: 260 seconds]
System_Error has joined #ffmpeg-devel
cone-166 has joined #ffmpeg-devel
<cone-166>
ffmpeg Martin Storsjö master:e30369bc1c68: aarch64: Use regular hwcaps flags instead of HWCAP_CPUID for CPU feature detection on Linux
<cone-166>
ffmpeg Andreas Rheinhardt master:f4df14b3545d: avutil/hwcontext_vdpau: Allocate public and priv device hwctx together
<cone-166>
ffmpeg Andreas Rheinhardt master:ac86b52f802b: avutil/hwcontext_vdpau: Don't use AVHWFramesInternal.priv
devinheitmueller has quit [Quit: devinheitmueller]
Krowl has quit [Read error: Connection reset by peer]
jarthur has joined #ffmpeg-devel
devinheitmueller has joined #ffmpeg-devel
MikhailAMD has joined #ffmpeg-devel
devinheitmueller has quit [Quit: devinheitmueller]
blb has quit [Quit: brb]
devinheitmueller has joined #ffmpeg-devel
mateo` has joined #ffmpeg-devel
blb has joined #ffmpeg-devel
Marth64 has joined #ffmpeg-devel
<mkver>
Is Stefano Sabatini on IRC?
<Marth64>
I don't think so afaik
<Marth64>
did I break something again
<mkver>
Marth64: This is not about you.
<Marth64>
cool
<cone-166>
ffmpeg Andrew Sayers master:6096c244bbd4: fix /// comments that should be ///<
<cone-166>
ffmpeg Andrew Sayers master:899e022cd7b2: avcodec/ivi: make comments more Doxygen-friendly
<cone-166>
ffmpeg Andrew Sayers master:0dde5c5efaef: avcodec/avcodec: downgrade multi-member comment to non-Doxygen comment
<cone-166>
ffmpeg Andrew Sayers master:dea1d7531d02: fix "param"s that should be "@param"
<cone-166>
ffmpeg Andrew Sayers master:b47b2c5b9125: fix "@param foo[in/out]" to "@param[in, out] foo"
<BtbN>
Marth64: btw., repeating my question about the dvd stuff from the other day: Why does configure check pkg-config, and on the same line supplies the -ldvdnav manually?
<mkver>
Too late...
<Marth64>
BtbN: had not seen the question, sorry. I had wrote that at least a year ago now, let me analyze and get back here.
<Marth64>
BtbN: builds fine without the -l on Linux at least, will test on Mac then email up the patch
<BtbN>
honestly, that patch is probably just fine to straight up push
<jdek>
mkver: shouldn't matter what's outside the initialised part as long as both buffers are the same. Whats required for valgrind to pass? just zero out buf0/buf1 beforehand?
<mkver>
Failing at link time (for non-Android, not for Android without JNI).
<mateo`>
Should we also do that for the other jni function ? It's only used on Android for now.
<mateo`>
Well in theory, JNI could be used on desktop platforms ...
<mkver>
We can't do this; legacy, you know.
rvalue has quit [Ping timeout: 256 seconds]
<Marth64>
BtbN: it ended up working on Mac. I am not familiar enough on Windows to test there. Sending up the patch now
<Marth64>
Interestingly, I get this warning at the end of the build: `ld: warning: ignoring duplicate libraries: '-lm', '-lz'` although it does not appear related to my changes
rvalue has joined #ffmpeg-devel
rajivharlalka has quit [Quit: Connection closed for inactivity]
<mateo`>
mkver: thx for the clarifications !
<Marth64>
I don't have ability to push but it's up for review.
ngaullier has quit [Ping timeout: 264 seconds]
<cone-166>
ffmpeg Marth64 master:59981b47db43: configure: do not specify -ldvdread, as pkg-config check is already done
<Marth64>
nice
devinheitmueller has quit [Quit: devinheitmueller]
Marth128 has joined #ffmpeg-devel
Marth64 has quit [Killed (NickServ (GHOST command used by Marth128!~Marth128@188.215.95.177))]
Marth128 is now known as Marth64
kurosu has quit [Quit: Connection closed for inactivity]
<JEEB>
with a quick check nothing in that file seems to utilize av_bprint correctly, so I'm going to skip noting anything about that (at least in theory you need to check if the bprint is complete). for the rest, were identifiers with spaces OK in that stuff?
<jamrial>
JEEB: there are line breaks in the compact output where there shouldn't without that patch
<JEEB>
yea that I agree with that that problem gets improved :)
<jamrial>
i think spaces in values are ok in compact output
<jamrial>
they are at least in xml
<JEEB>
ah it's values and not keys or whatever?
<jamrial>
fun thing, not all output formats print the .get_type() string
<JEEB>
yea
<jamrial>
JEEB: re av_bprint, the strings are pretty short, so they should fit in the stack buffer, with no reallocs happening
<JEEB>
yea
Marth128 has joined #ffmpeg-devel
Marth64 has quit [Ping timeout: 264 seconds]
<haasn>
jamrial: seems inconsistent to add a frame_side_data prefix to component/piece but not to timecode_list
<jamrial>
haasn: unique_name is when there's more than one element with the same name, afaik
<jamrial>
nothing else uses timecode
<jamrial>
but there's pixel_format using component
<haasn>
I see
Livio has joined #ffmpeg-devel
Krowl has joined #ffmpeg-devel
<Lynne>
I wanna write assembly
<jamrial>
there's lots missing for vvc
<jamrial>
elenril: added to my bump branch
Marth128 has quit [Ping timeout: 256 seconds]
Krowl has quit [Read error: Connection reset by peer]
cone-166 has quit [Quit: transmission timeout]
devinheitmueller has joined #ffmpeg-devel
devinheitmueller has quit [Client Quit]
jamrial_ has joined #ffmpeg-devel
tmm1_ has joined #ffmpeg-devel
mark4o has joined #ffmpeg-devel
markh has quit [Read error: Connection reset by peer]
jamrial has quit [Read error: Connection reset by peer]
mark4o is now known as markh
tmm1 has quit [Ping timeout: 272 seconds]
devinheitmueller has joined #ffmpeg-devel
MikhailAMD has quit [Quit: Leaving]
ngaullier has joined #ffmpeg-devel
devinheitmueller has quit [Client Quit]
devinheitmueller has joined #ffmpeg-devel
devinheitmueller has quit [Client Quit]
<JEEB>
elenril: I think in some previous iteration I had a different function to go over an array of things in general, and got comments about that so now it's in avcodec
<JEEB>
because I don't want to believe ffmpeg is going to be the only API client ever needing to for-loop over an array of side data and cloning it to another array
devinheitmueller has joined #ffmpeg-devel
<JEEB>
(since the first comment I received is "just have ffmpeg do the loopy-loop)
<JEEB>
so its current state is due to earlier reviews
<elenril>
that does not answer the question
<JEEB>
OK, so the first versions IIRC had it as a generic take array and take another array kind of thing
<JEEB>
then I got reviews that why is this not just a loop within ffmpeg so that there is no public symbol
<JEEB>
I objected on the premise that if ffmpeg needs to do this, then probably someone else would have to as well
<JEEB>
and the reviewer was happy putting it into avcodec
<elenril>
and so you made into a less general public symbol?
<JEEB>
because the reviewer seemed to be happy with that?
<elenril>
I fail to see the advantage
devinheitmueller has quit [Quit: devinheitmueller]
<JEEB>
well, if the original thing was bad then something needed to be done to it to applease the reviewer
<JEEB>
just like right now I'm going through the Nth iteration to appease some bits that you mentioned
<JEEB>
where I then propose that if this is bad, OK - is it better if I add a func specifically for avcodec
<JEEB>
yes
<JEEB>
you only objected to that at v5
<JEEB>
in the winter
<JEEB>
this is august
<jamrial_>
so mkver was the one that told you to make it avcodec specific?
<JEEB>
it happened between v4 and v5 where I probably had discussions with you and mkver most likely, since v5 is the first one where I then switched from extending sets to having an avcodec helper
ngaullier has quit [Ping timeout: 264 seconds]
<JEEB>
but yes, most likely I was discussing it then on IRC or otherwise with you and mkver
<JEEB>
trying to figure out what would be acceptable
<JEEB>
I wonder if this kind of back-and-forth is why I should just more easily push stuff that hasn't received a clear NAK :P me trying to do due dilligence just means I get thrown around
devinheitmueller has joined #ffmpeg-devel
<JEEB>
and while you get that dopamine rush at first that yay you made a FATE test that now tests new stuff and reads side data, at this point the enjoyment has rubbed out pretty well
<jamrial_>
i don't recall suggesting that, at least
<jamrial_>
personally, i don't think this needs an avcodeccontext specific function. you just have to call the add() function (from_sd, from_buf, or however it ends up being called) on avctx->decoded/raw_side_data
<jamrial_>
JEEB: nobody is trying to bikeshed your patches. i think that what happens is that a bit too much time passes between each iteration, and people forget what they suggested before, or new people comment
<JEEB>
yes, because I have limited brain power to then make the next version
<JEEB>
and/or since something is tiring I try to make something smaller that is actually enjoyable again
devinheitmueller has quit [Client Quit]
<JEEB>
and then I get the """enjoyment""" trying to figure out the context myself and attempting to apply as much of reviews received as possible
<jamrial_>
so what are the outstanding requests right now? remove the avctx specific function, plus some renames?
<JEEB>
renames I was just going through, when I noticed that comment that suddenly the avctx option was bad
<JEEB>
then not sure which nullptr etc checks are not required I haven't modified those either, just going through the stuff I can more or less agree with
<jamrial_>
removing the avctx function should be a matter of copying the relevant code from it to ffmpeg_enc.c
<JEEB>
but if ffmpeg needs to do it then most likely other things need to have that for loop as well...
<jamrial_>
if a similar loop ends up being duplicated in several places, then such a function can be added (that's not limited to avctx)
<JEEB>
so we first attempt to search for 3rd party API clients that would utilize this, and then add it?
<JEEB>
as opposed to the thing that you wanted which has zero users so far :P
<JEEB>
(internally evne)
<JEEB>
eh, sorry for putting it like that. could have worded it better.
<JEEB>
it has various renames and other comments applied
<JEEB>
not doing the avcodec change yet since I don't want to touch it right now. also it annoys me that I first add an extension function and I already once reworked it. then in winter I get told that the struct is not good. and then now apparently I need to go back to extending
<JEEB>
and I bet me being annoyed just makes it more annoying to others to read my crap :V
devinheitmueller has quit [Client Quit]
devinheitmueller has joined #ffmpeg-devel
<Daemon404>
it's the classic ffmpeg review cycle
devinheitmueller has quit [Quit: devinheitmueller]
devinheitmueller has joined #ffmpeg-devel
tonofclay has joined #ffmpeg-devel
tonofclay has quit [Remote host closed the connection]
hamzah has joined #ffmpeg-devel
devinheitmueller has quit [Quit: devinheitmueller]
devinheitmueller has joined #ffmpeg-devel
devinheitmueller has quit [Quit: devinheitmueller]
hamzah has quit [Quit: Client closed]
hamzah has joined #ffmpeg-devel
cone-569 has joined #ffmpeg-devel
<cone-569>
ffmpeg James Almer master:d78c75fe49a8: avformat/mov: fix setting disposition for the first iamf stream
<cone-569>
ffmpeg James Almer master:8f33d8fa9b11: avformat/mov: don't mark an item referenced by a grid as dependent if it's the primary item
<cone-569>
ffmpeg James Almer master:97e1480c8cb8: avformat/mov: set stream index field in tile grids