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
<compnn>
according to wikipedia, OK means oll korrect
<compnn>
thardin, make a wrapper to melt instead :)
khrbtxyz has quit [Ping timeout: 260 seconds]
<compnn>
ffmelt
mkver has joined #ffmpeg-devel
<ePirat>
michaelni, libavcodec/sonic.c:170:19: warning: unused function 'intlist_write' [-Wunused-function]
<ePirat>
170 | static inline int intlist_write(RangeCoder *c, uint8_t *state, int *buf, int entries, int base_2_part)
<ePirat>
| ^~~~~~~~~~~~~
<ePirat>
I quickly looked at sonic.c but now my eyes are bleeding from the bizarre ifdefery in there
iive has quit [Quit: They came for me...]
mkver has quit [Ping timeout: 252 seconds]
mkver has joined #ffmpeg-devel
cone-385 has quit [Quit: transmission timeout]
minimal has quit [Quit: Leaving]
Chagalle has joined #ffmpeg-devel
Chagall has quit [Ping timeout: 252 seconds]
Martchus_ has joined #ffmpeg-devel
Martchus has quit [Ping timeout: 265 seconds]
jamrial has quit []
novaphoenix has quit [Quit: i quit]
novaphoenix has joined #ffmpeg-devel
Richardcavell_ has joined #ffmpeg-devel
System_Error has quit [Remote host closed the connection]
TheVibeCoder has joined #ffmpeg-devel
System_Error has joined #ffmpeg-devel
wellsakus has joined #ffmpeg-devel
markh has quit [Read error: Connection reset by peer]
markh has joined #ffmpeg-devel
<thardin>
compnn: that's kind of how I usually do things
<thardin>
some devs appear to have forgotten the magic of the shell
<TheVibeCoder>
GUI
mkver has quit [Ping timeout: 260 seconds]
mkver has joined #ffmpeg-devel
mateo` has quit [Read error: Connection reset by peer]
mateo` has joined #ffmpeg-devel
dariusz has joined #ffmpeg-devel
<TheVibeCoder>
pross: get_bits_le() does not work?
<pross>
thx TheVibeCoder! today i learned something new
Teukka has quit [Read error: Connection reset by peer]
Teukka has joined #ffmpeg-devel
Teukka has quit [Changing host]
Teukka has joined #ffmpeg-devel
Richardcavell_ has quit [Ping timeout: 265 seconds]
Richardcavell_ has joined #ffmpeg-devel
jamrial has joined #ffmpeg-devel
wellsakus has quit [Quit: WeeChat 3.5]
wellsak27 has joined #ffmpeg-devel
wellsak27 is now known as wellsakus
Guest79 has joined #ffmpeg-devel
Guest79 has quit [Client Quit]
ccawley2011 has joined #ffmpeg-devel
<dariusz>
I have a question about FPS implementation in intra frame codec (APV). Raw bitstream does not contain any fps info and now we're setting some defatult value (30) for ffprobe for example. Is this approach is ok or it should have no fps info at all? I mean in raw bitstream encoding case, without container.
microchip_ has quit [Quit: There is no spoon!]
microchip_ has joined #ffmpeg-devel
ccawley2011 has quit [Ping timeout: 245 seconds]
minimal has joined #ffmpeg-devel
putacho has joined #ffmpeg-devel
rvalue has quit [Read error: Connection reset by peer]
rvalue has joined #ffmpeg-devel
microchip_ has quit [Read error: Connection reset by peer]
putacho is now known as microchip_
j45 has quit [Ping timeout: 276 seconds]
j45 has joined #ffmpeg-devel
j45 has quit [Changing host]
j45 has joined #ffmpeg-devel
Kei_N_ has joined #ffmpeg-devel
Kei_N has quit [Ping timeout: 248 seconds]
HarshK23 has joined #ffmpeg-devel
DVedaa3 has joined #ffmpeg-devel
<jamrial>
dariusz: i guess a framerate option could be added to the demuxer
<jamrial>
instead of hardcoding 30
ccawley2011 has joined #ffmpeg-devel
mkver has quit [Ping timeout: 244 seconds]
ccawley2011 has quit [Ping timeout: 260 seconds]
Kei_N has joined #ffmpeg-devel
Kei_N_ has quit [Read error: Connection reset by peer]
ccawley2011 has joined #ffmpeg-devel
Guest5460 has joined #ffmpeg-devel
Guest5460 is now known as lemourin
<fflogger>
[newticket] keks51: Ticket #11645 ([avcodec] pts_correction_num_faulty is not reset in ff_decode_flush_buffers) created https://trac.ffmpeg.org/ticket/11645
lemourin is now known as Guest5120
lemourin5 has joined #ffmpeg-devel
Guest5120 has quit [Killed (silver.libera.chat (Nickname regained by services))]
<compnn>
dariusz, if you are adding fps later when you put it in a container, then it doesnt matter. as long as you know the fps...
<ePirat>
Any reason to still support anything older than OpenSSL 1.1.0?
<ePirat>
(it was released 2016, so people using anything older should probably not...)
<BtbN>
I'd say on master it'd be fine to drop anything that no longet gets security updates anywhere
ccawley2011_ has quit [Ping timeout: 244 seconds]
<kasper93>
that ^, we should make it harder to build vulnerable software, even if it is dependency issue
<ePirat>
will prepare a patch then
<kasper93>
Versions 1.1.1 and 1.0.2 are no longer supported. Extended support for 1.1.1 and 1.0.2 to gain access to security fixes for those versions is available.
<kasper93>
Versions 1.1.0, 1.0.1, 1.0.0 and 0.9.8 are no longer supported.
<JEEB>
ah, repology is the one whose UI I recalled
<JEEB>
cheers
<JEEB>
if we go for example by ubuntu, there 20.04 was the last pre-3.0 LTS (as 3.0 was released in 2021), and it has run out of free support
<JEEB>
centos 8 is also out of support as of May 31, 2024
<JEEB>
so if we can get rid of a whole bunch of ifdeffery etc, limiting to 3.0+ is not that bad
<BtbN>
Ubuntu 20.04 has support until 2030
<kierank>
surely canonical does the backports for extended support
<JEEB>
yes, and I know you can technically enter that tier for a limited amount of systems for free (not sure if there were limits on if you were corporate entity VS private etc)
<JEEB>
but I would not really consider that standard support
<sfan5>
maybe it's intentionally misleading for some business reason ;)
<BtbN>
But yeah, in this case I'll have to do the mailman3 migration asap...
<JEEB>
on the page that BtbN linked I first jumped at the table "Ubuntu releases" and its hovering things
<BtbN>
cause mailman2 does not support python3, and Ubuntu has long dropped python2
<BtbN>
It'll also more than likely completely break postfix
<BtbN>
I had hoped I could wait until I could update all the way to 26.04 to do this
<JEEB>
it shows things for 20.04 as expanded security maintenance, which I assume is ESM. and then the May 2025 non-pro and Apr 2030 pro thing
<JEEB>
(pro including esm)
<sfan5>
well you can build py2 from source and it's not any worse "supported" than what 20.04 ships, but have fun with that
<BtbN>
That'd be a horrible hack that'd break on any other OS update. No.
<JEEB>
BtbN: but yea I can totally see by the "maintenance and security updates" part why you'd interpret things like that
Richardcavell_ has quit [Ping timeout: 252 seconds]
<BtbN>
The crazy part is, if I read that right, without paying, they will abandon the community repos from day 1
iive has joined #ffmpeg-devel
<BtbN>
and there's a lot of commonly used stuff in there
<JEEB>
yea, the actually officially supported part is not that large. stuff may get updates, but it also may not.
<JEEB>
I think in 20.04 you already had a command that could show a package's support state?
<BtbN>
I wonder if it'd be easier to just install an entirely new VM with Debian testing instead of Ubuntu. I think they recently went into freeze
<BtbN>
And then migrate everything ffmpeg.org to that, instead of trying to perform 6 years worth of upgrades on the current one
<JEEB>
does not sound like the worst of ideas
<BtbN>
There's a lot of random stuff collected on there as well
<BtbN>
and mailman seems to have been installed via "pip install" as root, so it's sprayed accross /usr/local and a lot of other places, virtually unmaintainable
<JEEB>
yea, I have one VM that has gone up from like 12.04 upwards, and I bet there's a magical buckload of stuff in there that is no longer required
<BtbN>
latest Ubuntu is an absolute shitfest anyway
<JEEB>
also that has been the only system where I've had so many successful ubuntu upgrades
<BtbN>
so a new Debian 13 machine wouldn't be the worst idea
ccawley2011_ has quit [Ping timeout: 260 seconds]
<JEEB>
on the other hand I've now been for like ten years on fedora and at this point I've gotten way too used to just doing an upgrade twice or once a year because it generally goes by without any drama. system reboots into upgrade mode for 10-20 minutes and boots right back up
<BtbN>
Well, server config file migration usually does not
<JEEB>
it even once failed during the upgrade and was able to revert...
<BtbN>
and there is so much random manual stuff done on that box, I don't think it'd come back up with anything but ssh running
<JEEB>
yea, config file migration is always fun
<BtbN>
specially for a mailserver
<BtbN>
half the components for stuff like dkim the current server uses are projects that have long died
<BtbN>
and are not available anymore on current distros
<JEEB>
yea, so figuring out things from scratch is really not a bad idea. adding automated templating of config files and other similar tasks. In other contexts I was utilizing ansible for that
<JEEB>
and yea, I've recently looked at someone else's deployment automation and stripped the hell out stuff like pip without a virtualenv
<BtbN>
a mailserver is also one of the things where putting it in docker does not help whatsoever
<BtbN>
it's so many moving parts that need to interact...
<BtbN>
on modern distros, they don't even let you run pip as root anymore
<BtbN>
cause of all the damage it can do
<BtbN>
on Gentoo it was a common theme that someone ran pip as root, and it uninstalled portage. The package manager.
<JEEB>
not even as root, at least on ubuntu 24 it will just not do it as normal user, either
<BtbN>
Yes, cause it's a security risk
<JEEB>
indeed
<BtbN>
the stuff installed into ~/.lib will override system site-packages
<BtbN>
and if you "sudo" something, the things in the user homedir still take effect
<JEEB>
and well, abused during the years to hell and back
<BtbN>
it's quite bad
leo60228 has quit [Read error: Connection reset by peer]
<JEEB>
I recall teaching certain people how to utilize a specific virtualenv for certain ansible tasks years ago
<JEEB>
but alas, that was not really utilized so I got to migrate stuff this year :P
<BtbN>
The latest hip thing seems to be "uv"
<BtbN>
cause of course when python needs a new tool, you write it in rust
vriska has joined #ffmpeg-devel
<TheVibeCoder>
michaelni: where is that file to reproduce infinite loop in filtering?
ccawley2011_ has joined #ffmpeg-devel
secondcreek has quit [Remote host closed the connection]
secondcreek has joined #ffmpeg-devel
<TheVibeCoder>
its soo hot, i can not code anything.
<ePirat>
also is it expected that -tls_verify 1 just leads to "certificate verify failed" when I try connect to anything?
<BtbN>
you need to point it to a CA store so it can verify
<ePirat>
why isnt it using my system one like it usually does?
<BtbN>
no idea
<BtbN>
Don't think it ever did that, outside of schannel on Windows
<ePirat>
well it works with the openssl command line tool
<wbs>
I think openssl got support for finding a system CA store at some point
<wbs>
and I faintly remember it being hooked up
ccawley2011_ has quit [Ping timeout: 272 seconds]
<kasper93>
BtbN: unless there is good upgrade path, doing it fresh and migrating just the configs, would indeed make sense. From what you said it would be nightmare to upgrade all those pip installed in root things
<BtbN>
well, it'd be much more than migrating just the configs
<BtbN>
the mailman3 upgrade is also very likely distruptive
<BtbN>
pipermail archives are no longer supported, they use hyperkitty now. And the links are different...
<ePirat>
BtbN, wbs, Ah, we never call SSL_CTX_set_default_verify_paths
<BtbN>
I think it might actually be documented to work like that
<BtbN>
But I think it's dumb, and if it's really just that, also easy to patch
<ePirat>
yeah
<ePirat>
I wondered if its intended to work like that but... it makes little sense IMHO
<ePirat>
at least if nothing else is set explicitly
<BtbN>
I've looked into adding dtls support to our schannel wrapper, but both ends of that are under-documented
<BtbN>
the new dtls interface on our end is, and dtls support in schannel is
<BtbN>
it's documented that it's supported. But that's about it
ngaullier has joined #ffmpeg-devel
Kei_N_ has joined #ffmpeg-devel
Kei_N has quit [Read error: Connection reset by peer]
ccawley2011_ has joined #ffmpeg-devel
ngaullier has quit [Remote host closed the connection]
zsoltiv has quit [Ping timeout: 260 seconds]
zsoltiv__ has joined #ffmpeg-devel
zsoltiv_ has quit [Ping timeout: 268 seconds]
zsoltiv__ is now known as zsoltiv
zsoltiv_ has joined #ffmpeg-devel
Guest97 has joined #ffmpeg-devel
ccawley2011_ has quit [Read error: Connection reset by peer]
Guest4 has joined #ffmpeg-devel
Guest4 has quit [Client Quit]
Guest97 has quit [Ping timeout: 272 seconds]
BradleyS has quit [Read error: Connection reset by peer]
BradleyS has joined #ffmpeg-devel
<BtbN>
Why is ff_url_read_all defined in tls.h?
<BtbN>
And tls.c
<fflogger>
[newticket] jorn: Ticket #11646 ([swscale] Full range RGB to YUV conversion using swscale has a slight green tint.) created https://trac.ffmpeg.org/ticket/11646
<BtbN>
the old archives from mailman2 are kinda iconic, would love to keep it
<beastd>
BtbN: IIRC thresh mentioned it's possible to keep using pipermail for the archives
<BtbN>
I can't find anything but dead links when looking for that
<beastd>
compnn: couldn't really find alternatives when i tried about half a year ago :(
<kasper93>
isn't hyperkitty just a web frontend?
<BtbN>
yes
<beastd>
though recently i stumbled, i think somehow via kernel.org, over mlmmj (which similar to ezmlm but for more MTAs than qmail): https://mlmmj.org/
<beastd>
maybe with public-inbox for the archives it could be an alternative to mailman3
* compnn
wonders when the web just abandoned accessibility html-only pages
<compnn>
blind people get fucked i guess :(
<BtbN>
Not like you can't make JS SPAs accessible
cone-806 has joined #ffmpeg-devel
<cone-806>
ffmpeg Michael Niedermayer master:0ffe97d9b9e1: avformat/mov: Check that sample_count is allocated in mov_parse_heif_items()
<compnn>
that hyperkitty thing annoys me. i'm going to file a bug on them