[Ffmpeg-devel-irc] ffmpeg-devel.log.20180304

[burek]

[00:25:28 CET] <cone-280> ffmpeg 03Philip Langdale 07master:dd3f1e3a1186: swscale: Introduce a helper to identify semi-planar formats
[02:07:38 CET] <jamrial> BtbN: no, see https://git.videolan.org/?p=ffmpeg.git;a=blob;f=configure;h=fe81ba31b5ed7b4cbee5b562c59eba955727777a;hb=HEAD#l3774
[02:07:50 CET] <jamrial> re zlib and such
[10:42:29 CET] <BtbN> hm "not disabled" should still work just fine though, or should I change the patch?
[11:04:29 CET] <nevcairiel> BtbN: components have a "neutral" state by default where both "enabled" and "not enabled" would return false, so basically, check that auto-detection still behaves as intended, that it doesn't end up being enabled anywhere it shouldn't be - or simpler, do what everything else does =p
[11:05:38 CET] <BtbN> well, there are examples for both, and with ! disabled it works as intended auto-detection wise
[11:05:54 CET] <BtbN> The non-supported platforms hard-disable it later in configure anyway
[11:09:50 CET] <nevcairiel> also that the disable-autodetect (or whatever that name was) switch keeps it disabled
[11:11:23 CET] <BtbN> it being in the hwaccel autodetect list should achive that
[11:34:08 CET] <BtbN> --disable-autodetect --enable-nvenc breaks though, unless you also add --enable-ffnvcodec manually
[11:43:57 CET] <gagandeep_> guys, i stumbled across the settings in vlc media player for deinterlacing, does the vlc media player post process the interlaced frames after decoding or is it the libav decoder doing the deinterlacing before passing it to the vlc
[11:44:18 CET] <BtbN> decoders, usually, don't deinterlace
[11:47:06 CET] <gagandeep_> does ffplay has a deinterlacing option like in vlc
[11:51:50 CET] <gagandeep_> nevermind its -vf <filter name>
[12:04:02 CET] <gagandeep_> kierank: i now understand the wavelet filter algorithm used after studying it in the wavelet.c and utils.c given in the cineform-sdk and have more understanding of cfhd.c, so how should i go about the bugs listed in the qualification tasks
[12:11:52 CET] <gagandeep_> atomnuker: i haven't been able to get a reply from kierank, so i would like to get some pointers on how to start working on the bugs listed in the qualification task of cineform decoder
[13:06:31 CET] <durandal_1707> gagandeep: by picking one of the bugs and fixing it...
[15:26:21 CET] <atomnuker> actual leaked atrac9 design meeting:
[15:26:28 CET] <atomnuker> "lets fuck everything up!"
[15:26:36 CET] <JEEB> :D
[15:26:37 CET] <atomnuker> "but how could we possibly do that?"
[15:26:54 CET] <atomnuker> "lets stick a shitload of frames into a packet and call it a superframe!"
[15:27:16 CET] <atomnuker> "oh and lets pack that up in .wav container, it rocks!11! lol"
[15:27:25 CET] <JEEB> also into mpeg-ps
[15:27:31 CET] <JEEB> they were using both in various places
[15:27:56 CET] <durandal_1707> you are now rock stars!
[15:29:01 CET] <thardin> .wav is a container so that part isn't so strange
[15:31:09 CET] <atomnuker> durandal_1707: I can decode and dequantize the first frame perfectly, thankfully the huffman tables aren't sparse so just plugging them into init_vlc works
[15:31:28 CET] <atomnuker> but the thing desyncs after the first frame's 5 or so blocks have been decoded
[15:33:01 CET] <durandal_1707> atomnuker: how do you split packets?
[15:34:42 CET] <atomnuker> in theory the wav demuxer should give me superframes, and in each superframe there ought to be a specified (4 in my case) in the extradata number of frames
[15:35:02 CET] <atomnuker> the frame size is constant throughout, in my case 112 bytes per frame
[15:35:12 CET] <atomnuker> but the packets I get are 4032 bytes big
[15:35:32 CET] <atomnuker> which means there are 36 frames in a single packet
[15:36:15 CET] <durandal_1707> atomnuker: do you set block align to reasonable value in wav demuxer?
[15:36:27 CET] <atomnuker> what's that?
[15:37:09 CET] <durandal_1707> the superframe size in bytes
[15:38:22 CET] <durandal_1707> 4096 is default value
[15:38:27 CET] <durandal_1707> when unset
[15:38:54 CET] <atomnuker> how do I change that? I just added support for atrac9's nonstandard guid
[15:39:03 CET] <atomnuker> (its 32 bits and not 16 bits like everything else)
[15:39:35 CET] <nevcairiel> how does a bigger thing even fit into the container
[15:41:07 CET] <durandal_1707> 32 bytes you mean?
[15:41:26 CET] <atomnuker> no, see ff_codec_wav_tags[]
[15:41:38 CET] <atomnuker> everything there has a 16 bit identifier
[15:41:46 CET] <nevcairiel> wav tags are TwoCC, the field in the container is only 16-bit, so how would they fit any more
[15:41:51 CET] <atomnuker> but to make it work with atrac9 I had to put in 0x47E142D2
[15:41:59 CET] <atomnuker> which is 32 bits
[15:42:04 CET] <durandal_1707> atomnuker: you are supposed to only set ff_codec_wav_guids
[15:42:25 CET] <atomnuker> well I had to set that + parse_waveformatex() to make it work
[15:42:25 CET] <durandal_1707> +    { AV_CODEC_ID_ATRAC9,   { 0xD2, 0x42, 0xE1, 0x47, 0xBA, 0x36, 0x8D, 0x4D, 0x88, 0xFC, 0x61, 0x65, 0x4F, 0x8C, 0x83, 0x6C } },
[15:43:49 CET] <atomnuker> https://github.com/atomnuker/FFmpeg/commit/b1c800a0c8d6b44c86a5604ba2b0fd3b4eddbd0e
[15:44:01 CET] <atomnuker> that's what I did
[15:45:15 CET] <nevcairiel> shouldnt the e lse cause capture the guid if you put it into ff_codec_wav_guids
[15:46:14 CET] <nevcairiel> no need to mess around with the too long twocc field then
[15:46:22 CET] <durandal_1707> yes, https://github.com/richardpl/FFmpeg/commit/655780387ea019ec15e798504b9ff42ff718867c
[15:51:50 CET] <atomnuker> yeah, that works, though I'm still getting huge frames
[15:53:34 CET] <durandal_1707> https://github.com/FFmpeg/FFmpeg/blob/master/libavformat/wavdec.c#L564
[15:57:32 CET] <atomnuker> there's a block align chunk in the container apparently
[15:57:33 CET] <atomnuker> https://github.com/Thealexbarney/VGAudio/blob/master/docs/audio-formats/atrac9/container.md#fmt-chunk
[15:59:17 CET] <durandal_1707> atomnuker: yes, so parse it and store it and use it in wav_read_packet()
[15:59:18 CET] <atomnuker> apparently its even used, in riffdec.c (par->block_align = avio_rl16(pb);)
[15:59:34 CET] <atomnuker> what's the difference between riff and wav?
[16:00:24 CET] <nevcairiel> its the same
[16:00:49 CET] <durandal_1707> atomnuker: apparently it will take multiples of block_align that fit into 4096
[16:01:07 CET] <atomnuker> so we have 2 wav demuxers? which one is being used?
[16:01:19 CET] <nevcairiel> RIFF is the underlying format that powers both wav and avi
[16:01:35 CET] <nevcairiel> so wavdec probably uses riffdec?
[16:01:44 CET] <JEEB> quite likely
[16:02:16 CET] <atomnuker> well if it does that it should hangle block aligning, no? riffdec sets it
[16:02:42 CET] <durandal_1707> block align is not same as block size
[16:03:01 CET] <durandal_1707> it handles it in wav_read_packet in libavformat/wavdec.c
[16:04:23 CET] <atomnuker> yeah, I can see that
[16:04:28 CET] <atomnuker> so wavdec uses riffdec
[16:04:43 CET] <atomnuker> riffdec (I checked) correctly sets codecpar->block_align from the bitstream
[16:04:59 CET] <atomnuker> but why isn't wavdec using it?
[16:05:27 CET] <durandal_1707> i told you that its used in wav_read_packet
[16:06:11 CET] <durandal_1707> if block_align is 512, it will still give 4096 packets, because 4096 is multiple of 512 last time I checked
[16:09:16 CET] <atomnuker> sorry, forgot that block_align != block_size
[16:09:27 CET] <atomnuker> any way to set the block size or should I just add a parser to do that
[16:15:49 CET] <durandal_1707> atomnuker: make atrac9 do like xma,wmalossless: .capabilities   = AV_CODEC_CAP_SUBFRAMES ?
[16:17:46 CET] <atomnuker> nah, that's last resort
[16:18:00 CET] <atomnuker> (IMO that should be deprecated and removed)
[16:19:04 CET] <atomnuker> well, it will have to be because once the new decoding api is the only one that flag will be meaningless
[16:19:50 CET] <durandal_1707> cant you just return exactly that many bytes you used?
[16:19:51 CET] <atomnuker> we'll have to rewrite every decoder pretty much anyway to deal with the old api disappearing
[16:20:10 CET] <atomnuker> I can return kinda big block sizes, that's not a problem
[16:20:33 CET] <atomnuker> I mean frame size
[16:21:28 CET] <durandal_1707> writing parser to just split into block_align seems overkill...
[16:27:29 CET] <atomnuker> well, having a parser would remove the need to decode an entire frame to get things like samplerate, which is more of an overkill I think
[17:46:10 CET] <atomnuker> JEEB / durandal_1707 / whoever has a native windows atrac decoder on hand: can you decode https://0x0.st/smm3.at9 and tell me how long it is
[18:04:56 CET] <RiCON> atomnuker: 00:02:53.05
[18:05:19 CET] <atomnuker> damn, I'm getting 11 seconds here
[18:06:03 CET] <RiCON> want an upload of the decode?
[18:07:25 CET] <RiCON> https://0x0.st/smmg.flac
[18:54:59 CET] Action: Compn drops troll grenade
[18:55:06 CET] Action: Compn runs
[18:57:59 CET] Action: techkid6 chases
[19:01:48 CET] <Compn> BBB : also i would argue its not even a political issue
[19:01:51 CET] <Compn> its the internet itself
[19:02:10 CET] <Compn> i think some companies have paid trolls to make it political to split people from the real issue
[19:02:37 CET] <durandal_1707> atomnuker: you receive all packets from demuxer and decode full data?
[19:03:12 CET] <durandal_1707> Compn: one of paid trolls is here, and guess who it is?
[19:03:31 CET] <atomnuker> yep, I decode everything but I have trouble figuring out why if I decode all frames I run out of bits
[19:03:48 CET] <atomnuker> it should be CBR and there shouldn't be overreading but there's a crash in entropy decoding
[19:04:15 CET] <durandal_1707> crash? you use unchecked bitreader?
[19:04:18 CET] <JEEB> anyways, the problem was re-adding the darn thing back to ffmpeg.org, which I am still strongly against even if that entity does keep its JS under control
[19:04:51 CET] <JEEB> at the very least the crap doesn't get added to the doxygen, but ffmpeg-all.html for example is something I utilized daily
[19:04:56 CET] <JEEB> or almost daily
[19:05:07 CET] <Compn> right
[19:05:11 CET] <JEEB> and the darn thing misbehaving was darn annoying, and I was happy for it get the heck out
[19:05:15 CET] <Compn> i agree should only happen on index page
[19:05:16 CET] <atomnuker> durandal_1707: no, I'm sending the result to the other decoder (to figure out where mine desyncs) which crashes
[19:05:17 CET] <Compn> if at all.
[19:05:25 CET] <Compn> not ffmpeg-all
[19:05:40 CET] <Compn> banner / news would be better
[19:17:01 CET] <BtbN> Wait, is Coverity dead?
[19:17:17 CET] <BtbN> https://scan.coverity.com/ the SSL error here is why the script just exits quietly
[19:17:33 CET] <Compn> are you using chrome ?
[19:17:41 CET] <BtbN> The cert is obviously bad...
[19:17:55 CET] <BtbN> for a whole different domain
[19:18:08 CET] <JEEB> curl: (51) SSL: no alternative certificate subject name matches target host name 'scan.coverity.com'
[19:18:12 CET] <JEEB> yup
[19:18:55 CET] <BtbN> https://www.synopsys.com/software-integrity/resources/datasheets/coverity.html it still seems to exist, but... it's broken?
[19:39:25 CET] <Compn> send mail to coverity contact
[19:39:27 CET] <Compn> ?
[19:39:59 CET] <BtbN> if you click the link in one of the older mail reports, you end up at https://brb.synopsys.com/
[19:47:13 CET] <Compn> ah
[19:48:28 CET] <Compn> https://twitter.com/CoverityScan
[19:48:49 CET] <Compn> scan.coverity.com redirects to the brb page now
[19:56:42 CET] <BtbN> that's from over 2 weeks ago, wow
[19:58:29 CET] <Compn> synopsys twitter has no info 
[19:58:47 CET] <Compn> someone should try private messaging them
[20:00:17 CET] <BtbN> I doubt that'll do anything useful
[20:00:23 CET] <BtbN> They are clearly aware
[20:02:59 CET] <nevcairiel> the coverityscan twitter has respones from the last couple days, its just down for the time being
[20:03:21 CET] <nevcairiel> last from March 1
[20:17:02 CET] Action: Compn cant remember if timo is on irc
[20:17:30 CET] <atomnuker> BtbN?
[20:17:36 CET] <Compn> ok yeah i was thinking that
[20:17:39 CET] <BtbN> hm?
[20:17:50 CET] <Compn> BtbN : why not just add a line to the error message of --enable-cuda explaining which lib user needs ?
[20:18:00 CET] <Compn> instead of renaming
[20:18:04 CET] <BtbN> Because it's a generic error message
[20:18:27 CET] <BtbN> it just goes "dependency *dependency_name* not found"
[20:18:49 CET] <Compn> no way to customize that error message ?
[20:18:58 CET] <BtbN> that would end up with a horrible mess of special cases
[20:19:00 CET] <Compn> catch it with a script or something
[20:19:01 CET] <Compn> hmm
[20:19:16 CET] <Compn> can we catch it at the start of configure with --enable-cuda ?
[20:19:20 CET] <BtbN> and not calling it cuda is more accurate as well
[20:19:25 CET] <Compn> i know and i agree
[20:19:39 CET] <Compn> i'm just wondering how to catch it with --enable-ffnvcodec
[20:19:40 CET] <Compn> heh
[20:19:49 CET] <BtbN> you never call configure with enable-cuda/ffnvcodec though.
[20:19:59 CET] <BtbN> it's autodetect, you only ever disable it
[20:20:19 CET] <BtbN> the case when it explodes is if someone does --enable-nvenc, and then it complains about nvenc enabled but required dependency missing: cuda
[20:20:19 CET] <Compn> er --enable-nvenc *
[20:20:20 CET] <nevcairiel> thats not really true
[20:20:33 CET] <nevcairiel> i use builds with --disable-autodetect and then explicitly turn shit on i want
[20:20:42 CET] <Compn> i mean we can break on script if they --enable-nvenc without the correct dependency
[20:20:45 CET] <Compn> nevermind 
[20:20:49 CET] <Compn> i give ups
[20:45:02 CET] <kierank> I'm glad I have Compn on ignore
[20:45:20 CET] <durandal_1707> lol
[20:51:11 CET] <BtbN> I mean, I totally can just add a special case for "if missing dep == cuda, print this instead: ..."
[20:51:17 CET] <BtbN> But I'm not sure if that's better?
[20:59:42 CET] <BBB> Compn: Im not sure that makes it different - in fact it may make my point more valid
[20:59:57 CET] <BBB> Compn: my point isnt that what were doing as a project is in itself invalid or untrue
[21:00:36 CET] <Compn> double negative in the sentence makes it hard to parse
[21:00:41 CET] <BBB> sorry
[21:00:58 CET] <BBB> let me try to explain by showing the other side
[21:00:59 CET] <Compn> no worries, please rephrase :)
[21:01:01 CET] <BBB> maybe then it makes sense
[21:01:12 CET] <Compn> sure
[21:01:45 CET] <BBB> lets assume that you are right (or not, Im not sure it matters), and that there are a substantial subset of people (be it on the other side of the political spectrum alone, or just in general), and they believe that we are hippies and that net neutrality is not a big deal
[21:01:57 CET] <BBB> with me so far?
[21:02:15 CET] <Compn> i need to see evidence that net neutrality is a political spectrum issue
[21:02:39 CET] <BBB> hm& ok, Im not sure I have that at this point
[21:02:46 CET] <BBB> can you imagine?
[21:02:49 CET] <Compn> for me to accept your argument, and be able to argue it honestly.
[21:02:52 CET] <BBB> or is evidence required?
[21:03:22 CET] <BBB> I havent prepared evidence, I didnt know this was going to be a trial
[21:03:22 CET] <Compn> without evidence this could just be your opinion, if that makes sense ?
[21:03:26 CET] <BBB> sure
[21:03:31 CET] <BBB> ok, so let it be my opinion then
[21:03:35 CET] <Compn> sure ok then
[21:03:42 CET] <BBB> can you imagine that theres people out there with such an opinion?
[21:03:53 CET] <BBB> so, now lets assume that we put out such a statement on our website
[21:03:57 CET] <Compn> no. i think there are a few people paid by verizon to stir shit :D
[21:04:13 CET] <BBB> my fear is that we basically chase half of the electorate away from contributing to (or caring about, or even willing to consider using) ffmpeg
[21:04:16 CET] <Compn> everyone else wants the internet to continue as it always has
[21:04:41 CET] <BBB> maybe it is a minority, I dont know
[21:04:48 CET] <BBB> but heres the thing: I dont want to argue that on our website
[21:04:57 CET] <BBB> I dont want our website to be the forum where that issue is decided
[21:05:02 CET] <BBB> let me take it one step further
[21:05:07 CET] <Compn> i mean, there are a minority of people who believe in flat earth. 
[21:05:09 CET] <BBB> you know guns? and abortion? and evangelism?
[21:05:15 CET] <BBB> and gays used to be an issue too
[21:05:19 CET] <BBB> Im not sure it still is
[21:05:22 CET] <BBB> maybe it is, maybe not
[21:05:27 CET] <BBB> but each of these were sway issues
[21:05:34 CET] <jamrial> Compn: for real, drop it
[21:05:44 CET] <BBB> single-issues that could entirely on their own decide a persons vote, each having a substantial number of voters behind them
[21:05:54 CET] <BBB> make net neutrality something like that!
[21:05:57 CET] <Compn> jamrial : are you tired of free speech ?
[21:05:57 CET] <BBB> I think it should be
[21:06:04 CET] <jamrial> Compn: we don't need, as you called it, more trolls throwing bombs
[21:06:05 CET] <BBB> but then ever more, we should not argue it on our website
[21:06:10 CET] <BBB> because I believe we should be apolitical
[21:06:16 CET] <BBB> I want verizon to use our software
[21:06:19 CET] <BBB> even their managers
[21:06:24 CET] <BBB> and also lefty hippies
[21:06:26 CET] <BBB> all of them
[21:06:28 CET] <Compn> i dont agree with your position that it is political
[21:06:39 CET] <Compn> but ok, you are worried about upsetting someone
[21:06:41 CET] <BBB> thats ok, I understand and agree I didnt back that up
[21:06:41 CET] <Compn> i get that
[21:06:47 CET] <rcombs> kinda hard to be apolitical when you're inherently anti-patent
[21:06:54 CET] <BBB> I think its the wrong place to upset people
[21:06:57 CET] <BBB> the forum is wrong
[21:07:03 CET] <BBB> I want to use another forum to upset these people
[21:07:06 CET] <BBB> ideally voting booths
[21:07:09 CET] <BBB> but any other place is fine also
[21:07:17 CET] <BBB> yes, I think most software patents are dumb
[21:07:19 CET] <Compn> rcombs : i said software patents
[21:07:21 CET] <BBB> of course I do
[21:07:37 CET] <BBB> but I doubt that banners on our website will do us any good
[21:07:41 CET] <BBB> make it a voting issue
[21:07:45 CET] <BBB> get people to vote
[21:07:47 CET] <Compn> regular patents are different in most cases, although rounded corners on cell phone designs seem like not such a patentable idea but who am i to judge
[21:07:58 CET] <rcombs> that's the point of the banners, though
[21:08:29 CET] <Compn> to bring people to see it as a voting issue, yes, that is the point of hte banner
[21:08:33 CET] <BBB> its true that were anti-patent, but I also believe that in some ways were orthogonal to them
[21:08:40 CET] <BBB> for example, chrome uses our h264 decoder
[21:08:46 CET] <BBB> yet they do pay h264 patent royalties
[21:08:47 CET] <BBB> so&
[21:08:56 CET] <BBB> in some way we are not fundamentally incompatible with patents, right?
[21:09:01 CET] <BBB> we dislike them, but it can still work
[21:09:07 CET] <BBB> so in some way I am OK with that
[21:09:08 CET] <Compn> we could be sued for violating patents, BBB
[21:09:25 CET] <Compn> it just hasnt happened YET, probably because we have no money
[21:09:31 CET] <BBB> ikr
[21:09:35 CET] <BBB> :)
[21:10:15 CET] <BBB> I think its great that youre upset about NN - and please vote the right party and people so that it becomes law
[21:10:28 CET] <BBB> I would be really happy if NN was established in law
[21:10:50 CET] <BBB> but I prefer that we dont use our website for that
[21:11:02 CET] <BBB> in the ideal world, we are apolitical
[21:11:09 CET] <Compn> who said i was upset about NN ?
[21:11:10 CET] <BBB> republicans use ffmpeg, democrats use ffmpeg
[21:11:15 CET] <BBB> gun owners use ffmpeg, gun haters use ffmpeg
[21:11:24 CET] <BBB> abortionistas and pro-lifers use ffmpeg
[21:11:27 CET] <BBB> everyone uses ffmpeg
[21:11:29 CET] <BBB> everyone loves ffmpeg
[21:11:31 CET] <Compn> you've been living in usa too long if thats how you think in binary terms :D
[21:11:32 CET] <BBB> thats the ideal world
[21:11:47 CET] <rcombs> so is your specific issue that you're worried that verizon or other ISPs will avoid ffmpeg if the website has net neutrality banners
[21:11:48 CET] <BBB> probably true, yes
[21:12:07 CET] <BBB> rcombs: no, I dont want us to be positioned along the political spectrum at all
[21:12:12 CET] <BBB> as a project
[21:12:39 CET] <Compn> your argument is that we could be targeted by the politicals if so ?
[21:12:39 CET] <BBB> in all honesty, if we literally pissed on them, they probably would still use use, largely because they have no alternate choice
[21:12:43 CET] <BBB> as silly as that sounds
[21:12:59 CET] <BBB> not target, just dragged into it
[21:13:24 CET] <BBB> its very easy nowadays to become a target of trolls, and the last thing we need is another set of trolls to hate us
[21:13:30 CET] <BBB> like we dont have enough of that already
[21:13:45 CET] <Compn> dragged into where? i'm sorry if it sounds dumb i just have no idea what you think the issue will be in the future
[21:14:08 CET] <Compn> if we take a side on this issue (which we already did, with the widget before)
[21:14:43 CET] <Compn> "if we put widget on website , XXX will happen"
[21:14:46 CET] <Compn> define xxx
[21:14:53 CET] <BBB> nah, too direct
[21:14:59 CET] <BBB> I dont know what will happen
[21:15:13 CET] <BBB> but I prefer a world in which we are not a political organization or a project with political leanings
[21:15:16 CET] <BBB> ok, fine
[21:15:31 CET] <BBB> if we put widget on website, we will be considered a project with political leanings
[21:15:39 CET] <BBB> does that make sense?
[21:16:03 CET] <BBB> its probably very cryptic, but maybe it makes sense
[21:16:29 CET] <Compn> i understand you think something bad will happen
[21:16:33 CET] <Compn> thats , i get it
[21:17:02 CET] <Compn> thanks for taking time to explain it
[21:17:02 CET] <rcombs> we're a group of people
[21:17:05 CET] <BBB> I think politicization is fundamentally a negative thing, yes
[21:17:07 CET] <rcombs> we're gonna have positions on things
[21:17:12 CET] <BBB> people, sure
[21:17:16 CET] <BBB> but as a group?
[21:17:42 CET] <Compn> are we a group ?
[21:17:47 CET] <durandal_1707> group?
[21:18:09 CET] <rcombs> I mean, we appear to be pro-free-software and anti-software-patent, at the very least
[21:18:11 CET] <durandal_1707> Carl and friends do not make a group
[21:18:32 CET] <Compn> rcombs : those sound very political
[21:18:53 CET] <rcombs> if you had a specific concern I'd find that reasonable but just saying "X shouldn't be political" seems fairly naïve
[21:19:19 CET] <rcombs> politics is the system by which the world is run
[21:19:27 CET] <durandal_1707> you are political all the time on mailing list
[21:19:32 CET] <thardin> this very project is political
[21:19:35 CET] <Compn> don't pile on him too hard rcombs :)
[21:19:35 CET] <rcombs> pretty difficult to completely separate anything from it
[21:19:53 CET] <BBB> thardin: how?
[21:19:59 CET] <Compn> thardin : we are anti-realplayer and anti-quicktime too :D
[21:20:01 CET] <rcombs> "we shouldn't take a public position on X" is reasonable
[21:20:15 CET] <BBB> rcombs: sort of like that, I guess
[21:20:18 CET] <rcombs> "we shouldn't be political" misunderstands what politics are
[21:20:24 CET] <thardin> "there should be a libre set of multimedia libraries" seems like a fairly political stance to me
[21:20:31 CET] <wm4> is ffmpeg pro-gun or anti-gun? I assume pro-gun, because it makes it so easy to shoot yourself in the foot
[21:20:36 CET] <rcombs> lol
[21:20:55 CET] <rcombs> note that I'm not arguing that the project should have a public stance on everything
[21:20:57 CET] <Compn> spacex rockets use ffmpeg , so we're definitely pro-space 
[21:21:03 CET] <BBB> rcombs: I would express it as we shouldnt take a position on X because that issue is politicized too much (i.e. Im giving a specific reason why that issue specifically is one Id like to avoid specifying a position on)
[21:21:13 CET] <BBB> rcombs: does it make more sense then? or am I just being more confusing then?
[21:21:17 CET] <rcombs> that doesn't mean anything
[21:21:30 CET] <BBB> hm...
[21:21:34 CET] <rcombs> is it "we shouldn't take a position on X because it's too controvertial"?
[21:21:36 CET] <thardin> also net neutrality is important. it angers me to no end that we've lost it in sweden
[21:21:45 CET] <BBB> rcombs: yeah I guess
[21:21:46 CET] <Compn> BBB : i am sorry, but it just sounds like concern trolling. "we shouldnt take a public position on x" would be fine though
[21:21:58 CET] <durandal_1707> Compn: rockets use ffmpeg? for what? fake earth pictures?
[21:22:06 CET] <thardin> durandal_1707: cameras
[21:22:14 CET] <rcombs> so, that seems like a fairly reasonable argument
[21:22:22 CET] <rcombs> personally I disagree, but I can see your point
[21:22:29 CET] <Compn> durandal_1707 : michaelni himself helped elon musk with his damaged microblocks :D
[21:22:52 CET] <BBB> thardin: I dont think (honestly) that opensource/libre/free software is a political leaning in and by itself
[21:23:12 CET] <thardin> I'd argue it is explicitly anticapitalist. at least free software is
[21:23:12 CET] <BBB> but maybe that doesnt matter so much
[21:23:17 CET] <BBB> really?
[21:23:28 CET] <BBB> me and kierank might disagree on that with you
[21:23:32 CET] <thardin> ok perhaps not explicitly
[21:23:34 CET] <BBB> possibly others also :)
[21:23:34 CET] <Compn> durandal_1707 : https://forum.nasaspaceflight.com/index.php?topic=34597.500
[21:23:53 CET] <thardin> but definitely adjascent
[21:24:03 CET] <Compn> libre and copyleft is totally anti-copyright :P
[21:24:22 CET] <Compn> especially if you look at GNU and GPL
[21:24:39 CET] <thardin> gpl is a marvellous copyright hack
[21:24:59 CET] <thardin> 10/10 praxis, would license again
[21:27:11 CET] <Compn> anyways it was a good discussion, i'm glad we were able to have it
[21:27:17 CET] <Compn> sorry jamrial if i was snarky towards you
[21:27:21 CET] <thardin> for the record, I don't have any strong feelings on the website thing
[21:27:22 CET] <rcombs> heh, TEI
[21:27:30 CET] <rcombs> I just recently added support for reading that in lavf
[21:29:12 CET] <thardin> is it just me or is the DASH code awful?
[21:29:19 CET] Action: kierank goes back to keyboard and sees BBB wall of text
[21:29:56 CET] <rcombs> enc or dec
[21:30:55 CET] <Compn> lol watching these spacex videos, the rocket keeps exploding if it falls over
[21:31:06 CET] <Compn> you think they would fix that first...
[21:32:00 CET] <rcombs> well yeah
[21:32:09 CET] <rcombs> the tanks are pressurized and full of explodium
[21:33:00 CET] <thardin> demuxer. at least based on the patches I've seen
[21:33:00 CET] <kierank> durandal_1707: Compn just wants ffmpeg to be his conspiracy theory dump
[21:33:10 CET] <kierank> next the NRA
[21:33:21 CET] <rcombs> thardin: on one hand& yeah it's a mess
[21:33:30 CET] <rcombs> on the other hand, DASH kinda inherently has to be to some degree
[21:33:42 CET] <rcombs> 'cause it's ridiculously complex and also built on top of XML
[21:40:53 CET] <BBB> kierank: sorry about that
[21:41:15 CET] <kierank> and the argument about videolan doing it is bogus
[21:41:17 CET] <kierank> if he makes it
[21:41:26 CET] <kierank> because videolan is an association around many OSS projects
[21:41:52 CET] <kierank> I have never cared for any of the political stuff (e.g snowden) at OSS confs either
[21:42:52 CET] <kierank> if people like Compn really cared about OSS, they'd make it apolitical like the Olympics
[21:43:05 CET] <kierank> But instead it's about pushing their views on others who don't care for it
[21:43:07 CET] <BBB> maybe some people do want it to be more political
[21:43:18 CET] <BBB> I can appreciate that theres different opinions on that, maybe?
[21:43:22 CET] <BBB> but I fundamentally disagree
[21:43:38 CET] <kierank> In the world as it stands you then lose ~52% of people
[21:43:49 CET] <BBB> per issue
[21:44:15 CET] <BBB> yeah, lets start getting ffmpeg involved in the gun debate
[21:44:18 CET] <BBB> see how well that goes...
[21:44:53 CET] <BBB> anywya, thats a silly argument, I think I tried to the best of my abilities over email and IRC to make a real case for why I dont like the widget
[21:47:53 CET] <sfan5> thardin: would you design dashdec significantly different from scratch or is it just the small issues that make it awful?
[21:48:33 CET] <thardin> it's just a feel
[21:49:11 CET] <thardin> coupled with my growing concern of the lack of langsec in most free software
[22:03:52 CET] <wm4> I feel like dashdec mostly copy&pasted from HLS
[22:04:25 CET] <sfan5> minus the XML bits, but yeah the code is very similar
[22:05:40 CET] <wm4> yeah, since HLS doesn't use XML, that part is obviously very different
[22:05:50 CET] <wm4> but I'd argue they could share some code
[22:06:06 CET] <JEEB> yea, I thought that we might as well make a common file for a lot of the functionality
[22:06:24 CET] <JEEB> there's quite a bit of specifics of being able to put various stuff on various levels of the XML, of course. but in the end you're creating a list of things to access
[23:13:35 CET] <Compn> kierank : the olympics are so highly politicized, what a bad example you've come up with.
[23:19:04 CET] <Compn> kierank : also your strawman arguments about me re: nra or videolanorg are bordering on COC violations
[23:19:55 CET] <wm4> since when did a COC violation have consequences in ffmpeg
[23:38:10 CET] <cone-397> ffmpeg 03Jerome Borsboom 07master:1160d1d47887: avcodec/vc1: add bitstream elements for VAAPI VC-1 interlaced decoding
[23:38:10 CET] <cone-397> ffmpeg 03Jerome Borsboom 07master:883bdc5fb7c4: avcodec/vaapi: add fields for VAAPI VC-1 interlaced decoding
[23:38:10 CET] <cone-397> ffmpeg 03Jerome Borsboom 07master:0ec7eb930520: avcodec/vaapi: mask unused bits in bitplane_present.value
[00:00:00 CET] --- Mon Mar  5 2018