IRCloggy #hts 2015-06-24

Logs Search ←Prev date Next date→ Channels Documentation

Provider of IRC logs since 2005.
WARNING: As Freenode became unjoinable and lost all warnings in topics, we cannot log channels on Freenode anymore.

2015-06-24

→ funky1 joined00:41
→ GaMeR64 joined01:04
← G64 left01:07
→ nuc-ninja_ joined01:14
← nuc-ninja left01:14
nuc-ninja_ → nuc-ninja01:14
← nilzen left01:54
← mETz left02:19
→ mETz joined02:25
→ nilzen joined02:30
→ adamsutton joined02:40
ChanServ set mode: +o02:40
→ arcus joined02:58
← arcus_ left03:01
← alexkn left03:05
→ alexkn joined03:06
← alexkn left04:37
→ alexkn joined04:54
→ jus7incase joined05:12
← jus7incase left05:12
→ G64 joined05:16
← GaMeR64 left05:18
← alexkn left05:20
→ alexkn joined06:07
← alexkn left06:15
→ robvh joined06:19
→ PrinceOfCats joined06:24
→ goroboro joined06:30
→ IvanG joined06:49
→ sekil joined07:00
→ xhaggi joined07:08
→ ckarrie joined07:16
← robvh left07:19
← linuxstb left07:35
→ dmarion joined07:55
→ dreamcat4 joined07:58
← ckarrie left07:59
→ Yepoleb_ joined08:00
← Yepoleb left08:00
Yepoleb_ → Yepoleb08:00
→ hans_s joined08:18
→ linuxstb joined08:19
→ loki666 joined08:19
loki666 hi, I have a question about HTSP08:20
implementation I've found, read HTSP packet as soon as they present themself on the socket08:21
so it's working in "push", but what about "pulling" packet. when needed, would that work?08:21
mpmctoo perexg: I'm still trying to setup a pi as a sat>ip server, but I'm not getting anywhere, all I'm getting now is 400 (bad request) whenever I try to reconfigure it I end up getting "missing tuner settings - global config" Is there some secret to getting Sat>IP with a single LNB working?08:22
dreamcat4 hi, when i set 0775,0664 recording permissions it doesn't make group-writable. instead the recordings are writted 0775 and 064008:26
tvheadend: version 4.0.5-4~g7a7b3b308:28
does anyone else see that issue ?08:28
hans_s dreamcat4: is your tvheadend running in background? (-f --fork)08:30
otherwise umask applies08:30
dreamcat4 hans_s: uhm... that's some thing i didn't expect to hear :)08:33
hans_s: it seems to be running as foreground. there is no -f --fork option08:34
hans_s dreamcat4: even the help states, that umask applies ;)08:34
dreamcat4 hans_s: it sounds like an entirely unreleated feature to me08:35
hans_s it's -f or --fork, -f is the shortopt for --fork08:35
→ ckarrie joined08:35
hans_s well, about the fork, that's not documented, I just search the code for umask and saw, that umask is set to 0, when -f or --fork is used08:36
dreamcat4 ok. thanks for doing that for me! it's not a bug then you don't believe ?08:37
hans_s umask is a standard unix feature08:37
perexg mpmctoo: it should be really simple: create network with "SAT>IP Source" set to 1 and in the global config - set the number of exported virtual tuners.. like 1 for DVB-T etc...08:37
mpmctoo: the network must be attached to a tuner...08:38
→ Bumble-Bee joined08:38
hans_s dreamcat4: you can call "umask 0" on shell before starting tvheadend to set the umask for tvheadend to 008:39
nuc-ninja mpmctoo: and do not forget to add a multicast route to the routing table i.e. route add -net 224.0.0.0 netmask 240.0.0.0 eth008:39
perexg mpmctoo: anyway, it might be easier to configure minisatip on pi to save resources... if you don't expect to stream from it..08:39
dreamcat4 hans_s: thanks! - that sounds like a good thing for me to try now08:40
→ zipleen joined08:41
dreamcat4 hans_s: so it's 'umask 0 ; tvheadend $opts' ?08:41
hans_s loki666: can you explain further? what are you trying to accomplish?08:41
dreamcat4: yes08:41
dreamcat4 seem like i shall need to rebuild my container. will take a while. in meantime can you point me to the souce code where umask / fork in main? or no bother i'll find it myself08:42
mpmctoo perexg: A few config shots, PI server: http://i.imgur.com/wb7KWhh.png?1 http://imgur.com/2iX8B2x Client (main tvh server) http://i.imgur.com/w2Nm5h7.png?108:43
dreamcat4 hans_s: nevermind - found it08:43
perexg mpmctoo: go to 'global config' there's 'SAT>IP server' block (you need to click on it to expand) and configure Exported tuners...08:44
loki666 hans_s: what's happening on hts, if I don't read HTSP packet as soon as they presents on the socket, but only when my video/audio sink requires to fill its buffers. will the server queue the packets so I can read them in burst ?08:45
andoma loki666: it will queue up to a limit, then it will start dropping frames intelligently08:45
loki666 filtering no stream packets that i don't need08:45
andoma so initially it drops non-ref frames08:45
at least that's how it worked when i wrote it many years ago :)08:46
loki666 oh ok, what's the queue limit? can the client configure it?08:46
mpmctoo perexg: Do you mean this? http://i.imgur.com/1NEr0Vm.png?108:46
hans_s loki666: also your os and the server os will probably buffer some data in it's cache08:46
perexg andoma: this code is there unchanged :-)08:47
andoma loki666: ok, it can be sent as part of the HTSP subscribe message08:47
perexg mpmctoo: no - configuration / general / SAT-IP Server ....08:48
mpmctoo: on rpi...08:48
nuc-ninja perexg: I think I have nailed down the CCerr culprit…I have added some LAG connections to the network and the multicast traffic is load-balanced across the LAG group so I believe I am getting out of order packets…will run a tcpdump to see if this is the case….08:48
loki666 andoma: ok works for me, thanks for the info08:48
mpmctoo perexg: Ooohh, I see it now >.<08:49
mpmctoo feels like an idiot for not noticing it sooner!08:49
perexg nuc-ninja: it might be.. there are not order specific checks in tvh.. but it is not too difficult to add some..08:49
andoma loki666: "queueDepth" parameter part of the subscription message controls the buffer size08:50
default is 500000 (ie, half meg buffer)08:50
perexg mpmctoo: I know that it's not ideal - I'll create next-level for panels to split the general config to more tabs..08:50
ProfYaffle hans_s, dreamcat4 - I remember scratching my head a little about the umask when I added that code. I assumed (but never really tested, I must confess) that root umask applies to the daemonised process (022 vs 002 for a normal user).08:53
negge nuc-ninja: so your CC errors were due to out of order packets?08:54
→ Immanuel joined08:55
Immanuel hi08:55
mpmctoo perexg: Oh btw, is there anything you can do about getting newer builds into the apt.tvheadend.org repo, seems to be causing a lot of issues for people (namly those that use i386 builds as the dockerhub doesn't support them)?08:56
perexg mpmctoo: it's probably question for adamsutton ..08:57
nuc-ninja perexg: am still checking if that is the case…I am analysing a packet trace at the moment and will report the findings…08:59
← IvanG left09:02
perexg nuc-ninja: are you using UDP or RTP/UDP ? for raw MPEG-TS in UDP, there's missing any stamping to reorder packets back..09:03
nuc-ninja perexg: mpeg-ts in udp…I can switch it to rtp to check if the CCerrs occur as well09:04
perexg nuc-ninja: there are no packet order checks for RTP in tvh either, but for RTP - they can be added..09:05
dreamcat4 ProfYaffle: i'll investigate umask behaviour later today. raise an issue if needed09:05
← loki666 left09:06
perexg nuc-nunja: but if you analyze packet flow - you should detect the reordering using the RTP headers..09:06
ProfYaffle dreamcat4, or just update the help, as I might have just got it wrong in the first place. the code certainly calls for file creation with the stipulated permissions, I know that09:06
so it might just be about clarifying what to expect09:06
nuc-ninja perexg: i will switch to rtp and recheck….09:07
dreamcat4 ProfYaffle: nah it's most likely a bug. the documentation may be fixed afterwards to reflect any new change in the umask behaviour (depending what's decided / done about it)09:09
ProfYaffle: it was pretty good that you mentioned umask at all to begin with actually. very thorough work good job :)09:11
← nilzen left09:20
→ nilzen joined09:20
DoozerBot Commit in 'master ' by Jaroslav Kysela [ files modified ] http://bit.ly/1K9KBV809:22
IPTV: Add RTP sequence checks09:22
perexg nuc-ninja: with ^^^, if you enable "--trace iptv" you should see the RTP sequence discontinuities in log now...09:23
"RTP discontinuity" lines...09:23
DoozerBot Build #11502: v4.1.296 for precise-amd64 completed http://bit.ly/1K9KBVj 09:24
nuc-ninja perexg: am not getting any RTP discontinuities (at least at the ethernet interface level)…will turn on trace as you have suggested and let you know I am also getting tons of “uncorrected blocks” errors in the gui but am not seeing anything in the log…09:25
DoozerBot Build #11503: v4.1.296 for precise-i386 completed http://bit.ly/1K9KIzW 09:25
mpmctoo perexg: Great work on the SAT>IP code, it works fantastic! :)09:26
DoozerBot Build #11504: v4.1.296 for trusty-amd64 completed http://bit.ly/1K9KPeY 09:27
Build #11505: v4.1.296 for trusty-i386 completed http://bit.ly/1K9KYiw 09:28
perexg nuc-ninja: "uncorrected blocks" or "continuity errors" ? before my patch, the "uncorrected blocks" should be always zero for IPTV..09:28
nuc-ninja before patch…cannot get your patch to run …core dump09:29
perexg: iptv services have tons of uncorrected blocks….09:30
perexg nuc-ninja: really ? ok, then the errors are because UDP packets does not contain the MPEG-TS packets...09:32
nuc-ninja: it has nothing to do with reordering...09:33
nuc-ninja: this counter increases with 0x47 sync is lost..09:33
also09:33
nuc-ninja perexg: OK…but am still getting CCerrs…no out of order packets…wireshark RTP shows all is in order, jitter is nominal (10 micro seconds max)…sequence numbers are all in order…wierd09:35
perexg nuc-ninja: do you have backtrace for the crash with my patch ? I don't see any issue in my code...09:38
nuc-ninja perexg: will get you one…it is not because of your patch…I updated to the latest git this morning and got core dumps so I reverted to yesterday morning’s version09:39
perexg: should I also compile in tsdebug?09:41
perexg: never mind…I compiled it in anyway...09:41
perexg it won't help probably; only for confirmation - do you see both cc errors and uncorrected block errors ?09:42
nuc-ninja perexg: am recompiling and installing at the moment…will get back with the results09:42
adamsutton mpmctoo: ??09:44
nuc-ninja perexg: message is “ tvheadend[30434]: segfault at 1142000 ip 00000000004a8370 sp 00007fff9c2a79b0 error 4 in tvheadend[400000+93a000]”09:44
mpmctoo adamsutton: The apt repo isn't being updated properly =P09:45
adamsutton oh, I thought that was all abandoned, I thought someone had set something else up?09:45
BtbN for Ubuntu09:45
mpmctoo adamsutton: with no support for i386 builds which is causing a few issues.09:45
BtbN Who still has 32bit hardware?09:46
adamsutton ah ok, sorry I'm falling further and further out of the loop09:46
BtbN: it's a very good question :p09:46
mpmctoo I do actually lol.09:46
adamsutton but apparently people do!09:46
mpmctoo Not that I use it anymore but still.09:47
adamsutton mpmctoo: I'll buy you a new PC, might be more efficient than me fixing apt :p09:47
mpmctoo adamsutton: Lol, I have too many systems as it is! I rarely buy new systems though so most of mine are old.09:47
perexg nuc-ninja: are you run tvh from a package or directly from sources ? the debug package should be probably installed to get source files instead RAM addresses..09:48
mpmctoo adamsutton: Would be nice to have the apt repo working again though, I think the version number change broke it somehow.09:48
nuc-ninja perexg: running from sources….09:48
adamsutton fair enough, I'm probably similar, I rarely upgrade stuff, just my last round of upgrades (mainly for TVH) was not that long ago09:49
mpmctoo: ah, what changed?09:49
BtbN it's at 4 now09:49
perexg nuc-ninja: do you have EXECINFO in .config.mk ? grep EXECINFO .config.mk09:49
← ckarrie left09:49
BtbN i think you can safely remove ubuntu builds from it, and only build for debian(i think jessie is missing, too)09:49
mpmctoo adamsutton: versions went from n.n.n to n.n.n-n (or something like that).09:49
BtbN The git revision has allways been in there.09:50
perexg nuc-ninja: perhaps, it might be easier to run tvh binary under gdb (gdb --args <binary_and_args).. then type 'r' and 'bt' after crash..09:50
adamsutton rebuildd web interface is broken, that could just be the web interface or the whole build system could be down.09:50
it was always flakey09:50
hans_s might have to do with the concatenating/comressing js/css resources09:52
adamsutton actually build system itself is still running and utopic builds are happening, not sure if they're reaching apt?09:53
but others are failing09:53
mpmctoo: perexg: I'll take a look09:55
hans_s there's 4.1.18 for utopic in the repo, that seems to be one of the last versions before the change to the 4.1-nnn version scheme happended09:56
adamsutton ah, has something more significant changed?09:56
and yes, I misread that as June, not May, so was thinking it was only a few days old09:57
hans_s well, the verion scheme now has a '-' instead of a '.' to separate tag from revision count in the output of git describe09:59
see recent history of support/version file10:00
adamsutton hans_s: what's the logic behind that? just trying to avoid big jumps in the patch number?10:09
hans_s I'm not really sure, perexg should be able to explain10:12
adamsutton perexg: ?10:13
hans_s But I think it has to do with the tags on the release/4.0 branch, as there are 4.0.0, 4.0.1 etc. tags10:13
adamsutton struggling to find any obvious reasons why things are failing10:13
← linuxstb left10:20
perexg ok?10:22
adamsutton perexg: what was the purpose of the version change? not criticising, just curious about thinking?10:22
perexg the new numbering follow the git - I use tags now to create .zip files...10:22
for the stable versions like 4.0.1 etc...10:22
so everything after '-' means 'number of development changes' now./.10:23
adamsutton ah ok10:23
yeah, makes sense10:23
perexg follow the github - not git...10:24
adamsutton well I can run the pbuilder code locally, so no idea why the auto builds have stopped10:24
perexg adamsutton: https://github.com/tvheadend/tvheadend/releases - nice isn't ? so we don't need to create own source packages...10:25
adamsutton yeah, that's cool10:25
I think that's kinda what I'd been aiming for, but as usual got it a bit cockeyed10:25
so yeah, makes much more sense now10:26
clandmeter perexg: are those real releases, or just git tags?10:52
perexg clandmeter: release = git tag... that's the whole idea of releases on github...11:00
clandmeter perexg: i think you can also do real releases with real release names and release notes? filenames would be $project-$version.tar.gz (i never tried it though)11:01
perexg clandmeter: any docs for it ? but this current scheme is very handy for me.. if you click on any package, you'll get tvheadend-version name, too...11:04
→ Spr0cket joined11:17
nuc-ninja perexg: core dump was my fault…wrong perms on a dir. as to the CCerrs, I have moved the tvh server onto the same switch as the multicast source (no more LAG to worry about) and I am still getting CCerrs on the udp streams (moved back to UDP as some of my STBs cannot do RTP)..wireshark analysis of the TS Stream shows all Continuity Counters in order (no missing packets and no out of order counters) the CCerrs show up more frequently on H11:30
than SD streams but I am getting them on both....11:30
→ linuxstb joined11:44
negge nuc-ninja: how can you simply switch between UDP and RTP? Are you generating the multicast locally?11:52
clandmeter perexg: i checked, but it seems you are already doing it correctly, would still be nice if there was a release notes though (gives you more trouble/work though)11:53
nuc-ninja perexg: yes, i am generating locally (vlc is your friend :) )…11:54
perexg: am also testing ffmpeg/opencaster but that is really screwy....11:55
→ hyponic joined11:57
perexg nuc-ninja: so in tvh - continuity errors increase and/or uncorrectable ?12:03
nuc-ninja perexg: correct12:06
perexg nuc-ninja: if you look to the UDP packets in wireshark - do they all have same size ?12:08
nuc-ninja perexg: since I moved the tvh server to the other switch, I have no more Uncorr. Block errs..only CCerrs now…as mentioned, the pcap shows that all the packets coming in on the wire are in order and C. Counter is showing 1-16 in order but tvh is showing CCerrs… have tried to play with the sysctl (net.ipv4.udp_mem/udp_rmem values) but they have no effect on tvh…on the wire, all is correct…do not have access to a TS analyiser though :12:10
perexg: yep…all have 1316 as the size…running a tcpdump again to verify (tcpdump -i eth0 host 239.255.1.155 | grep -v 1316)12:12
→ dreamcat6 joined12:15
nuc-ninja perexg: one thing noticed is when an IGMP report is sent out from the kernel, I immediately get a CC error12:16
← dreamcat5 left12:17
nuc-ninja perexg: scrap the last statement, was a coincedence...12:18
perexg nuc-ninja: could you grab the stream using 'socat' to check if received data are correct ? the support/pid-count.py utility in tvh source tree should verify the CC problems..12:20
the command should be like: socat STDIO UDP4-DATAGRAM:224.1.0.1:6666,bind=:6666,range=192.168.10.0/24,ip-add-membership=224.1.0.1:192.168.10.2 > a.ts12:22
http://www.dest-unreach.org/socat/doc/socat-multicast.html12:22
→ ckarrie joined12:25
nuc-ninja perexg: output is here: http://pastebin.com/5ELmRVub it is showing cc errors12:39
perexg: ouch12:40
perexg nuc-ninja: could you play with rcvbuf for socat ?12:42
nuc-ninja perexg: will do12:43
perexg Try to set it to 1048576 bytes12:43
← ckarrie left12:47
→ ckarrie joined12:50
nuc-ninja perexg: does not seem to matter what I set the rcvbuf size to…command I am using is socat -u UDP4-RECV:1234,ip-add-membership=239.255.1.155:0.0.0.0,rcvbuf=262144 CREATE:a.ts…have tried 262144, 524288, 1048576,2097152, 4191308...12:55
← ckarrie left12:57
nuc-ninja perexg: looks like a vlc streaming problem…will do a capture on the source machine….13:01
perexg yes, the pid-count.py utility will show you issues....13:01
nuc-ninja perexg: am also using the tsccc program from opencaster and it is showing issues as well….13:02
perexg ok13:02
→ ckarrie joined13:04
← ckarrie left13:04
nuc-ninja perexg: it seems like I will never find a simple application that streams .ts files without interuption and loop the list where I can change the source file on the fly and all the info is corrected (PTS/DTS/PCR) etc…. have been trying vlc, ffmpeg, multicat, opencaster…and nothing seems to do what i need….13:05
BtbN Write one yourself, it's not even remotely simple.13:11
looping them is not easy, as you have to fix the timestamps13:11
nuc-ninja BtbN: I know it is not simple…I am just wondering that no one other than myself is looking for something along those lines…13:12
BtbN Why would someone want to loop a playlist with ts files?13:12
Seems quite unusual13:12
and changing the playlist on the fly... at what point should it read the new one? When it's done reading it? Imediately?13:13
nuc-ninja BtbN: simple home IPTV with my own video files…no commercials, etc….13:13
BtbN Why do you want to loop ts files for that?13:14
andoma play using HLS?13:14
nuc-ninja BtbN: vlc comes real close with the vlm functionality, just am fighting with CCerrs at the moment…13:14
BtbN Just remux them to mkv or one big ts file, and play them via http/smb/...13:14
nuc-ninja BtbN: not what I want…I have multiple STBs that can only receive UDP multicast streams and I wish to have the setup like IPTV and only worry about the files on the server…I am not a fan of “client” based media management…WAF is then very low…13:16
BtbN: and all STBs (whether UDP multicast receivers or XBMC clients) should have the same “place” in the broadcast…13:17
BtbN Broadcast-Style media consumption is deprecated imo13:17
It's slowly dying in favor of on-demand video13:17
Why not just use some media player in fullscreen, and let it loop over the list?13:19
nuc-ninja BtbN: that is why it is an “opinion”…linear TV still has its appeal, especially if it is using local content13:19
BtbN not realy, TV still has its appeal with actual live content, like sport events13:19
for recorded stuff like movies it's just outdated and annoying13:19
nuc-ninja BtbN. not an option as as I have 12 clients accessing the same content…13:19
BtbN how many damn wifes do you have?!13:20
negge wouldn't you be better of with 12 Kodi clients and a central tvheadend server?13:20
BtbN: :D13:20
nuc-ninja BtbN. I do not use it for movies…I have a “dedicated channel” for each TV-Series; i.e. The big Bang Theory Channel or the Stargate SG-1 channel.13:20
BtbN But why? So you can experience missing stuff when having to take a break?13:21
nuc-ninja negge: no possibility to serve “files” from tvh in a uninterrupted form to multiple clients…13:21
negge nuc-ninja: so you're doing something like the "pseudo-tv" addon for Kodi?13:21
you should look into it, sounds like exactly what you want/need13:21
BtbN The only idea i have for that is an actual broadcasting setup13:21
nuc-ninja BtbN: not an issue as I know the episodes by heart and it does not matter…13:21
BtbN wtf13:22
Well, you are kinda alone with that, so either pay someone to implement what you want, or do it yourself. I don't think that exists yet.13:22
nuc-ninja negge: sort of but I then need to configure every bloody client…looking for a central server solution...13:22
BtbN There is no such thing13:23
build it yourself13:23
nuc-ninja BtbN. vlc comes very, very, very close but it is giving me CC errors…other than that, it does exactly what I want…13:23
negge nuc-ninja: you could most likely distribute the pseudo-tv addon configuration somehow13:23
ProfYaffle nuc-ninja, can you use path substitution or something on the Kodi clients to share the config? Are they all the same? Could you pre-configure an OpenElec installation, and then update it from a central location in some way?13:24
BtbN yes, that's what happens when you loop13:24
like i said, you need an actual broadcasting server, as you want to do an actual broadcast.13:24
nuc-ninja negge: being a network guy, I do not like “distributing” something to each and every client…not when it can be solved elegantly centrally….13:24
negge I'm just thinking there has to be a better way than multicasting stuff with vlc13:24
BtbN a cheap one would be some mediaplayer looping what you want + ffmpeg screen-capturing it and sending the screen capture as continous stream.13:25
negge nuc-ninja: I was thinking of a network share mounted where the addon config resides13:25
nuc-ninja negge: it works, just not 100% corretly (damn CC errors in tvh) but the streams are watchable on Kodi clients as well as my STBs…13:25
negge what are these STBs?13:25
nuc-ninja when I meant damn CC errors, this is not specific to tvh (which is great) but to vlc generating them in the first place!!!!13:26
negge: Dune HD boxes13:26
negge: older ones....13:26
BtbN: screen capture for 110 channels is not an option :D .. the answer as to why I want to do this is “because I can”13:28
BtbN It is, just build yourself a suitable solution13:29
PulseAudio with propper routing, a huge virtual screen with automatic layouting13:29
not impossible at all13:29
someone just has to build it13:29
nuc-ninja BtbN: my C skills are 30 years out of date :D and I have a “day job” as well :D13:29
BtbN you don't need any C for that13:29
nothing a shell script can't do13:30
nuc-ninja BtbN: I was referring to your comment about building the application myself…13:30
BtbN: as you mentioned, the timestamp correction is the most major headache….13:31
← PrinceOfCats left13:36
→ perexg joined13:51
[-T-] hey13:56
it looks like there is a bug in the descambler/cwc part13:56
when a chan has multiple ecm pids13:57
← linuxstb left14:20
← perexg left14:22
nuc-ninja perexg: is there any way to implement a tsfixcc type of functionality in tvh? see opencaster-3.2.2 fixccc app for an example: http://ftp.de.debian.org/debian/pool/main/o/opencaster/opencaster_3.2.2+dfsg.orig.tar.gz14:27
→ UmbraMalison joined14:28
hans_s nuc-ninja: why would you want to implement that in tvheadend?14:29
→ perexg joined14:31
hans_s do you mean this? https://github.com/nexgenta/opencaster/blob/master/tools/tsfixcc/tsfixcc.c14:31
nuc-ninja hans_s: yep14:37
perexg the problem is that this complely throws off the TS checking....14:38
completely14:39
hans_s I still don't see, why this should even be in tvheadend14:39
nuc-ninja perexg: ok…I did a test with piping the vlc udp output into a fifo and re-multicasting from the fifo via ffmpeg…cc errors are gone…14:39
perexg: and the transcoding session in tvh is a lot happier…hence my question :)14:40
hans_s and you don't have cc errors in ffmpeg?14:40
or does ffmpeg correct the cc errors?14:41
nuc-ninja perexg: no more “Detected framedrop in audio” amd libav errors complaining about “skipped MB in I Frame at xxx xxx”…when there are no CCerrs, the transcode session is super happy....14:41
hans_s: if there were cc errors on the ffmpeg input, ffmpeg would complain…will do a socat session to remove ffmpeg from the equation and report results....14:42
hans_s: when I strem from a pure .ts file with ffmpeg, i get other sorts of errors, but from the fifo, no errors at all (other than ffmpeg complaining about thread_queue_size on startup but this is due to ffmpeg reading from a fifo)14:44
hans_s if ffmpeg doesn't have cc errors on input, then this would probably mean that vlc does something wrong when udp multicasting...14:45
nuc-ninja hans_s: yes, you are correct…I am just trying to find a “workaround” for the issues I am facing :D14:45
→ KingDaveRa joined14:46
nuc-ninja hans_s: getting something fixed in vlc is a “nightmare”…took the vlc team 2 years to fix the PGS subtitle bug I reported….14:46
hans_s if the fifo works with vlc, you could also use something simpler than ffmpeg, like netcat14:49
← sekil left14:53
← freebs left14:56
→ sekil joined14:57
nuc-ninja hans_s: I am now doing this; vlc->fifo->tsfixcc->fifo->tsudpsend on a CBR encoded .ts file…no cc errors14:57
→ linuxstb joined15:00
[-T-] ok I found my issue15:07
→ freebs joined15:08
[-T-] in the ECM filters, if you set USE for a provider and you don't explicitly set USE on another one, on the same chan15:08
it will only use the first one15:08
perexg [-T-]: depends if you reject all other providers...15:09
[-T-] i don't15:09
let's put numbers15:09
i have 000ABC and 000DEF both working for given chan15:10
i have a line that says USE 000ABC for this chan15:11
+ i line that says IGNORE 000ABC15:11
and that's it15:11
000DEF will never be used15:11
no filter is created for it15:12
but if I add a line at the end of the list saying : USE 000DEF15:12
then it's ok, it create both filters15:12
← sekil left15:12
perexg [-T-]: ok, looking to sources, it's how to code should work.. when USE or IGNORE is detected, only specified rules are active; all other streams are ignored....15:14
[-T-] looks so :)15:15
if you have 2 USE lines though15:15
both are used15:15
(for the same matching pid)15:15
thanks for checking ...15:16
perexg yes, you may use 'EXCLUSIVE' to filter to select only one (first hit - first won)15:16
[-T-] yep15:18
the behavior if good, when you use it well :)15:18
it's a bit different from oscam.dvbapi, which confused me15:18
perexg I wrote the code, but I always need to check again what it does :-)15:19
[-T-] what lead me to this study, is when I wanted to force it to use 000DEF, so I unchecked the USE line on 000ABC15:21
and it never tried to filter 000DEF15:21
because of the prefered CA PID ...15:21
this is maybe a little bug15:21
→ nils_2 joined15:48
→ DereC joined15:55
← xhaggi left16:00
← linuxstb left16:04
→ nils__2 joined16:08
← nils_2 left16:08
→ andoma joined16:12
ChanServ set mode: +o16:12
→ nils_2 joined16:12
→ andoma_ joined16:13
← nils__2 left16:13
← hans_s left16:21
→ robvh joined16:44
← KingDaveRa left16:58
→ ant_thomas joined16:59
→ KingDaveRa joined17:01
← zipleen left17:05
→ tuxuser joined17:14
← freebs left17:16
→ freebs joined17:28
→ alexkn joined17:34
DoozerBot Commit in 'master ' by Jaroslav Kysela [ files modified ] http://bit.ly/1J5Oj0B17:42
muxer/makedirs: fix mode (ulimit) - fixes #296917:42
Build #11514: v4.1.297 for precise-amd64 completed http://bit.ly/1J5OjxN 17:43
Build #11515: v4.1.297 for precise-i386 completed http://bit.ly/1J5OB7J 17:45
Build #11516: v4.1.297 for trusty-amd64 completed http://bit.ly/1J5OWaA 17:47
Build #11517: v4.1.297 for trusty-i386 completed http://bit.ly/1J5PhKa 17:48
Commit in 'master ' by Jaroslav Kysela [ files modified ] http://bit.ly/1J5S8Tx17:59
WEBUI JS: more tooltip additions for grid headers17:59
← robvh left18:04
DoozerBot Build #11522: v4.1.298 for precise-amd64 completed http://bit.ly/1J5TT33 18:10
Build #11523: v4.1.298 for precise-i386 completed http://bit.ly/1J5Uf9K 18:11
Build #11524: v4.1.298 for trusty-amd64 completed http://bit.ly/1J5Uu4V 18:13
Build #11525: v4.1.298 for trusty-i386 completed http://bit.ly/1J5UMsn 18:15
→ hans_s joined18:23
→ GaMeR64 joined18:33
← G64 left18:34
→ G64 joined18:38
← GaMeR64 left18:39
→ linuxstb joined18:47
← DereC left19:12
→ IvanG joined19:25
→ DereC joined19:26
→ ChrisSc joined19:31
ChrisSc hi all19:31
→ GaMeR64 joined19:31
ChrisSc since todays update of tvheadend to current head tvh wont start as a service anymore19:32
DoozerBot Commit in 'master ' by Jaroslav Kysela [ files modified ] http://bit.ly/1Lv4Crf19:32
WEBUI JS: Show the disk space (free/total)19:32
ChrisSc im able to start it manually but without my prevoius config19:32
any known bug in the current verion or should i purge my config and start with a new one19:33
← G64 left19:33
perexg ChrisSc: where it stops ?19:33
DoozerBot Build #11534: v4.1.299 for precise-amd64 completed http://bit.ly/1Lv4Crl 19:33
perexg ChrisSc: could you run tvh from cmdline using the old config ?19:34
ChrisSc perexg: syslog isnt showing anything but that the old process got killed19:34
if i start tvh without any argument from shell, it uses a new config19:34
DoozerBot Build #11535: v4.1.299 for precise-i386 completed http://bit.ly/1Lv4Ezs 19:35
perexg ChrisSc: copy the config to new directory and use '-c <thisdir>' argument..19:35
← adamsutton left19:36
ChrisSc sry, i was wrong: syslog reports "killed by SEGV signal when attempting to start tvh19:36
DoozerBot Build #11536: v4.1.299 for trusty-amd64 completed http://bit.ly/1Lv4M1S 19:36
ChrisSc config was located in .hts/tvheadend, wasnt it?19:36
DoozerBot Build #11537: v4.1.299 for trusty-i386 completed http://bit.ly/1Lv4Oqm 19:38
→ adamsutton joined19:38
ChanServ set mode: +o19:38
ChrisSc perexg: using your suggested command, tvh starts but reports [ ERROR] access: No access entries loaded19:39
and seems to create a new config19:39
perexg you yould point '-c ' to directory with 'config' file.... and probably 'epgdb.v2' should be there..19:40
also check permissions..19:41
ChrisSc path should be correct19:41
how should the permissions look like?19:42
perexg ChrisSc: if you run tvh as an user - just ensure that this user is owner of all files in this directory (chown -R)..19:43
(assuming you're working with a config backup)19:43
ChrisSc in fact it seems like my config got already overwritten, epgdb is empty, config states a different server name than before19:44
perexg https://tvheadend.org/projects/tvheadend/wiki/Development19:46
ChrisSc ...I dont have a backup...19:47
nevermind, if i completely reset the config, tvh should be able to start as a service again, shouldnt it?19:48
perexg ChrisSc: I wrote copy the config to new directory..19:48
ChrisSc: It should but it does not resolve the bug..19:48
ChrisSc yeah, i did that. the original dir must habe been overwritten before19:49
i've purged the .hts/tvheadend dir (and therefore expect tvh to start with a new config), however it still reports the following in syslog after "sudo service tvheadend start"19:51
Jun 24 21:48:42 Server kernel: [ 1241.040154] init: tvheadend main process (5840) killed by SEGV signal19:51
Jun 24 21:48:42 Server kernel: [ 1241.040178] init: tvheadend main process ended, respawning19:51
perexg try to gather coredump / backtrace : https://tvheadend.org/projects/tvheadend/wiki/Debugging19:53
ChrisSc will try19:55
← DereC left19:59
← nils_2 left20:09
→ nils_2 joined20:10
→ mac0815 joined20:10
mac0815 hi someone a hint what this log entry could mean ?20:11
satip: SAT>IP DVB-S Tuner #1 (192.168.100.65) - RTSP error -5 (Input/output error) [6-404]20:11
← mac0815 left20:51
← IvanG left20:51
DoozerBot Commit in 'master ' by Jaroslav Kysela [ file modified ] http://bit.ly/1GIEWmx21:03
WEBUI JS: Add tooltips for 'Play' column - DVR21:03
Commit in 'master ' by Jaroslav Kysela [ file modified ] http://bit.ly/1GIEWmA21:03
WEBUI JS: Add free/total qtip for disk space21:03
Commit in 'master ' by Jaroslav Kysela [ file modified ] http://bit.ly/1GIEVzd21:03
WEBUI JS: Add free/total qtip for disk space (small correction)21:04
Commit in 'master ' by Jaroslav Kysela [8 files modified ] http://bit.ly/1GIEWmC21:04
intl: another transifex update21:04
Build #11538: v4.1.303 for precise-amd64 completed http://bit.ly/1GIEVzf 21:05
Build #11539: v4.1.303 for precise-i386 completed http://bit.ly/1GIF31v 21:07
Build #11540: v4.1.303 for trusty-amd64 completed http://bit.ly/1GIFfOh 21:08
Build #11541: v4.1.303 for trusty-i386 completed http://bit.ly/1GIFsRy 21:10
→ ProfYaffleAsWell joined21:16
← ChrisSc left21:19
ProfYaffleAsWell Was just looking at the new disc space function and wondering about kB vs KB... found this... https://xkcd.com/394/21:19
and then melted my head with WIkipedia's kibibytes, mebibytes, and gibibytes, so just surrendered - KB, MB and GB it will remain...21:23
mpmctoo ProfYaffleAsWell: :|21:25
bow hehe21:36
it is annoying... as if there are any valid reasons to use the base 10 prefixes in any context thats computing related (unless you're trying to sell bandwidth or storage and wish to inflate the numbers)21:37
[-T-] perexg, maybe the diskspace values fonts are too big :)21:41
← hans_s left21:45
ProfYaffleAsWell bow: I'm old-school, I struggle with this new decimal nonsense... what was wrong with binary, it was all ones and zeroes when I was a boy...21:50
I did like the 'drivemaker's kilobyte' from that xkcd cartoon, though21:50
← ProfYaffleAsWell left21:54
mpmctoo ProfYaffleAsWell/ProfYaffle: Must have been awful to be alive when computers used punch cards ;)21:59
← ant_thomas left22:28
ProfYaffle mpmctoo, naah, it was all beads on sticks to me :)22:45
← goroboro left23:34

Logs Search ←Prev date Next date→ Channels Documentation