[mythtv] iptv recorder with enigma2 box

[Torbjorn Jansson]

there is another problem.
whenever i try to open up my iptv card in mythtv-setup the setup program 
freezes for a very long time.

in the log i can see:

2016-02-21 18:13:13.063067 I  VBox: Found 1 possible VBoxes
QIODevice::read (QNetworkReplyHttpImpl): device not open
2016-02-21 18:15:13.144049 I  VBox: GetDeviceDesc() failed for <Unknown>
2016-02-21 18:15:13.144087 I  VBox: Using UPNP to search for Vboxes (3 secs)
QIODevice::read (QNetworkReplyHttpImpl): device not open
2016-02-21 18:15:16.416803 I  VBox: Found 1 possible VBoxes
QIODevice::read (QNetworkReplyHttpImpl): device not open
QIODevice::read (QNetworkReplyHttpImpl): device not open
2016-02-21 18:17:16.494787 I  VBox: GetDeviceDesc() failed for <Unknown>
2016-02-21 18:17:16.497308 E  V4L2(): Could not open '':
                         eno: No such file or directory (2)

it took nearly 5 minutes from when i pressed return on the card until 
its settings opend.

what is VBox ?


On 2016-02-21 17:48, Torbjorn Jansson wrote:
> Hello
>
> i have tested the iptv support in mythtv (master and not fixes/0.27) and
> found some problems.
> my testing was done against my dvb box that is based on enigma2 and its
> web interface.
>
> iptv recorder was pointed at one of the urls from the web interface
>
> problem 1: url too long
> url got cutoff since it was longer than 128 character limit of
> capturecard.videodevice in the database.
>
> so i had to download the m3u file manually and put it on a web server to
> make url shorter
>
> question: can we extend that field, for example by doubling it so long
> urls work?
>
>
> problem 2: "#EXTM3U "
> channel scanning didn't work at all, it just complained about the file.
> apparently there is an extra space in the first line of the m3u file
> like this without quotes: "#EXTM3U "
>
> a simple trim near the beginning of: IPTVChannelFetcher::ParsePlaylist
> should do the trick
>
>
> problem 3: format of #EXTINF:
> next the format of EXTINF: is a bit different, like this:
> #EXTINF:-1,Animal Planet
>
> normally there is another number after the comma sign (channel number),
> like this:
> #EXTINF:0,1 Animal Planet
> Where 1 is the channel number to use
>
>
> fixing this would require changing the for example the first regexp in
> parse_extinf()
> but i don't fully understand the existing regex and what happens if we
> don't have any channum?
> in my case it doesnt exist in the m3u file, so allocating channums based
> on order in the file is probably best bet (maybe already done?)
>
>
> fixing above three problems results in successful scanning of channels.
> BUT then there is a 4th issue.
>
> problem 4: channel.serviceid=0
> field serviceid in channel table is 0, result is that the backend can't
> find the channel once it has successful started receiving data.
> there is errors in the log about not able to find program 0 in PAT
>
> only way to workaround this is specify #EXTVLCOPT:program= in the m3u file.
>
> i assume there is no way of convincing the backend to tune to the only
> working service there is?
>
> i think the streaming service in enigma2 webui is very simple, it takes
> a service ref url and based on that tunes the mux and applies pid
> filtering and then sends the data as is over http
> so there is no modification of PAT/PMT before sending it resulting in
> backend receiving lots of programs in the PAT but only one will work.
>
>
> i just want to see what you think about the problems above.
> i can probably make some patches for some of the issues, but problem 4 i
> dont know how to fix other than modifying the m3u file so it contains
> #EXTVLCOPT:program