123 replies to this topic

[littlesat]

Always worked perfect != Is perferct...

Edited by littlesat, 29 November 2015 - 15:33.

[athoik]

We removed from tuneExt the use decoding index variable. Maybe hbbtv uses tuneExt with that variable? We can create a wrapper if so.

[christophecvr]

We removed from tuneExt the use decoding index variable. Maybe hbbtv uses tuneExt with that variable? We can create a wrapper if so.

Well that I dont't, But that is also why I only could produce a very brutal patch to just reverse.

 

But I do have no objection to try out a patch with a wrapper .

[Erik Slagter]

We removed from tuneExt the use decoding index variable. Maybe hbbtv uses tuneExt with that variable? We can create a wrapper if so.

VU+' HbbTV uses a piece of python code to scan for the AIT, I don't think it will access a function parameter (at most a class member).

 

I threw my VU+Duo2 somewhere in the basement a few weeks ago, so I can't easily test. Do you have one handy to try to revert this.

[athoik]

Here is the tuneExt with extra parameter use_decode_demux. Can you give a try?

diff --git a/lib/dvb/pmt.cpp b/lib/dvb/pmt.cpp
index 98fc9a5..76a38ab 100644
--- a/lib/dvb/pmt.cpp
+++ b/lib/dvb/pmt.cpp
@@ -769,6 +769,11 @@ int eDVBServicePMTHandler::tune(eServiceReferenceDVB &ref, int use_decode_demux,
        return tuneExt(ref, s, NULL, cue, simulate, service, type, descramble);
 }

+int eDVBServicePMTHandler::tuneExt(eServiceReferenceDVB &ref, int use_decode_demux, ePtr<iTsSource> &source, const char *streaminfo_file, eCueSheet *cue, bool simulate, eDVBService *service, serviceType type, bool descramble)
+{
+       return tuneExt(ref, source, streaminfo_file, cue, simulate, service, type, use_decode_demux);
+}
+
 int eDVBServicePMTHandler::tuneExt(eServiceReferenceDVB &ref, ePtr<iTsSource> &source, const char *streaminfo_file, eCueSheet *cue, bool simulate, eDVBService *service, serviceType type, bool descramble)
 {
        RESULT res=0;
diff --git a/lib/dvb/pmt.h b/lib/dvb/pmt.h
index 9c3949e..171e85a 100644
--- a/lib/dvb/pmt.h
+++ b/lib/dvb/pmt.h
@@ -149,6 +149,7 @@ public:

        /* deprecated interface */
        int tune(eServiceReferenceDVB &ref, int use_decode_demux, eCueSheet *sg=0, bool simulate=false, eDVBService *service = 0, serviceType type = livetv, bool descramble = true);
+       int eDVBServicePMTHandler::tuneExt(eServiceReferenceDVB &ref, int use_decode_demux, ePtr<iTsSource> &source, const char *streaminfo_file, eCueSheet *cue, bool simulate, eDVBService *service, serviceType type, bool descramble);

        /* new interface */
        int tuneExt(eServiceReferenceDVB &ref, ePtr<iTsSource> &, const char *streaminfo_file, eCueSheet *sg=0, bool simulate=false, eDVBService *service = 0, serviceType type = livetv, bool descramble = true);

[littlesat]

Sounds reasonable to me... but this is still a work-a-round... but more a backwards compatible work-a-round that might fix more "plugin" issues. there is nothing in it like if VU+ then... When this indeed solve the issue we should add a "general" comment to it why this is really needed.... (backwards compatible as previously this function needs a use_decode_demux integer)...

 

On the "other hand" I still think it is better to solve the function call in the HbbTV plugin of VU when this is really the case...(I'm not aware of any other plugin with the same issue yet).

 

But my Vu's are now at the Attic... so I cannot quickly check aswell... Sorry....

Edited by littlesat, 29 November 2015 - 19:10.

[Erik Slagter]

There must be surely SOMEONE that has a VU+ receiver ready?

[christophecvr]

@athoik I tried but it even doers not compile

[athoik]

In header file use just tuneExt without eDVBServicePMTHandler.

[christophecvr]

I do have very last clean e2 but there is a problem with you're patch

*** Manual patch resolution mode ***
Dropping to a shell, so patch rejects can be fixed manually.
Run "quilt refresh" when patch is corrected, press CTRL+D to exit.

Applying patch e2-demuxer.patch
patching file lib/dvb/pmt.cpp
Hunk #1 succeeded at 769 with fuzz 1.
patching file lib/dvb/pmt.h
Hunk #1 FAILED at 149.
1 out of 1 hunk FAILED -- saving rejects to file lib/dvb/pmt.h.rej
Applied patch e2-demuxer.patch (forced; needs refresh)
christophe@stefan2l:~/openpli40-gst1/openpli-oe-core/build/tmp/work/vuduo2-oe-linux/enigma2/2.7+gitAUTOINC+d8b2ebe8b8-r0/git$

I first to try (before i used a patch file to do this little changes on a fresh e2 pull it in the mean time commit d8b2ebe8b8ae5a6757e0c104f25dfbb106d34a66) but compile failure.

 

I did a new fresh pull with You're file in a patch faile and like You can see it fails to patch.

[Erik Slagter]

Just checkout enigma itself, change it in the source and compile enigma, no problem then.

[athoik]

Use the following patch:

Attached Files

•  tuneExt.patch   1.6KB   6 downloads

[christophecvr]

Tried with above patch

 

Sorry no hbbtv.

[athoik]

Are there any enigma2 logs?

What exacly hbbtv calls that fails on vu?

How demuxer is called from hbbtv?

[christophecvr]

I do not have a clue , it's all official licensed software . Not a unofficial illegal hack like is used on et boxes. But it just needs all and all channel demux pid info before it will open any url there is no debug about it its' protected. With You're change you somehow filtered out extreme important dvb service channel info required due to licence issues before a hbbtv channel is allowed. There may be not any basic channel info filtered that's it. and point end of line. Else the some needed info which is just simply required to stay official can't be used. The official channel info is there for a reason. Do not filter it out as we can see now that it is required. Off course if the intention is to only use et boxes and then non unofficial hacked actually illegal hbbtv services off you go.

 

But .....  vu is paying licence fees for every customer. Et uses vu hacked illegal software that is what's happening.

 

Simply do not filter any channel service info pid info or whatever they are needed and after all do not harm anything.

 

By the way pli4 does not build att all anymore due to a problem in enigma2-plugins ? (with or without patch)

 make[3]: Leaving directory `/home/christophe/openpli40-gst1/openpli-oe-core/build/tmp/work/vuduo2-oe-linux/enigma2-plugins/experimental-gitAUTOINC+3ac116e130-r0/git/transcodingsetup/meta'
| Making all in po
| make[3]: Entering directory `/home/christophe/openpli40-gst1/openpli-oe-core/build/tmp/work/vuduo2-oe-linux/enigma2-plugins/experimental-gitAUTOINC+3ac116e130-r0/git/transcodingsetup/po'
| make[3]: *** No rule to make target `all'.  Stop.
| make[3]: Leaving directory `/home/christophe/openpli40-gst1/openpli-oe-core/build/tmp/work/vuduo2-oe-linux/enigma2-plugins/experimental-gitAUTOINC+3ac116e130-r0/git/transcodingsetup/po'
| make[2]: *** [all-recursive] Error 1
| make[2]: Leaving directory `/home/christophe/openpli40-gst1/openpli-oe-core/build/tmp/work/vuduo2-oe-linux/enigma2-plugins/experimental-gitAUTOINC+3ac116e130-r0/git/transcodingsetup'
| make[1]: *** [all-recursive] Error 1
| make[1]: Leaving directory `/home/christophe/openpli40-gst1/openpli-oe-core/build/tmp/work/vuduo2-oe-linux/enigma2-plugins/experimental-gitAUTOINC+3ac116e130-r0/git'
| make: *** [all] Error 2
| ERROR: oe_runmake failed

[athoik]

Just run enigma2 and post the outputs.

There should be something in the logs.

We are making our hobby after all, right?

[wortellica]

I hope you can solve this problem 

[christophecvr]

Just run enigma2 and post the outputs.

There should be something in the logs.

We are making our hobby after all, right?

Well that's now just THE problem it does show nothing.

 

Also (into /var/volatile/log/messages) 0,0 info if it was a drivers issue usually you find a problem or warning there.

 

That wit the black and white that before the commit https://github.com/O...9bb9b8a9c96eff8

 

All was 100 % ok. After not.

 

And basically really do not understand any piece off code you changed but well that some info is filtered out that the only thing i see in it. This filtering out was not required.

 

And the longer the more I think it's just some info which is crucial to let work the hbbtv off vu plus which is missing now. At the same time vuplus does it not like that because they want to annoy the persons, but cause they HAVE to due to licensing issues.

 

wortellica has two boxes. And with two vu boxes You can use streamconverter. A vu utility. Whit that utility You can connect to the other box using the ip of the box and then download the bouquets list or simply one bouquets. On that bouquets you can open a channel. This channel is then just streamed on the client vu box while the tuner off the host box is used. Well then the client box can use the hbbtv ... gues that's a valuable info. Unfortunatelly I can't check that by me since i do have only one vu box my other box is a dm8000 and then that streamconverter does not work.

 

So to resume all

 

I already tested all possible options from old drivers till most recent all possible off hbbtv applications and opera-hbbtv. Just many times a downgraded the drivers o yes and I forgot even the linux kernel. That far I went.

Due to this downgrades yes I really reflashed the box many many times even in such extend that it looks like my duo2 want's to run away when I approach him with a usb stick.

 

result off all noop (enigma2 self was well kept on autorev).

 

I finally started then with looking if it was from some e2 revision. For that I used a specific vu hbbtv revision off which I knew it worked before inclusif the dumpait and opera-hbbtv (this last for gst-1 image i always work with gst-1 branch)

Then I hit after more then one day trials the exact commit where it was ok and the first where it was not ok (also to downgrade e2 ok it's simple using opkg remove enigma2 --force-depends but to have a clean search i prefered to do a flash and I finished to have nightmare's that my duo2 attacked me cause i did not respect him).

 

It is black and white ok by every commit before https://github.com/O...9bb9b8a9c96eff8 .

 

After that I used again gradually a upgrade off vu plus kernel and drivers untill the very recent ones hbbtv just working perfect. Then i upgraded hbbtv self till last openvuplus version.

 

SRC_URI = "git://code.vuplus.com/git/dvbapp.git

 

simply SRC_URI = "git://code.vuplus.com/git/dvbapp.git;protocol=http;branch=${BRANCH}  and the branch is : vuplus_experimental

the srcrev delivered was so what the last avbl perhaps now after one week its' a bit further but does give nothing extra.

 

Then opera tv yes comes out off openvuplus_3.0  (there the gst version is important i do have the very last date SRC_DATE = "20151001_1") The dumpait to last srcrev "c0616d5d37f32977d944a38d76b5cf7c2e5dc227" master branch.

 

With all enigma2 before commit https://github.com/O...9bb9b8a9c96eff8 .

 

It runs perfect and actually the last hbbtv versions are much faster then the previous one.

 

After commit https://github.com/O...9bb9b8a9c96eff8 .

 

Nothing works unless I apply my brutal reverse patch.  And yes I call it brutal since it is and not met to be applied by a beta version however its the collection off all possible culprit code. And for testers who build there own images this is a tempo solution actually almost for all who are using pli4 with gst-1 except the ones who are nice testers with the MastaG images.

 

The Erik Slagter comments sorry but ......... ok he could be right about the drivers but the longer i searched checked and cross-checked I'm shure he's wrong .

 

And its' a service info needed before the hbbtv may open any url, end off line period . The reason license issue. This info is now filtered by the dvb service demuxer while it was not before.

 

And a hacked illegal version of hbbtv and browser on et work's ???? Yes ET the users do not give a shit about the license which vu MUST do.

Edited by christophecvr, 29 November 2015 - 23:41.

[athoik]

 234                 try:
 235                         from Screens.InfoBarGenerics import gHbbtvApplication
 236                         self.m_vuplus = gHbbtvApplication.getUseAit()
 237                 except:
 238                         self.m_vuplus = False


 315         def get_autostart_application(self):
 316                 if self._applicationList is None:
 317                         service = self.session.nav.getCurrentService()
 318                         info = service and service.info()
 319                         if info is not None:
 320                                 sid = info.getInfo(iServiceInformation.sSID)
 321                                 onid = info.getInfo(iServiceInformation.sONID)
 322                                 tsid = info.getInfo(iServiceInformation.sTSID)
 323                                 name = info.getName()
 324                                 vbcfg.g_channel_info = (sid, onid, tsid, name)
 325 
 326                                 pmtid = info.getInfo(iServiceInformation.sPMTPID)
 327                                 demux = info.getInfoString(iServiceInformation.sLiveStreamDemuxId)
 328                                 vbcfg.DEBUG("demux = %s, pmtid = 0x%x, sid = 0x%x" % (demux, pmtid, sid))
 329 
 330                                 from aitreader import eAITSectionReader
 331                                 reader = eAITSectionReader(demux, pmtid, sid)
 332                                 if reader.doOpen(info, self.m_vuplus):
 333                                         reader.doParseApplications()
 334                                         reader.doDump()
 335                                 else:
 336                                         vbcfg.ERR("no AIT")
 337 
 338                                 try:
 339                                         self._applicationList = reader.getApplicationList()
 340                                 except:
 341                                         pass


  14 class eAITSectionReader:
  15         def __init__(self, demux, pmtid, sid):
  16                 self.mVuplusBox = False
  17                 self.mInfo = None
  18                 self.mAppList  = []
  19                 self.mDocument = None
  20                 self.mCommand  = "%s --demux=%s --pmtid=%x --serviceid=%x"%(DUMPBIN, demux, pmtid, sid)

VU+ uses dumpait utility to get the url.

The dumpait utility is in /usr/lib/enigma2/python/Plugins/Extensions/HbbTV/dumpait

The dumpait utility uses demux information from iServiceInformation.sLiveStreamDemuxId (line 327 on plugin.py)

The vbcfg.DEBUG doesn't print something because debug is disabled on hbbtv plugin (can be enabled on vbcfg.py)


Now should be clear where to look and what it fails.

[athoik]

christophecvr, can you test this one?

diff --git a/lib/dvb/pmt.cpp b/lib/dvb/pmt.cpp
index 98fc9a5..eec22f3 100644
--- a/lib/dvb/pmt.cpp
+++ b/lib/dvb/pmt.cpp
@@ -710,7 +710,7 @@ int eDVBServicePMTHandler::getDecodeDemux(ePtr<iDVBDemux> &demux)
        int ret=0;
                /* if we're using the decoding demux as data source
                   (for example in pvr playbacks), return that one. */
-       if (m_use_decode_demux)
+       if (m_pvr_channel)
        {
                demux = m_demux;
                return ret;