Customized rtmpdump binaries with patch file

[lorus]

Originally Posted by KSV

Try adding -v (--live) switch.

OMG, stupid me ... big thx

[RedPenguin]

I have compiled this release into librtmp's for various systems (iOS (Apple TV, iPhone, iPod Touch), RasPi, Linux 32&64) for anyone wanting this release in their XBMC.

I can easily put exectuables up also if anyone needs but I don't know of any practical reasons off the top of my head for devices like RasPi and iOS.

http://www.mediafire.com/?x4cvp5hl4m9xr

[KDN165]

Thank you Redpenguin.

I have a problem with your xbmc apk for android.

(Sorry for my english because I´m Spanish). I installed your xbmc apk and when I click to run it crash.

Say: La aplicación XBMC (proceso org.xbmc.xbmc) se ha interrumpido inesperadamente. Inténtelo de nuevo

[bentzi]

very weird problem: I'm using unofficial KSV version. on one of my boxes it doesnt work properly: I investigated and found that the server asks for an answer which on other boxes is calculated well, but on this box no... checked python, and there it calculates well
so, on this thing:
enc = AMF_EncodeNumber(enc, pend, exp(atan(sqrt(VerificationNumber))) + 1)

the "bad" box returns ((VerificationNumber) + 1)

any idea? I'm totally cluless

btw, it's a linux box

[RedPenguin]

Originally Posted by KDN165

Thank you Redpenguin.

I have a problem with your xbmc apk for android.

(Sorry for my english because I´m Spanish). I installed your xbmc apk and when I click to run it crash.

Say: La aplicación XBMC (proceso org.xbmc.xbmc) se ha interrumpido inesperadamente. Inténtelo de nuevo

At first I had a mix up with signing the modified APK but I fixed it and most folks report it's working great.

http://forum.xbmc.org/showthread.php?tid=162307 is the support thread I maintain for my librtmp files and the APK if you wanna check it out.

[KSV]

Originally Posted by bentzi

very weird problem: I'm using unofficial KSV version. on one of my boxes it doesnt work properly: I investigated and found that the server asks for an answer which on other boxes is calculated well, but on this box no... checked python, and there it calculates well
so, on this thing:
enc = AMF_EncodeNumber(enc, pend, exp(atan(sqrt(VerificationNumber))) + 1)

the "bad" box returns ((VerificationNumber) + 1)

any idea? I'm totally cluless

btw, it's a linux box

Post the full debug log (-V) on pastebin. if it works on everything else then it can't be code's problem.

[RedPenguin]

Also KSV or somebody else not sure if you have a cable TV subscription or not to test:

http://www.cnn.com/video and test their "Live TV" section.

But I have noticed that the rtmpdump code must have changed sometime after 2011-07-11 because like I said before if I use KSV or normal versions since then it constantly results in "Invalid AMF Data" with rtmpsrv from RTMPDumpHelper.

From what I can tell, they (edgecfs) are simply changing the connection data on every play request and using weird characters like the Spanish upside down exclamation point.

Not sure why but the older rtmpdumps can simply flow with these odd connection data but the newer cannot and just error out, so I was wondering if this could be fixed via a patch.

Here is a sample of what I mean:

rtmpdump -r "rtmpe://cp99921.live.edgefcs.net:1935/live" -a "live?auth=daEdabcbS
bfdrdlcNa.dudNbld5cebQaKan-brBoEc-hca-lxy_GFs&aifp=v001&slist=tvehln104_1_low@s8
9305" -f "WIN 11,7,700,169" -W "http://z.cdn.turner.com/xslo/cvp/plugins/akamai/
streaming/osmf1.6/2.11.3/AkamaiAdvancedStreamingPlugin.swf" -p "http://www.cnn.c
om/video/#/video/cvptve/cvpstream2" --live -C Θ: -y "tvehln104_1_low@s89305?auth
=daEdabcbSbfdrdlcNa.dudNbld5cebQaKan-brBoEc-hca-lxy_GFs&aifp=v001&slist=tvehln10
4_1_low@s89305" -o "tvehln104_1_low@s89305.flv"

ERROR: Invalid AMF parameter: Θ:

It's as if the command would easily work with zero issues if it was not for the strange ASCII characters.

Though not sure if this would be something the actual RTMPDump would need to code.

[bentzi]

here is the log:

the "bad" box-

DEBUG: RTMP_ClientPacket, received: invoke 34 bytes
DEBUG: (object begin)
DEBUG: Property: <Name: no-name, STRING: verifyClient>
DEBUG: Property: <Name: no-name, NUMBER: 120781448.00>
DEBUG: Property: NULL
DEBUG: Property: <Name: no-name, NUMBER: 96.73>
DEBUG: (object end)
DEBUG: HandleInvoke, server invoking <verifyClient>
DEBUG: VerificationNumber: 96.73
DEBUG: (object begin)
DEBUG: Property: <Name: no-name, STRING: _result>
DEBUG: Property: <Name: no-name, NUMBER: 120781448.00>
DEBUG: Property: NULL
DEBUG: Property: <Name: no-name, NUMBER: 97.73>
DEBUG: (object end)

the number given back in response is 97.73 (should be 5.35)
here is the "good" box-

DEBUG: RTMP_ClientPacket, received: invoke 34 bytes
DEBUG: (object begin)
DEBUG: Property: <Name: no-name, STRING: verifyClient>
DEBUG: Property: <Name: no-name, NUMBER: 120781448.00>
DEBUG: Property: NULL
DEBUG: Property: <Name: no-name, NUMBER: 86.61>
DEBUG: (object end)
DEBUG: HandleInvoke, server invoking <verifyClient>
DEBUG: VerificationNumber: 86.61
DEBUG: (object begin)
DEBUG: Property: <Name: no-name, STRING: _result>
DEBUG: Property: <Name: no-name, NUMBER: 120781448.00>
DEBUG: Property: NULL
DEBUG: Property: <Name: no-name, NUMBER: 5.32>
DEBUG: (object end)
DEBUG2: RTMP_SendPacket: fd=4, size=29

it's a mediadirect.ro server, so the number is a random number given by server to get the response
rest of the log is the same for both boxes (until hang up)

[KSV]

@bentzi

your problem seems like a bug in math library (libm.so). try upgrading glibc to newer version. it's also possible that your compiler optimization is breaking it. try compiling static binary with no optimizations.

LDFLAGS="-static" OPT="-g -O0"

@RedPenguin

does this happens also with rtmpsuck? also post the verbose log of successful and failed attempt.

-z 2> Debug.txt

[RedPenguin]

Originally Posted by KSV

@RedPenguin

does this happens also with rtmpsuck? also post the verbose log of successful and failed attempt.

-z 2> Debug.txt

For some reason rtmpsuck has never worked with these streams (rtmpsuck claims it fails and video continues in browser) this is what it always does with 2011-07-11 and identical error with your latest compile.

http://pastebin.com/VkbqL0vQ