List Info

Thread: CR: Fix for Assert failure at clip end while playing mpeg4v/3gp clip with FCS URL




CR: Fix for Assert failure at clip end while playing mpeg4v/3gp clip with FCS URL
country flaguser name
United States
2007-09-26 04:17:11
Synopsis

========

This CR include fix for assertion failure at the clip end
when playing 
mpeg4v/3gp video clip with FCS URL.

Branch : HEAD, SERVER_BORABORA_RN
Suggested Reviewer : Jamie

Description
=========
While testing the mpeg4v/3gp video clip switchability, I
found that when 
we play the 3gp video clip, mp4fileformat notify a premature
stream 
done. Actually mp4fileformat maintains a cache of one packet
and on 
every get packet it kept coming packet in its cache and send
the present 
packet of the cache in function
CQTFileFormat::PacketReady(). But if an 
error status is passed to function
CQTFileFormat::PacketReady, it should 
send the  packet from its cache with status = HXR_OK to
function call 
m_pFFResponse->PacketReady(HXR_OK, pPacketToSend),
because it is sending 
a valid packet. But it is sending a valid packet with an
error status to 
function CPacketHandler::PacketReady() (and so a valid last
packet is 
discarded), which in turn causes the
PPM::Session::StreamDone get 
called. This PPM::Session::StreamDone call further become
the cause of 
failure of assertion "HX_ASSERT(!m_bIsLive &&
m_uEndPoint > 0);" at line 
5286 in file ppm.cpp. Note that we are not setting
m_uEndPoint in case 
of FCS, so it is always 0.

Now this fix will now keep the assertion and also save last
packet which 
was being lost.

Files Affected
==============
datatype/mp4/fileformat/qtffplin.cpp

Testing Performed
=================

Unit Tests:
-None

Integration Tests:
- Verified that no assertion is failing and last packet is
being displayed.

Leak Tests:
- I will test for leaks after mpeg4/3gp video switchability
changes.

Performance Tests:
- None

Build verified: win32-i386-vc7
Platforms Tested: win32-i386-vc7

QA Hints
=======
- None

Regards
-Dalvinder



_______________________________________________
Datatype-dev mailing list
Datatype-devhelixcommunity.org
http://lists.helixcommunity.org/mailman/listinfo/da
tatype-dev

  
Re: CR: Fix for Assert failure at clip end while playing mpeg4v/3gp clip with FCS URL
country flaguser name
United States
2007-09-26 15:08:04
Looks good.

Thanks,
Jamie

Dalvindar Singh wrote:
> Synopsis
> 
> ========
> 
> This CR include fix for assertion failure at the clip
end when playing 
> mpeg4v/3gp video clip with FCS URL.
> 
> Branch : HEAD, SERVER_BORABORA_RN
> Suggested Reviewer : Jamie
> 
> Description
> =========
> While testing the mpeg4v/3gp video clip switchability,
I found that when 
> we play the 3gp video clip, mp4fileformat notify a
premature stream 
> done. Actually mp4fileformat maintains a cache of one
packet and on 
> every get packet it kept coming packet in its cache and
send the present 
> packet of the cache in function
CQTFileFormat::PacketReady(). But if an 
> error status is passed to function
CQTFileFormat::PacketReady, it should 
> send the  packet from its cache with status = HXR_OK to
function call 
> m_pFFResponse->PacketReady(HXR_OK, pPacketToSend),
because it is sending 
> a valid packet. But it is sending a valid packet with
an error status to 
> function CPacketHandler::PacketReady() (and so a valid
last packet is 
> discarded), which in turn causes the
PPM::Session::StreamDone get 
> called. This PPM::Session::StreamDone call further
become the cause of 
> failure of assertion "HX_ASSERT(!m_bIsLive
&& m_uEndPoint > 0);" at line 
> 5286 in file ppm.cpp. Note that we are not setting
m_uEndPoint in case 
> of FCS, so it is always 0.
> 
> Now this fix will now keep the assertion and also save
last packet which 
> was being lost.
> 
> Files Affected
> ==============
> datatype/mp4/fileformat/qtffplin.cpp
> 
> Testing Performed
> =================
> 
> Unit Tests:
> -None
> 
> Integration Tests:
> - Verified that no assertion is failing and last packet
is being displayed.
> 
> Leak Tests:
> - I will test for leaks after mpeg4/3gp video
switchability changes.
> 
> Performance Tests:
> - None
> 
> Build verified: win32-i386-vc7
> Platforms Tested: win32-i386-vc7
> 
> QA Hints
> =======
> - None
> 
> Regards
> -Dalvinder
> 
> 
> 
>
------------------------------------------------------------
------------
> 
> _______________________________________________
> Datatype-dev mailing list
> Datatype-devhelixcommunity.org
> http://lists.helixcommunity.org/mailman/listinfo/da
tatype-dev

_______________________________________________
Datatype-dev mailing list
Datatype-devhelixcommunity.org
http://lists.helixcommunity.org/mailman/listinfo/da
tatype-dev

[1-2]

about | contact  Other archives ( Real Estate discussion Medical topics )