[asterisk-dev] [Code Review] Force stream to really stop when ast_stopstream() is called

[Russell Bryant]

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviewboard.asterisk.org/r/372/#review1100
-----------------------------------------------------------

Ship it!


This looks like a good change to me.  Nice work, Tilghman!

- Russell


On 2009-09-18 12:07:04, Tilghman Lesher wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviewboard.asterisk.org/r/372/
> -----------------------------------------------------------
> 
> (Updated 2009-09-18 12:07:04)
> 
> 
> Review request for Asterisk Developers.
> 
> 
> Summary
> -------
> 
> In issue 15129, the reporter has an issue where GSM frames are getting queued after the writeformat has changed to ulaw.  I tracked this down to the use of reference counts within filestreams, whereby when the stream is interrupted, we no longer immediately cut off queuing frames from that stream; instead we wait until the last reference count from the filestream is decremented to stop queuing frames.  By forcing the stream to stop immediately in ast_stopstream() (which calls ast_closestream), even though the filestream is not closed until later, we avoid this potential for getting an invalid format frame written to the channel.
> 
> 
> This addresses bug 15129.
>     https://issues.asterisk.org/view.php?id=15129
> 
> 
> Diffs
> -----
> 
>   /branches/1.6.1/main/file.c 219412 
> 
> Diff: https://reviewboard.asterisk.org/r/372/diff
> 
> 
> Testing
> -------
> 
> Playback runs fine, and is interrupted by DTMF fine.  Unfortunately, since I cannot replicate the reporter's scenario, I am unable to confirm whether this fixes his issue.
> 
> 
> Thanks,
> 
> Tilghman
> 
>