[asterisk-dev] [Code Review] 3734: media formats: Add a 'none' format to prevent NULL format derefences; update unit tests

[Matt Jordan]

> On July 9, 2014, 5:40 p.m., Corey Farrell wrote:
> > /team/group/media_formats-reviewed-trunk/main/channel.c, lines 2936-2940
> > <https://reviewboard.asterisk.org/r/3734/diff/1/?file=62623#file62623line2936>
> >
> >     Looks like fmt is not used other than the if statement, do we really need a variable?  Also as discussed on IRC I'd like to see an assert so we can identify NULL format issues.

Since channels now have the none type, that shouldn't be needed any more. Asserts added as well.


- Matt


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviewboard.asterisk.org/r/3734/#review12533
-----------------------------------------------------------


On July 9, 2014, 5:17 p.m., Matt Jordan wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviewboard.asterisk.org/r/3734/
> -----------------------------------------------------------
> 
> (Updated July 9, 2014, 5:17 p.m.)
> 
> 
> Review request for Asterisk Developers.
> 
> 
> Repository: Asterisk
> 
> 
> Description
> -------
> 
> This patch does two things:
> 
> * It updates a few of the unit tests for some of the API changes. In particular, it focuses on adding some tests for formats with attributes and their expected behaviour. A few other non-format related unit tests were updated as well to handle off nominals detected during testing.
> 
> * It adds an 'ast_format_none' format. This format is a dummy format that can be used instead of a NULL pointer to prevent having to put NULL dereference checks into every place in the codebase. Channels are no assigned this format immediately upon creation, and their default capabilities are set to have it. As this format's codec has no translation (nor a representation in the RTP engine), it _shouldn't_ cause harm.
> 
> * A few NULL checks were put in anyway into key areas in a few modules. These were ones that were hit hard by the unit tests and prone to crashing if presented a NULL format.
> 
> 
> Diffs
> -----
> 
>   /team/group/media_formats-reviewed-trunk/tests/test_format_cap.c 418254 
>   /team/group/media_formats-reviewed-trunk/tests/test_format_cache.c 418254 
>   /team/group/media_formats-reviewed-trunk/tests/test_core_format.c 418254 
>   /team/group/media_formats-reviewed-trunk/tests/test_cel.c 418254 
>   /team/group/media_formats-reviewed-trunk/main/format_cap.c 418254 
>   /team/group/media_formats-reviewed-trunk/main/format_cache.c 418254 
>   /team/group/media_formats-reviewed-trunk/main/codec_builtin.c 418254 
>   /team/group/media_formats-reviewed-trunk/main/codec.c 418254 
>   /team/group/media_formats-reviewed-trunk/main/channel.c 418254 
>   /team/group/media_formats-reviewed-trunk/main/bridge_channel.c 418254 
>   /team/group/media_formats-reviewed-trunk/include/asterisk/format_cache.h 418254 
>   /team/group/media_formats-reviewed-trunk/include/asterisk/codec.h 418254 
> 
> Diff: https://reviewboard.asterisk.org/r/3734/diff/
> 
> 
> Testing
> -------
> 
> Unit tests pass.
> 
> There is a FRACK on shutdown, but it doesn't appear to be caused by this patch (things didn't run long enough without this patch before)
> 
> 
> Thanks,
> 
> Matt Jordan
> 
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.digium.com/pipermail/asterisk-dev/attachments/20140710/11570dd4/attachment.html>