[asterisk-dev] [Code Review] Add UUID support to Asterisk

[David Lee]

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviewboard.asterisk.org/r/2217/#review7461
-----------------------------------------------------------



/trunk/include/asterisk/uuid.h
<https://reviewboard.asterisk.org/r/2217/#comment14251>

    Since UUID strings are a fixed size (37 chars, including null terminator), I would be fine with a bare char* here.
    
    Either way, you should define AST_UUID_STRLEN so callers can properly size their strings before calling.



/trunk/include/asterisk/uuid.h
<https://reviewboard.asterisk.org/r/2217/#comment14252>

    This should at least take a char*. It seems more straightforward to do that than force callers to wrap their pointers in ast_str.



/trunk/main/uuid.c
<https://reviewboard.asterisk.org/r/2217/#comment14253>

    The concerns about a random generator would be 1) performance and 2) thread safety. You may want to address those here, too.



/trunk/main/uuid.c
<https://reviewboard.asterisk.org/r/2217/#comment14254>

    It would be useful to test for the existing of /dev/urandom and log an error if it doesn't exist.
    
    /dev/random is a blocking entropy source, and could dramatically affect performance if that's the only generator available.
    
    If neither are available, libuuid uses rand(), which isn't thread safe.



/trunk/tests/test_uuid.c
<https://reviewboard.asterisk.org/r/2217/#comment14255>

    UUID string format is defined in RFC 4122, section 3. Any UUID library should return a 36 character string.


- David


On Nov. 28, 2012, 10:16 a.m., Mark Michelson wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviewboard.asterisk.org/r/2217/
> -----------------------------------------------------------
> 
> (Updated Nov. 28, 2012, 10:16 a.m.)
> 
> 
> Review request for Asterisk Developers, Matt Jordan and David Lee.
> 
> 
> Summary
> -------
> 
> This is the first of many parts of the overall goal of improving Asterisk APIs. This adds a UUID API to Asterisk, eventually to be used as a method of uniquely identifying channels and potentially other objects.
> 
> The implementation here uses libuuid, which is packaged by all major Linux distributions. The API is basically a wrapper around libuuid's API, but it hides all implementation details in case we want to switch out libraries at some point.
> 
> Licensing for libuuid is a bit confusing. On one hand, this online man page (http://linux.die.net/man/3/libuuid) claims that the library is distributed under the LGPL version 2. However, when I downloaded the source for the library on my machine, the code had the 3-clause BSD license printed in the source. In either case, I believe use of external library is license-compatible with Asterisk.
> 
> Regarding specific implementation decisions, I think comments in the code should explain why I chose to do things certain ways. Please let me know if anything should be done differently or if you have ideas for tests beyond what I have already included.
> 
> I have no idea what is up with the changes listed in autoconfig.h.in. Those changes showed up when I ran the bootstrap.sh script after changing configure.ac. 
> 
> 
> This addresses bug ASTERISK-20725.
>     https://issues.asterisk.org/jira/browse/ASTERISK-20725
> 
> 
> Diffs
> -----
> 
>   /trunk/configure UNKNOWN 
>   /trunk/configure.ac 376724 
>   /trunk/include/asterisk/autoconfig.h.in 376724 
>   /trunk/include/asterisk/uuid.h PRE-CREATION 
>   /trunk/main/Makefile 376724 
>   /trunk/main/asterisk.c 376724 
>   /trunk/main/uuid.c PRE-CREATION 
>   /trunk/tests/test_uuid.c PRE-CREATION 
> 
> Diff: https://reviewboard.asterisk.org/r/2217/diff
> 
> 
> Testing
> -------
> 
> I have included a set of unit tests that exercise all the APIs. They pass when run locally.
> 
> 
> Thanks,
> 
> Mark
> 
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.digium.com/pipermail/asterisk-dev/attachments/20121128/a96ccc11/attachment.htm>