[Ffmpeg-devel] Coverity defect scan

[Corey Hickey]

Mike Melanson wrote:
>>>>Fixing these defects is also going to be a status thing I guess.  The
>>>>xine people loudly say how horrible the MPlayer code is at every
>>>>possible occasion.  Nevertheless their defect rate is considerably
>>>>higher than ours.
>>>
>>>	0.416 > 0.465? Back to school, Diego.
>>
>>
>>Haha, that comment sure motivated some people :)
>>
>>Nevertheless read the following in the article about Coverity on
>>lwn.net:
>>
>>  The projects with the lowest defect density include Ethereal, OpenVPN,
>>  Perl, and xmms; the all-time winner is xmms, with a total of six
>>  detected errors. At the other end of the scale, one finds Amanda,
>>  Firebird, NetSNMP, OpenLDAP, Samba, X, and Xine.
>>
>>http://lwn.net/Articles/174769/
>>http://lwn.net/SubscriberLink/174769/07c92421c084afbb/
>>
>>;-p
> 
> 
> 	Maybe you should read the authoritative site, such as it is, which 
> reports actual defect density figures.

I think those pages weren't incorrect -- they're just out-of-date by now.

The density appears to be calculated as:
1000 * current_defects / lines_of_code
rather than:
1000 * original_defects / lines_of_code

So, it's entirely possible that, before many bugs were fixed or marked
invalid, Xine had a higher reported density than MPlayer. Or, the
statistics were read at a time when many MPlayer bugs were
fixed/invalidated and relatively few Xine bugs were.

As another example, at this time AMANDA (see above) has one of the
lowest densities because the defect count has been reduced from 108 to 1.

Not that it really matters much, but I didn't want to see a flamewar
over data sampled at different times. :)

-Corey