[PATCH] dix: use BUG_WARN for xi2mask size issues

[Chase Douglas]

On 11/09/2011 03:55 PM, Alan Coopersmith wrote:
> On 11/08/11 21:15, Peter Hutterer wrote:
>> Yes, we're likely corrupting memory here but really this is unlikely to be
>> triggered other than a real bug in the server. In which case a stacktrace is
>> going to be more useful than any silent error handling.
> 
> Is it going to cause static analyzers to issue false alarms because we're
> checking for cases that should cause failures but allowing them to happen?

I think this would be better for static analyzers. If there's a return
on error, the analyzer may see that as "correct" error handling.
However, allowing it to go through may be caught by the analyzer as a bug.

That's assuming static analyzers could see this type of bug at all.

-- Chase