[PATCH libXi] Allocate enough memory for raw events + extra data.

[Jeremy Huddleston]

On May 2, 2011, at 11:00 PM, Jeremy Huddleston wrote:
> On May 2, 2011, at 10:00 PM, Peter Hutterer wrote:
>> 
>> re "Idempotent operation"
>> I'm staring hard at those but don't really see the issue.
>> SetReqLen(req, foo, foo); is valid and for any foo that is non-zero
>> shouldn't have the effect of "Assigned value is always the same as the
>> existing value". SetReqLen essentially does req->length += foo.
> 
> Yeah... "within expansion of macro..." results are a bit bad with clang.  I just opened a bug about that.  Hopefully future analytics will be better.  For now, at worse it's a dead assignment, and we can come back to it when clang gets better at printing those results.

Ah, I see it.  There's a "len = len" in the expression.  

I dislike macros.  Why don't we make all those inline functions instead?  That is so much cleaner.