RFC 17: Dynamic Array Sizing

[Frank Warmerdam]

Umberto Nicoletti wrote:
>> If you want to propose use of APR, you will need to write up an RFC.  I
>> think it would be relatively straight forward to re-engineer MapServer
>> to use APR services for stuff like mutexes and perhaps it would help
>> us with thread local data as well.
> 
> I have proposed apr just to avoid yet another reimplementation of
> dynamic arrays.

Umberto,

Ah, I hadn't realized this was the objective.  I very certainly would
not support introduction of a substantial additional dependency for
something this trivial.

 >  This would maybe save time and apr seems to be good
> for other things too, like you have suggested. I don't think I need to
> write another rfc, we could all agree on using apr for dynamic arrays
> as that is the subject of rfc17 and then write other rfcs for
> re-engineering other bits of mapserver to use the APR apis.

Well, I'm not willing to expand RFC17 to include APR.  Adding a
depenendency on APR definately deserves it's own RFC.

> <dreaming>
> We could even make mapserver an apache module and get a super fast
> mapserver interface.
> </dreamin>

I don't see that anything stands in the way of implementing this except
for someone motivated to make it happen.  I am personally dubious about
how performance benefit there would be in this, but it might be an
interesting experiment.

Best regards,
-- 
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush    | President OSGF, http://osgeo.org