[geos-devel] Thoughts on in/outlining and robustness

[Mateusz Loskot]

Martin Davis wrote:
> strk, while I commend your valiant efforts to reconcile the behaviour
>  of inlined VS outlined functions and their effect on GEOS 
> robustness, I'm wondering whether this effort can ever pay off.


IMO considering inline/non-inline stuff at this point is a premature
optimization, and it will cause more problems than it can help at the
moment.
Especially, that optimization made by inlining can help in 5% ? roughly
estimating, because GOES core is full of virtual polymorphism, and in
most cases it can't be inlined even if requested.
But calls that can be inlined, won't introduce remarkable increase of
performance.
That's what I'm almost sure from my intuition.


> I think the best place to direct effort is in solving the overall 
> problem of robustness in the fundamental algorithms.  Hopefully this 
> will be solid enough that it will be immune to the inline/outline FP 
> differences.


Definitely agreed.
Less virtual polimorphism, more usage of compile-time checks and strong
type-safety + less classes but clean API.
That's what I'd focus on regarding GEOS future.

I'm still really affraid of threads like:
http://geos.refractions.net/pipermail/geos-devel/2005-January/001197.html


> My impression is that it would be a lot of incredibly finicky work to
>  "solve" the inline/outline difference, without improving robustness
>  all that much.  Or am I missing something?


Martin, I agree with you.


Cheers
-- 
Mateusz Loskot
http://mateusz.loskot.net