Re: [AD] malloc

[aj <aj@xxxxxxxxxx>]

> > i know the whole thing is going to get a design, and changes (as 
> > mentioned already), but is there any reason why we still cant change all 
> > the malloc's to _al_malloc() in the mean time ?
> >    
>
> It will be a waste of your time if we have to change it again later.

I do not believe it is a waste of my time.  I want to see it done.  It 
is very important for me.  Without trying to sound to whingy, im not too 
impressed with the rate of development of allegro, I know things take 
time, and making a good design is very important, but i have been 
asking, requesting, trying, to have the powers that be listen to my 
needs for quite a long time now, and i dont seem to be making much 
progress, each time i have brought up this particular issue, i was told 
it would be done in the future, or no, that will have to wait until 
after 420 was released.  And rightly so, all the reasons for delaying my 
requests are justified, im not asking for anything that will derail the 
roadmap, Im trying to have my needs met without 'rocking the boat', but 
am getting ever more frustrated.  Im not sure if im reading this right, 
but now it seems i am again being politely declined the chance to get 
anything done.
The release schedule (based on past experience) tells me that if i have 
to wait for 4.3 to get this done, i might be waiting another 6 months 
(conservative estimate :P )

I hope this isn't the case, that someones future design for a non-C GC 
language implementation is going to stop C related changes that are 
valid for users (me) _right now_.
Im not asking to change any functionality, just  malloc replaced with 
_al_malloc(),  that is all.

[deleted paragraph]  (one step short of a Korval style rampage) 
[/deleted paragraph]

As it stands at the moment, allegro can be severly performance hit, if 
someone compiles it with SSE turned on, P4, and AMD CPUs both have SSE.  
Intel are now recommending that code favour SSE instead of MMX (cpu 
manufacturers have already signalled the death of MMX), so allegro and 
its i386 code  + some MMX, is now closer to being a 20th century library 
than ever before.  If allegro is a games library, then it needs to start 
targeting the types of CPUs that gamers have, and its not Pentium1's.   
Im trying to develop SSE stuff, that may benefit allegro, might even 
drag it into this century.


aj.