Carsten Haitzler (The Rasterman) wrote:
> On Tue, 06 Sep 2005 17:29:30 +0200 Kim Woelders <[EMAIL PROTECTED]> babbled:
> 
> 
>>Actually, I fixed this in the e16 versions of these functions *long* 
>>ago. I think the implementation there (e16/e/src/ecore_e16.c) is more 
>>efficient than the current ecore one, as it avoids the unneeded malloc() 
>>when sizeof(int) == sizeof(long).
>>I did mention to raster that I believed there were 64 bit issues with 
>>this code, but he didn't agree, and I didn't persue the matter.
> 
> 
> i don't rememebr that bit - mind you i did leartn my lesson about the long
> thing for 32bit long (hahahah - nice pun) ago. i would be most surprised if i
> made the msitake again. i caught it as it was about to be "fixed" wrongly too
> today. i think we may have crossed wires as to which code particularly was not
> 64bit clean - different bits of code. yeah - i know the malloc isnt that
> efficient - but it beats special casing it in all calling code. you need to
> copy it - and in todays libc's the malloc is not so expensive really. e17/efl
> is insanely malloc heavy, and it's actually amazing how little it affects it.
> 
Ok, now I am going to persue this one :)
It seems to me that you say that we always have to malloc and copy.
I say we only have to malloc and copy if sizeof(int) != sizeof(long).
In my opinion it is worth while to avoid malloc/copy in this kind of
low-level code, if possible.

/Kim


-------------------------------------------------------
SF.Net email is Sponsored by the Better Software Conference & EXPO
September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
_______________________________________________
enlightenment-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/enlightenment-devel

Reply via email to