Jean-Marc Lasgouttes a écrit :
"Martin" == Martin Vermeer <[EMAIL PROTECTED]> writes:
Would you care to explain the difference between what has been
there and what you have now?
Martin> Good question. I didn't see either that this had anything to
Martin> do with caching as Abdel claims, but didn't feel certain
Martin> enough to bring the matter up ;-)
Me too! ;)
Maybe my code is wrong but I'd really like to know why. All I am doing
is saving the pointer in the table. The second time a font that has the
same characteristics is required, instead of creating it again I pass
the one created before. So the contruction of a new font is skipped.
If the pointer retrieved in the table is equal to zero that means that
this particular font was not used before and we have to create it (via
new) and save its pointer to the table.
The older code was doing nothing as the table was filled with zero and
was never filled in anywhere!
Seriously Abdel, I think it is time to think again about profiling.
I am doing just that. This patch is the direct analysis of the profiling
done by Bennett, please read that and the analysis I made. Then, please
explain me why I am wrong when I think there is a problem with font
caching under MacOSX.
Do you have an AMD chip? In this case, you could try
http://developer.amd.com/cawin.aspx
Nope, intel here.
Otherwise, you could maybe find someone who has vtune...
I was trying to help MacOS port so I don't see how profiling under
wintel will help that as the two implementions of Qt4.1 have very
different performances.
Abdel.
