Soeren Kuklau wrote:
> Yes, the current experimental build is quite good. But apparently some 
> other performance bugs were fixed after this build was made, which 
> causes this build to be worse than 2002-04-28 in five out of six tests 
> (as we've already been discussing via e-mail):
> 
> <http://soeren.mystfans.com/mozilla/dhtml-perf.html>

I don't see how you can draw that conclusion.... You have tested two builds:

1) April 22, with reflow patch
2) April 28, without reflow patch

and noted that #1 is slower than #2 at some things.  There are two 
possibilities here:

a) Something got fixed on the trunk between April 22 and April 28.
b) The reflow patch makes some things slower.

Since it's known for a fact that in its current state the reflow patch 
_does_ make some things slower, I find option (b) to be the more likely 
one...

In any case, the only way to differentiate options (a) and (b) is to 
test an April 22 build _without_ the reflow patch.  I'd be quite 
interested in the results of such testing, myself....


Reply via email to