More dogfood tests

[Segin]

I have also started using Firefox in Wine for everything, namely because 
native lost a lib, old mozilla and opera and native firefox cause X to 
malloc() like mad (pkill -9 the broswer to cause X to free() ).
I have found very few problems, mainly just images being color-reversed, 
and only on PNG of certain types.

Also, Wine is MUCH faster if you have NPTL as your only threading method 
in glibc. (think a 15x improvement in speed here...)

Wine also seems to know about the XDND X extension, cause it knows what 
GTK+2 widgets can be dragged and dropped onto, but otherwise the actual 
drag-and-drop procedure is a null effect.

Ahh to hell with it, let's get to the speed tests.
time wine firefox.exe, cold cache:
real    0m48.895s (about 45s, really)
user    0m4.116s
sys     0m1.589s
hot cache:
real    0m9.287s (about 7.5s, really)
user    0m3.966s
sys     0m1.026s
(Please note that the "real" time is off by about 1-2 seconds, to allow 
for me to quit Firefox.)
Here we do the 7-Zip File Manager. Errors:
err:listview:LISTVIEW_WindowProc unknown msg 2005 wp=00000001 lp=00000000
err:rebar:REBAR_AdjustBands Phase 1 failed, x=1000, maxx=-4, start=1, end=1
time wine 7zFM.exe, cold cache:
real    0m34.815s (about 30s)
user    0m1.044s
sys     0m0.672s
and with a hot cache:
real    0m5.105s (about 4s, actually)
user    0m0.997s
sys     0m0.378s

And now we do PuTTY:
(no errors)
Cold cache:
real    0m39.382s (37s)
user    0m1.023s
sys     0m0.868s
Hot cache:
real    0m3.170s (no ajustment)
user    0m0.978s
sys     0m0.395s

On a normal Windows system, the cache is always hot (all the core DLLs 
are always loaded), so the hot cache times actually rival a real Windows 
system, especially the one for PuTTY (which beats every test on a real 
Windows box I have done except those on more powerful hardware)