Hi, I was working on bug 3885 tonight, and am stuck now.. Could someone
take a look and test a couple of things?
I'm running 2.6 series kernel, and it keeps dieing into the debugger
before the main window opens.
I think that it will sart with builtin msvcrt with 2.4 series kernel,
but not 2.6, but since I'm in the middle of another project, I cant
reboot to find out. It will at least start under both series for sure
with native msvcrt.
Louis Lenders removed a trace line from wine and recompiled and it
started, even with builtin, which is kinda weird, since he didnt run it
with any traces. So it may be a combination of a bug in that trace, and
a bug in the 2.6 kernel..
Once running the app, with either builtin or native msvcrt, it goes thru
the first benchmark and then spits out an error dialog box when it goes
to start the next one, regardless of which benchmarks the app is told to
test.. IOW, you can tell it to run #1 and #2, and it will spit out the
error before starting #2, and take you back to the main app window. Or
you can tell it to run #1 and #3, and same happens, or you can tell it
to run #2 and #3, and same happens, so its not something to do with the
way wine is handling the benchmark itself.
Louis also hard coded D3D_OK into wine in the right spot and managed to
get the program to go thru all benchmarks from 1-16, so wine is
mis-reporting something there.
If someone could take a look and propose a fix, it would be appreciated,
seems to me that the 2 fixes should be fairly trivial to come up with
since those 1 line hacks worked to get around the aforementioned problems.
If we can get it to start on 2.4 and 2.6 with builtin and get it to go
all the way to mark 16 before it spits out an error, then that will be
another bug closed, and I will file a separate one for mark 16..
Tom
begin:vcard
fn:Tom
n:Spear;Tom
note:a.k.a. Dustin Navea
x-mozilla-html:TRUE
version:2.1
end:vcard
