http://bugs.winehq.org/show_bug.cgi?id=13555
--- Comment #9 from foobard jens@porup.com 2008-05-31 12:55:40 ---
"make" is only compiling wine. You still have to install wine with sudo make install.
hmmm... well, "make depend && make" was enough to generate a ./wine executable in ~/wine-git, so I don't know. Either way the regression docco should be updated.
So... I ran the regression test using git bisect good wine-0.9.53 (the bug was fixed in 0.9.56) and git bisect bad wine-1.0-rc3. I ran the test with the installer of both the full version of Acrobat Pro 6 and the trial version of Acrobat Pro 7.
I was unable to install either version at any point during the regression testing. Every test at every version resulted in the error:
wine: Unhandled page fault on read access to 0x00007ffc at address 0xb7ce648d (thread 0032), starting debugger...
with the hex addresses varying at different times.
What I can speculate:
The debugger never used to start when this install failed. Perhaps fixes to the debugger have caused this regression?
Also, a good place for a developer to start might be the original patch that made it work in the first place (or rather subsequent changes to the same module):
http://winehq.org/pipermail/wine-cvs/2008-February/040180.html
which is mentioned in bug 11514, the original bug report for this issue.