Posted in bugzilla, I'll add the executable later on, and the patch... But it's not a solution, it's only make app believing the page has a different protection flag, doesn't correct the protection itself.... Should cause errors elsewhere.
Date sent: Wed, 17 Jul 2002 19:45:36 -0600 From: Tony Lambregts <[EMAIL PROTECTED]> To: [EMAIL PROTECTED] Copies to: [EMAIL PROTECTED] Subject: Re: Virtual Pages difference - Spotted with VirtualQuery() > Max wrote: > > >As usual, AutoCAD seems to be a good test-bench for wine :-) > >I spotted the problem that seemed coming from VirtualQuety function, but > >that is (I guess... hmmmm) more a problem of memory handling in wine. > > > >I made a little test app; here the results in wine and win2k, later on > >the app itself : > > > >WINE RESULTS : > > > >Memory info about '004012F0' : > > Base Address : 00401000 > > Allocation Base : 00400000 > > Allocation Protect : PAGE_READONLY | PAGE_WRITECOPY | > > PAGE_EXECUTE_READWRITE | PAGE_NOACCESS > > Region Size : 00017000 > > State : MEM_COMMIT > > Protect : PAGE_WRITECOPY <----- HERE > > Type : MEM_PRIVATE > > > >WIN2k RESULTS : > > > >Memory info about '004012F0' : > > Base Address : 00401000 > > Allocation Base : 00400000 > > Allocation Protect : PAGE_EXECUTE_WRITECOPY > > Region Size : 00001000 > > State : MEM_COMMIT > > Protect : PAGE_EXECUTE_READ <---- AND HERE > > Type : MEM_IMAGE > > > >Well, dunno too much about Type field, I didn't find an app that check it yet; > >I did find the app that uses Protect flag; I think it's poor coded, but as > >usual we should make bad apps work too :-) > >The app put some data in code page, not yet spotted if hard or soft-coded, > >then before reading it checks Protect field to see if it has not set the > >PAGE_GUARD, NO_ACCESS and WRITECOPY flags; if t has, it hangs some 10.000 > >lines later (sigh....) > >I think wine should set Protect flag quite as like as possible as windoze > > > >For AutoCAD2000 I put a dirty hack in process.c, but is really ugly; i'd > >prefer not to put my hands (bytes ?) too deeply inside some code that I don't > >understand completely... > > > >Attached is the test source (compiled in Borland CBuilder, but should work > >with any compiler, apart some unuseful pragma's). > >If needed I can post the compiled too :-) > > > >Regards > > > >Max > > > > > This is an excellent bug report but even if someone starts working on it > right away it would be nice to have this stuff in bugzilla. That > includes the exe and your (maybe not-so) ugly hack. > > > Tony Lambregts > >
