December 2001 - Wine-Devel - list.winehq.org

RE: Installshield 6 (inter-proc) patches
by Patrik Stridvall Dec. 13, 2001

Dec. 13, 2001

> It's been quite fascinating watching this discussion; it's > nice to see that > such issues are covered. It raises two issues for me: > 1. Where does Lindows fit into all this? Does anyone know > what's happening > there? No idea. > 2. Is the current wine license as loose as it appears to me? Probably. > My > understanding is that "the license" is the file "LICENSE" at > the root of the > cvs tree and basically allows anyone to do anything with it > provided they > keep the license in there (and respect copyright). Yes. > I keep > seeing quotes and > things that suggest wine is under GPL, LGPL, BSD etc.; I > presume that they > are simple ill-informed. It is under the modified BSD license, ie the BSD license without the advertisment clause, more often called the X11 license.

1 0

Re: Patch to memory/virtual.c
by Uwe Bonnes Dec. 13, 2001

Dec. 13, 2001

>>>>> "Robert" == Robert Baruch <autophile(a)starband.net> writes: Robert> Description: map_image now sets up the protections in the view Robert> of the memory-mapped executable image according to the Robert> protections for each section in the image. Before the patch, Robert> running an app under WinDbg on W2K/VMware showed a protection Robert> violation at a certain address, which Wine did not fault at that Robert> address. After the patch, the app faults in the same place in Robert> Wine as it does in W2K. I think the protection setting needs more care. With some shrinker file (targ2001.exe) I get ... trace:module:map_image mapping section .idata at 0x75b69000 off 34000 size c00 flags 40000040 ... trace:virtual:map_image Set protection for section .idata 40000040 to 41. trace:virtual:VIRTUAL_SetProt 0x75b69000-0x75b69fff c-r-- ... Unhandled exception: page fault on write access to 0x75b6903c in 32-bit code (0x40091c2f). In 32-bit mode. 0x40091c2f (PE_fixup_imports+0x2af [pe_image.c:330] in libntdll.so): movl %eax,0x0(%ecx) 333 if (!thunk_list->u1.Function) { Wine-dbg>Backtrace: =>0 0x40091c2f (PE_fixup_imports+0x2af(wm=0x40382b48) [pe_image.c:330] in libntdll.so) (ebp=405b5dc8) Giving VPROT_WRITE to .idata breaks when native oleaut32 is loaded: trace:module:MODULE_LoadLibraryExA Trying native dll 'D:\wine95\system\OLEAUT32.DLL' trace:module:PE_LoadImage loading D:\wine95\system\OLEAUT32.DLL trace:virtual:CreateFileMappingW (18,(nil),01000000,0000000000000000,(null)) trace:module:map_image mapped PE file at 0x65340000-0x653d2000 trace:module:map_image mapping section .text at 0x65341000 off 2000 size 82000 flags 60000020 ... trace:virtual:map_image Set protection for section .text 60000020 to 45. trace:virtual:VIRTUAL_SetProt 0x65341000-0x653c2fff c-r-x ... Unhandled exception: page fault on write access to 0x65341000 in 32-bit code (0x40091c2f). In 32-bit mode. 0x40091c2f (KERNEL32.DLL.__wine_call_from_16_thunk+0x61cf): movl %eax,0x0(%ecx) Wine-dbg>Backtrace: =>0 0x40091c2f (KERNEL32.DLL.__wine_call_from_16_thunk+0x61cf) (ebp=405b6cbc) 1 0x400924e2 (KERNEL32.DLL.__wine_call_from_16_thunk+0x6a82) (ebp=405b6d48) 2 0x40092632 (KERNEL32.DLL.__wine_call_from_16_thunk+0x6bd2) (ebp=405b6d70) ... Running with builtin oleaut32 and friends, I again get to the shrinker error box. Thanks for your efforts so long! Bye -- Uwe Bonnes bon(a)elektron.ikp.physik.tu-darmstadt.de Institut fuer Kernphysik Schlossgartenstrasse 9 64289 Darmstadt --------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

1 0

RE: Installshield 6 (inter-proc) patches
by Patrik Stridvall Dec. 13, 2001

Dec. 13, 2001

> Patrik Stridvall <ps(a)leissner.se> writes: > > > In short: > > Should the Wine project wait until you release or should it not? > > That's certainly a question we have to think about, but I think there > is a deeper issue: should we continue to release under a license that > allows people to use our own code to hurt the project? Umm. I feared that question would come. The "protection" the LGPL (or GPL) that Marcus proposed is IMHO largely an illusion when it comes to libraries. Sure we might use a strict interpretion as a weapon in a PR campaign against possible voilators but we don't have the resources to sue somebody and I very much doubt we would succed either. If I was I judge I wouldn't want to touch the issue of what constituted "linking" with a ten foot pole. Beside since where was no real damages since Wine was already free I would probably look for an excuse to dismiss the case as early as possible. No, a license change is not the answer to our problems. > My concern is not so much about Transgaming, I trust that Gav means to > do the right thing, Agreed and I think we should allow them considerable time to consider their position as well no need to rush something. > even if I don't entirely agree with his methods. Well, money makes the world go round whether you like it or not. :-) > But I'm worried that if Transgaming succeeds, it will set a precedent > that others will follow, who may have no desire at all to do the right > thing for Wine. Succeed with what? Now that Transgaming has done the hard work of getting InstallShield to work and even been kind enough provide the source code eventhough under a propritary license it can't be that difficult to look at it and provide an alternative implementation. Note that Gav said: < To say that our work is preventing anyone from doing a free version of DCOM is < completely incorrect. To the contrary: we have already contributed < substantial amounts of this code to the public tree - thousands of < lines, in fact. That can only help anyone with a sudden and unexpected < interest in recreating our work. We could quite easily make InstallShield work ourselves now that they have don't all the tedious debugging and we know what needs to be done. Note copyright only protects expression not fact or ideas (or hard work for that matter). In light of this it might be advantagous for Transgaming to release their InstallShield code. But as I said lets no rush things. I'm prepared to give them the benifit of the doubt and wait for them to decide. However my (and presumable others) patience in not boundless. > What will happen if 5 different dlls are improved and > released by 5 different companies under 5 different non-free licenses? When we had to decide which of the companies we trusted to release their code in the not so far away future and which we don't trust. The work of companies that we don't trust is ignored and we work on as we always have.

2 1

Improved redraw performance when application loaded
by Dave Hawkes Dec. 13, 2001

Dec. 13, 2001

changelog: Reduced the time for a window with many child windows to be displayed. More info: A large application was taking up to 10 seconds to display its main window. This appears to be due to X individually exposing every child window (of which there are very many) of the main application window and then wine forcing a redraw on the parent of every child and then all the subchildren again (!) resulting in the accumulation of a huge redraw list. This patch will force a redraw only on the child and not the parent as X will eventually expose that anyway. After applying this patch the initial redraw dropped from around 10 seconds to about 2 seconds and I could not find any painting problems If anyone else is having slow application loading problems of this nature, then please try this patch. Dave Hawkes

1 0

Re: Uncleared WM_PAINT loop problem
by Gerard Patel Dec. 13, 2001

Dec. 13, 2001

At 02:13 PM 12/12/2001 -0500, you wrote: >I have a problem where an edit control get stuck in an infinite WM_PAINT >loop. > >BeginPaint never clears the paint count because of this code: > > if(((hrgnUpdate = wndPtr->hrgnUpdate) != 0) || (wndPtr->flags & >WIN_INTERNAL_PAINT)) > add_paint_count( hwnd, -1 ); > > >if wndPtr->hrgnUpdate is 0 then the paint count is never decremented here >and nowhere else attempts to decrement it and we land up in an infinite >paint loop. I'm not sure how we get to this point the a NULL update region >(which may be a bug anyway), but there should still be a method of breaking >out of this loop once we are there. > >I tried removing the test and always calling add_paint_count and this >appears to fix the problem. > With current Cvs code the infinite loop should have gone away... Alexandre Julliard has changed a strangeness of BeginPaint : when the update region is null, Wine used to paint the whole window. New code set an empty painting rectangle. However, if the empty update region comes from an earlier Wine bug, the new code replaces excessive repainting with no repainting at all. I have already an app that shows this problem. As I was saying in reply to Andi Mohr, better code can reveal better old bugs... Gerard

2 1

Improved redraw performance when application loaded
by Dave Hawkes Dec. 13, 2001

Dec. 13, 2001

changelog: Reduced the time for a window with many child windows to be displayed. More info: A large application was taking up to 10 seconds to display its main window. This appears to be due to X individually exposing every child window (of which there are very many) of the main application window and then wine forcing a redraw on the parent of every child and then all the subchildren again (!) resulting in the accumulation of a huge redraw list. This patch will force a redraw only on the child and not the parent as X will eventually expose that anyway. After applying this patch the initial redraw dropped from around 10 seconds to about 2 seconds and I could not find any painting problems If anyone else is having slow application loading problems of this nature, then please try this patch. Dave Hawkes

1 0

Great ! [Re: wine/ dlls/user/Makefile.in dlls/user/wnd16.c ...]
by Andreas Mohr Dec. 13, 2001

Dec. 13, 2001

On Wed, Dec 12, 2001 at 08:03:31PM -0500, Alexandre Julliard wrote: > Started rewrite of painting functions for multiple processes. Moved > BeginPaint to dlls/user. Fixed bug in paint count handling. Oh yeah. *Much* better now :-) OrCAD PSpice can actually be looked at now without vom****** :-) -- Andreas Mohr Stauferstr. 6, D-71272 Renningen, Germany Tel. +49 7159 800604 http://home.nexgo.de/andi.mohr/

2 1

Re: Installshield 6 (inter-proc) patches
by Igor Furlan Dec. 13, 2001

Dec. 13, 2001

On Thu, 13 December 2001, Alexandre Julliard wrote: it will set a precedent > that others will follow, who may have no desire at all to do the right > thing for Wine. What will happen if 5 different dlls are improved and > released by 5 different companies under 5 different non-free licenses? > Wasn't (L)GPL designed to ___prevent___ this kind of things to happen ? just my 0.02euro Igor

1 0

Re: Great ! [Re: wine/ dlls/user/Makefile.in dlls/user/wnd16.c ...]
by Gerard Patel Dec. 13, 2001

Dec. 13, 2001

At 02:08 PM 13/12/2001 +0100, you wrote: >On Wed, Dec 12, 2001 at 08:03:31PM -0500, Alexandre Julliard wrote: >> Started rewrite of painting functions for multiple processes. Moved >> BeginPaint to dlls/user. Fixed bug in paint count handling. >Oh yeah. >*Much* better now :-) > >OrCAD PSpice can actually be looked at now without vom****** :-) Well, at least one of my test apps is now completely broken, certainly a following of this patch (repainting problems). This does not mean that the patch is wrong of course - maybe 2 bugs were compensating each other before. Gerard

1 0

Re: OSS fixes
by Duane Clark Dec. 12, 2001

Dec. 12, 2001

Eric wrote: > could the people who had issues with the latest OSS patch(es) test > if this one fixes all their issues ? This patch causes both Riven and Myst to immediately freeze for me, and no sound at all. Without the patch, the sound in Myst works pretty good (occasional brief dropouts). In Riven without the patch, only the opening sequence has sound, and that has many large dropouts. I'm using ALSA drivers in Redhat. Duane

6 7