- Wine-Devel - list.winehq.org

Re: urlmon: Prevent possible use of freed memory
by Jacek Caban May 22, 2005

May 22, 2005

Troy Rollo wrote: >This fixes Bugzilla bug 2969 <http://bugs.winehq.org/show_bug.cgi?id=2969> > >ChangeLog: > Prevent a URLMonikerImpl from being freed before its work in > BindToStorage is done. > > >------------------------------------------------------------------------ > >Index: dlls/urlmon/umon.c >=================================================================== >RCS file: /home/wine/wine/dlls/urlmon/umon.c,v >retrieving revision 1.52 >diff -u -r1.52 umon.c >--- dlls/urlmon/umon.c 5 May 2005 09:50:46 -0000 1.52 >+++ dlls/urlmon/umon.c 21 May 2005 23:57:36 -0000 >@@ -456,6 +456,13 @@ > *ppvObject = (void *) This->pstrCache; > IStream_AddRef((IStream *) This->pstrCache); > >+ /* We are about to start calling back into the application. It is >+ * possible that the application will release its reference to us in >+ * these callbacks, so we need to add a reference to This to make sure the >+ * URLMonikerImpl is not freed before we reach the end of this method. >+ */ > > The right solution is separating IBinding and IMoniker interfaces. They shouldn't be implemented in the same object. I'll send the patch. Jacek

2 1

Re: urlmon/umon.c: Fix corruption
by Jacek Caban May 22, 2005

May 22, 2005

Uwe Bonnes wrote: >Changelog: > dlls/urlmon/umon.c: URLMonikerImpl_Release() > Release structure elements before Freeing structure > > My patch I sent today fixes this too. > Changelog: > wine/dlls/urlmon/umon.c:URLMonikerImpl_BindToStorage() > Don't use a HeapAllocated pointer for other purposes ... > + if (path) > + HeapFree(GetProcessHeap(), 0, path); You don't need this check. > Changelog: > dlls/wininet/internet.c: CrackURL, dlls/wininet/tests/http.c > Add some escape/unescape cases and code thor that cases ... > + ok(InternetCrackUrlA(TEST_URL3, 0, ICU_DECODE, &urlComponents),"InternetCrackUrl failed with GLE 0x%lx\n",GetLastError()); > + fprintf(stderr,"BON:%*s\n", (int)urlComponents.dwUrlPathLength, urlComponents.lpszUrlPath); > > } You shouldn't use fprintf() here. Use trace() instead. Jacek

1 0

Re: [OLE #90] Marshal Objects & Monikers into the ROT
by Alexandre Julliard May 22, 2005

May 22, 2005

Robert Shearman <rob(a)codeweavers.com> writes: > Changelog: > - Marshal objects & monikers into the ROT. > - Test for this behaviour. This fails make test here: ../../../tools/runtest -q -P wine -M ole32.dll -T ../../.. -p ole32_test.exe.so moniker.c && touch moniker.ok fixme:ole:MkParseDisplayName (0x77c518a0, L"clsid:20D04FE0-3AEA-1069-A2D8-08002B30309D:", 0x77acfe68, (nil)): stub. err:ole:CoUnmarshalInterface IMarshal::UnmarshalInterface failed, 0x80004005 moniker.c:145: Test failed: Number of monikers should be equal to 2 not 0 err:ole:CoUnmarshalInterface IMarshal::UnmarshalInterface failed, 0x80004005 moniker.c:164: Test failed: Number of monikers should be equal to 2 not 0 fixme:ole:CreateClassMoniker {0002e005-0000-0000-c000-000000000046} err:ole:get_unmarshaler_from_stream Failed to read common OBJREF header, 0x00000001 make[3]: *** [moniker.ok] Error 2 -- Alexandre Julliard julliard(a)winehq.org

2 1

WineHQ css; binfmt && Mono && wine
by Brian Vincent May 22, 2005

May 22, 2005

Something broke on WineHQ. For some reason <br /> tags aren't getting parsed right - see this week's WWN for an example. It's displaying "<BR>". I haven't looked into it at all, but I figured someone (Dimi?) might remember messing with this stuff recently. Also, anyone happen to know if Wine can be used in conjunction with binfmt alongside Mono? Are there any unique C# identifiers in the first 128 bits of a .Net executable that could identify it? It appears the Mono guys recommend looking for the magic MZ bytes, which is exactly what Wine requires also. We can't look for "PE" since binfmt can only look at the first 127 bytes of a file. -Brian

8 10

console behaviour of tty vs. x11 driver
by Kuba Ober May 22, 2005

May 22, 2005

Hi, I'm running vc.net as a cross-compiler under wine. Everything looks very nice when you run it with X server running. Yet, when there's no X present, winetty driver kicks in and corrupts the console output. Is there a simple remedy? Here are the problems that I observe: 1. Pointless Xlib error message about it not being able to connect to the X server -- this wastes 2 lines of screen real estate and the apps in question are console-only *anyway* It'd be nice to have an environment variable or a registry setting to disable showing of this message. 2. Winetty seems to use ncurses for simple line-oriented output where winex11 did the right thing and didn't wrap the output with escape sequences in such a weird way 3. Winetty clears the screen upon startup (due to ncurses maybe), so e.g. running things from nmake.exe makes the screen get wiped everytime a compile/link step is invoked. This has a nasty side effect: with winetty driver a simple wine foo.exe > log gives you corrupt output with escape sequences that weren't reallyoutput by foo.exe , whereas with winex11 wine foo.exe > log gives you the *correct* output capture. I will file a bug for it if it's not just something weird on my side. Kuba

2 1

WineConf todo list
by Francois Gouget May 21, 2005

May 21, 2005

I'm trying to gather a list of stuff we said needed to be done at WineConf. Of course the initial list is going to be horribly incomplete but I'm sure others will contribute the missing pieces. * Wine Wiki This seems to be done already. It's available there: http://wiki.winehq.org/ * Add a list of companies providing Wine services and support This discussion/flameware is in full swing. * Putting the videos online Done. Maybe some retouching / cleanup to do if we find someone knowledgeable and willing. * Moving the Wine doc to SourceForge The documentation, i.e. the SGML files in documentation/, uses completely different tools from the rest of Wine. Also Alexandre does not want to deal with its translations. So move the SGML documentation to its own CVS repository, winedoc, which we could host on SourceForge. The English version would go in the en subdirectory so we can easily add translations. * Check in the French translation and corresponding infrastructure. The translation will go in the fr subdirectory of the winedoc CVS repository. At the same time the infrastructure for adding other translations should be put in place. * Make an easy to use cxtest script. The idea is to have a script which could be put in a crontab so it can be run entirely automatically and which would: - checkout / update and recompile CXTest - checkout / update and recompile the latest Wine - run regression tests using Wine's Winelib applications - maybe download some free applications and run tests to make sure they install and start up Jozef Stefanka is planning to work on this. * Get rid of the configuration file Well that's the 0.9 todo list. Maybe it's best to keep it separate from this list. * Find a way to document the registry I had a discussion with Alexandre and lamented that, now that we're going to move things out of the configuration, we won't be able to put comments describing what each setting does. However we already have the problem with all of the Windows settings. Furthermore he noted that it's hard to find where a given setting is used in the Wine code (because the registry path is encoded as a Unicode string?). And the problem is compounded by the lack of proper documentation for registry settings in MSDN. So the idea is to create a way to document each registry setting in the source code so it can be extracted with a tool similar to c2man.pl (or with a hacked c2man.pl). -- Francois Gouget fgouget(a)free.fr http://fgouget.free.fr/ Avoid the Gates of Hell - use Linux.

4 3

Re: dwarf2 support progress step 4
by Mike Hearn May 21, 2005

May 21, 2005

On Fri, 20 May 2005 23:16:12 +0200, Raphael wrote: > We only have to build a correct symt model to get it working perfectly What exactly will this work allow - will winedbg get better as a result? At the moment backtraces already give function information... thanks -mike

4 3

Serialization of DEBUGMSG needed
by Uwe Bonnes May 21, 2005

May 21, 2005

Hallo, running CVS wine on my dual pentium machine with debugmessages enabled, I now see that both CPUs spit out debug output at the same time, leading to garbled output, like: 000b:Call gdi32.GetCurrentObject(0000136c,00000005) ret=7f0013:Ca000b:Ret gdi32.GetCurrentObject() retval=00000044 ret=7fef3e40 0013:Ret ntdll.NtUnmapViewOfSection() retval=00000000 ret=77b000b:R0013:Ret kernel32.UnmapViewOfFile() retval=00000001 re000b:Call g0013 :Ret wininet.CommitUrlCacheEntryW() re000b:Ret gdi32.GDI_ReleaseObj() retval=00000000 ret=7fef3f1d Is this an artefact of my machine or a wine problem? In the latter case, as Hyperthreading amd Multicore machines will get more available with time, a fix would be appreciated. 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

[RFC] upmfs - page-unaligned mmap() for Linux
by Dan Aloni May 21, 2005

May 21, 2005

Hello All, I'd like to bring this to your attention and review, hopefully to spawn some constructive discussions. The code was a result of experiments I've been doing lately concering some new idea not especially related to WINE. To quote myself from http://wiki.winehq.org/UnalignedMmap: ---- Large majority of PE executables have sections that are page-aligned in memory but not in the PE file image. Linux's mmap() doesn't support mappings that are not aligned to page boundries with regard to file offsets. Currently it forces wineserver to fake mmap() by allocating anonymous pages and using read() to load the sections. In dlls/ntdll/virtual.c, comment above map_file_into_view() says: "Linux kernels before 2.4.x can support non page-aligned offsets, as long as the offset is aligned to the filesystem block size. This is a big performance gain so we want to take advantage of it." These days are long gone, we are now using Linux 2.6, and it doesn't support unaligned mmap(). Enter upmfs. This simple kernel module exports this function to userspace using '/proc': * int mmap_offset(int fd, unsigned long offset); Given an fd and an offset, a new fd is returned. Using mmap() on that new fd, the mapping will be aligned with the given offset, which can be smaller than PAGE_SIZE. A wineserver implementation can check the existance of that module in the kernel (i.e "/proc/upmfs/interface" exists) and use its functionality. ---- The code for the upmfs source package is attached to this post. -- Dan Aloni da-x(a)colinux.org

1 0

environment-priority
by Detlef Riekenberg May 21, 2005

May 21, 2005

Hi While working with wine, I found the way how wine handles the environment. example for "TMP": 1. HCU\Environment\TMP 2. HKLM\System\CurrentControlSet\Control\Session Manager\Environment\TMP 3. $WINETMP 4. $WINEPREFIX/config (TEMP) 5. Buildin Default (c:\windows\temp) During "wineprefixcreate", "TMP" and "TEMP" are created for entry 2 (HKLM\..) from entry 4 ($WINEPREFIX/config : TEMP) with a fallback to entry 5 (c:\windows\temp), but entry 3 ($WINETMP) is ignored. When I delete the entry 2 (HKLM\..), then entry 3 is used ($WINETMP). IMHO, "$WINETMP" should overwrite all other locations. PS: Tested only "WINETMP", but there are also $WINETEMP, WINEHOME and WINEPATH. -- By By ... ... Detlef

1 0