March 2005 - Wine-Devel - list.winehq.org

Re: OLEAUT32: fix detection of GIF palette index >= 128 for transparency

by Andreas Mohr

Hi, On Tue, Mar 29, 2005 at 12:43:38PM -0500, Alex Villacis Lasso wrote: > I traced this problem to a one-line bug in the GIF loading code. The > code is supposed to initialize a transparency flag to -1 when no > transparency is used, or else the palette index of the transparency > color. The sample image I was using indicates that the transparency > index is 255. Due to an unintended sign extension, the transparency > index (held in a Byte array of basic type "signed char") is > sign-extended to the variable "transparency" of basic type "signed int". > This causes the flag to be assigned the value -1, which happens to be > the same value used as a no-transparency indicator. So transparency was ....^^^^^^^^^^^^ Heh, 3 times "the transparency" in exactly the same line position, 4 times "transparency" in exactly the same position - what did you smoke?? ;-) (or worse, what did *I* smoke to recognize this?? :-) OK, I promise I'll stop that useless fun email here. Anyway, thanks for all those nice patches, Andreas Mohr

March 29, 2005

Re: MSVCRT: implement _chsize (RESEND)

by Alexandre Julliard

Hans Leidekker <hans(a)it.vu.nl> writes: > Anything wrong with this one? > > This should fix apps built with MinGW (bug report 2826). You should at least restore the file position before returning, though it would probably be better to not change the file position at all, this is not thread safe. -- Alexandre Julliard julliard(a)winehq.org

March 29, 2005

Re: Resent^2 : Zero out return string in GetMenuStr

by Alexandre Julliard

Uwe Bonnes <bon(a)elektron.ikp.physik.tu-darmstadt.de> writes: > Anything wrong with this patch? > > Changelog: > dlls/user/menu.c: GetMenuStrA|W > Initialize return string as an empty string You need to take the specified length into account before writing to the string. -- Alexandre Julliard julliard(a)winehq.org

March 29, 2005

Re: KERNEL: improve stubs for DnsHostnameToComputerName{A,W} (RESEND)

by Alexandre Julliard

"Dmitry Timoshkov" <dmitry(a)baikal.ru> writes: > "Hans Leidekker" <hans(a)it.vu.nl> wrote: > > > +BOOL WINAPI DnsHostnameToComputerNameA(LPCSTR hostname, > > + LPSTR computername, LPDWORD size) > > > + for (i = 0; i <= len; i++) > > + computername[i] = toupper(hostname[i]); > > You can't use toupper() from glibc on a name in a Windows encoding. > Why not just call a W version to do the job? Actually, why do you want to convert to uppercase at all? Is there really an app that breaks if computer name contains lowercase chars? -- Alexandre Julliard julliard(a)winehq.org

March 29, 2005

Re: BadDrawable and install shield.

by Oliver Stieber

On Tuesday 29 March 2005 16:04, Marcus Meissner wrote: > On Tue, Mar 29, 2005 at 03:19:42PM +0100, Oliver Stieber wrote: > > Hi, > > I've been having a lot of problems with install shield and BadDrawables. > > > > 1: Does anyone know what's causing the problem? > > 2: Are there any plans to implement XSetErrorHandler and manager the errors better, even if it's an int3 that winedbg can catch. > > Are you using desktop mode? > > InstallShield works fine in Managed mode... > Turning of desktop mode fixed the problem. > Oh, and there is a X Errorhandler, just enabled Synchronous mode to > get a backtrace. > > Ciao, Marcus > -- For DirectX9 on Linux www.oliverthered.f2s.com/projects/wine ICQ 169222899 MSN oliverthered(a)hotmail.com

March 29, 2005

discusion about a message loop in listview.c

by Dietrich Teickner

I have some weeks before reported, FlashFXP v3.02 loops with a message-loop in listview.c after login. It does this in Odin and in Wine. In Odin it stops in smaller time (deep 1000). Wine has a bigger stack, and so needs wine more time for this, bat I can see in the loog, the args for the function are more and more locatet at lower stack adresses. I have written this workarout for testing this problem in wine and in odin. with this FlashFXH 3.02 will display the contence of the server dir very fast at both environment. In the log I can see, that the message-loop was reported and stopped. I thing, we need a flag inside any item to report every function, that can called recursive, this item is in work (function-separart, or global ?), that can stop recursions of this type in the future or many programms. I have also found this try does also helps Agent 2.0 in the groups->defailt propertys (only odin-tested). This does also every crashs in earlyer builds with stack overflow in odin. Dietrich Index: wine/dlls/comctl32/listview.c =================================================================== RCS file: /home/wine/wine/dlls/comctl32/listview.c,v retrieving revision 1.406 diff -u -r1.406 listview.c --- wine/dlls/comctl32/listview.c 25 Mar 2005 20:49:00 -0000 1.406 +++ wine/dlls/comctl32/listview.c 29 Mar 2005 16:31:48 -0000 @@ -139,6 +139,7 @@ * WM_CREATE does not issue WM_QUERYUISTATE and associated registry * processing for "USEDOUBLECLICKTIME". */ +#define __T__ #include "config.h" #include "wine/port.h" @@ -7050,10 +7051,23 @@ * SUCCESS : TRUE * FAILURE : FALSE */ +#ifdef __T__ +static int zIndex = -2; +typedef struct { + PVOID lTol; + HWND cHwnd; + HWND cHwndParent; + INT cItem; + } SetItemStateLink, *PSetItemStateLink; +#endif static BOOL LISTVIEW_SetItemState(LISTVIEW_INFO *infoPtr, INT nItem, const LVITEMW *lpLVItem) { BOOL bResult = TRUE; LVITEMW lvItem; +#ifdef __T__ + int saveLastError = GetLastError(); + SetItemStateLink cTol = { NULL, infoPtr->hwndSelf, infoPtr->hwndNotify, nItem}; +#endif lvItem.iItem = nItem; lvItem.iSubItem = 0; @@ -7062,6 +7076,34 @@ lvItem.stateMask = lpLVItem->stateMask; TRACE("lvItem=%s\n", debuglvitem_t(&lvItem, TRUE)); +#ifdef __T__ + if (-2 == zIndex) + { + zIndex = TlsAlloc(); + SetLastError(saveLastError); + } /* endif */ + if (zIndex >= 0) + { + cTol.lTol = TlsGetValue(zIndex); + if (cTol.lTol < (PVOID)&cTol) + { + cTol.lTol = NULL; + } /* endif */ + SetLastError(saveLastError); + if (cTol.lTol + && (((PSetItemStateLink)cTol.lTol)->cHwnd == infoPtr->hwndSelf) + && (((PSetItemStateLink)cTol.lTol)->cHwndParent == infoPtr->hwndNotify) + && (((PSetItemStateLink)cTol.lTol)->cItem == nItem)) + { + TRACE_(listview)("(DT) Hwnd:%x,%x Parent:%x,%x Item:%d,%d, stop the recursion\n", infoPtr->hwndSelf, ((PSetItemStateLink)cTol.lTol)->cHwnd, + infoPtr->hwndNotify, ((PSetItemStateLink)cTol.lTol)->cHwndParent, + nItem, ((PSetItemStateLink)cTol.lTol)->cItem); + return TRUE; + } + TlsSetValue(zIndex,(LPVOID)&cTol); + SetLastError(saveLastError); + } +#endif if (nItem == -1) { /* apply to all items */ @@ -7070,6 +7112,14 @@ } else bResult = LISTVIEW_SetItemT(infoPtr, &lvItem, TRUE); +#ifdef __T__ + if (zIndex >= 0) + { + saveLastError = GetLastError(); + TlsSetValue(zIndex,(LPVOID)cTol.lTol); + SetLastError(saveLastError); + } +#endif /* *update selection mark

March 29, 2005

Re: BadDrawable and install shield.

by Marcus Meissner

On Tue, Mar 29, 2005 at 03:19:42PM +0100, Oliver Stieber wrote: > Hi, > I've been having a lot of problems with install shield and BadDrawables. > > 1: Does anyone know what's causing the problem? > 2: Are there any plans to implement XSetErrorHandler and manager the errors better, even if it's an int3 that winedbg can catch. Are you using desktop mode? InstallShield works fine in Managed mode... Oh, and there is a X Errorhandler, just enabled Synchronous mode to get a backtrace. Ciao, Marcus

March 29, 2005

Re: [PATCH] MapiSendMail() implementation

by Kuba Ober

On poniedziałek 07 marzec 2005 02:40 pm, Mike Hearn wrote: > On Mon, 2005-03-07 at 20:21 +0100, Michael Stefaniuc wrote: > > Gnome has the notion of "default email client", i suspect KDE has it too. > > Unless you use Bonobo+Evolution (and maybe KMail+DCOP) though you can't > match the abilities of MAPI, which lets you create arbitrary emails. Using dcop with kmail gives you more power, but for basic functionality command line suffices: echo "Hello!" > contents_file kmail \ --subject "Test email" \ --header "In-Reply-To: <1110224432.3769.40.camel(a)littlegreen>" \ --msg contents_file \ --composer Cheers, Kuba

March 29, 2005

Re: KERNEL: improve stubs for DnsHostnameToComputerName{A,W} (RESEND)

by Dmitry Timoshkov

"Hans Leidekker" <hans(a)it.vu.nl> wrote: > +BOOL WINAPI DnsHostnameToComputerNameA(LPCSTR hostname, > + LPSTR computername, LPDWORD size) > + for (i = 0; i <= len; i++) > + computername[i] = toupper(hostname[i]); You can't use toupper() from glibc on a name in a Windows encoding. Why not just call a W version to do the job? -- Dmitry.

March 29, 2005

BadDrawable and install shield.

by Oliver Stieber

Hi, I've been having a lot of problems with install shield and BadDrawables. 1: Does anyone know what's causing the problem? 2: Are there any plans to implement XSetErrorHandler and manager the errors better, even if it's an int3 that winedbg can catch. e.g. int myXErrorHandler(Display *display, XErrorEvent *xError){ switch(xError->type){ case BadDrawable: break; default: int3(); return XErrorHandler(NULL); } Xerror = xError->error_code; return 0; /* return value ignored */ } oid setErrorHandler(void){ XSetErrorHandler(myXErrorHandler); } void dosomething(void){ Xerror = 0; SomeXCall(); if(Xerror){ /* call failed, take appropriate action */ } } Some examples are.... http://www.download.com/Aliens-vs-Predator-2-single-player-demo/3000-7563_4… fixme:sync:SetNamedPipeHandleState 0xa8 0xb5bd50d4 (nil) (nil) wine: Unhandled exception (thread 002f), starting debugger... WineDbg starting on pid 0x11 X Error of failed request: BadDrawable (invalid Pixmap or Window parameter) Major opcode of failed request: 67 (X_PolyRectangle) Resource id in failed request: 0x6200007 Serial number of failed request: 989 Current serial number in output stream: 1049 http://games.softpedia.com/get/Games-Demo/Crazy-Taxi-Demo.shtml X Error of failed request: BadDrawable (invalid Pixmap or Window parameter) Major opcode of failed request: 62 (X_CopyArea) Resource id in failed request: 0x6000007 Serial number of failed request: 3252 Current serial number in output stream: 3258 http://www.download.com/1503-A-D-The-New-World-demo/3000-7485_4-10228730.ht… fixme:x11drv:X11DRV_SetWindowRgn not supported on other thread window 0x3002e X Error of failed request: BadDrawable (invalid Pixmap or Window parameter) Major opcode of failed request: 62 (X_CopyArea) Resource id in failed request: 0x5a00007 Serial number of failed request: 3152 Current serial number in output stream: 3158 http://www.download.com/Colin-McRae-Rally-2005-demo/3000-7531_4-10312485.ht… crashes at the end with fixme:propsheet:PROPSHEET_UnImplementedFlags PSH_RTLREADING X Error of failed request: BadDrawable (invalid Pixmap or Window parameter) Major opcode of failed request: 62 (X_CopyArea) Resource id in failed request: 0x6600007 Serial number of failed request: 28770 Current serial number in output stream: 28799 -- ICQ 169222899 MSN oliverthered(a)hotmail.com

March 29, 2005