September 2007 - Wine-Devel

Re: Always set DIOBJECTDATAFORMAT.pguid to GUID_Button for joystick buttons

by Vitaliy Margolen

Antonio Cuni wrote: > At least one game (Pro Evolution Soccer 6) rely on the pguid field of > the DIOBJECTDATAFORMAT struct to respond to joystick button's events. > Setting pguid to GUID_Button solves the problem. > > 2007-09-20 Antonio Cuni <anto.cuni(a)gmail.com> > > * dlls/dinput/joystick_linux.c, > dlls/dinput/joystick_linuxinput.c > Always set DIOBJECTDATAFORMAT.pguid to GUID_Button for > joystick buttons > Thank you for the patch. I have few comments about it: > -- /tmp/wine-0.9.45/dlls/dinput/joystick_linux.c 2007-09-14 19:30:55.000000000 +0200 > +++ ./joystick_linux.c 2007-09-20 17:02:15.000000000 +0200 You should diff from the top of the tree (above line should say dlls/dinput/joystick_linux.c). > + df->rgodf[idx].pguid = &GUID_Button; // some games require the GUID to be GUID_Button C++ comments are not allowed in Wine. Please put comments above the code. > newDevice->buttons[i] = 0x80 | newDevice->numButtons; > + df->rgodf[idx].pguid = &GUID_Button; // some games require the GUID to be GUID_Button > df->rgodf[idx++].dwType = DIDFT_MAKEINSTANCE(newDevice->numButtons++) | DIDFT_PSHBUTTON; Please don't use tabs. Use spaces to keep formating. Vitaliy.

Sept. 20, 2007

Re: Implemented PsCreateSystemThread in ntoskrnl.exe

by Juan Lang

> Actually, NtCreateThread is too low-level and so isn't implemented in > Wine. RtlCreateUserThread might be better to use though. Whoops :) Thanks, Rob. --Juan

Sept. 20, 2007

Re: Implemented PsCreateSystemThread in ntoskrnl.exe

by Robert Shearman

Juan Lang wrote: > Hi Carroll, > > + if (ProcessHandle!=0) > + localThreadHandle = > CreateRemoteThread(ProcessHandle,0,0,StartRoutine,StartContext,0,&ThreadId); > + else localThreadHandle = > CreateThread(0,0,StartRoutine,StartContext,0,&ThreadId); > + > + if (localThreadHandle==0) return STATUS_SEVERITY_ERROR; > > You should use NtCreateThread rather than CreateThread. It also > returns a proper status, so you don't have to guess like you do here. > Actually, NtCreateThread is too low-level and so isn't implemented in Wine. RtlCreateUserThread might be better to use though. -- Rob Shearman

Sept. 20, 2007

Re: Implemented PsCreateSystemThread in ntoskrnl.exe

by Juan Lang

Hi Carroll, + if (ProcessHandle!=0) + localThreadHandle = CreateRemoteThread(ProcessHandle,0,0,StartRoutine,StartContext,0,&ThreadId); + else localThreadHandle = CreateThread(0,0,StartRoutine,StartContext,0,&ThreadId); + + if (localThreadHandle==0) return STATUS_SEVERITY_ERROR; You should use NtCreateThread rather than CreateThread. It also returns a proper status, so you don't have to guess like you do here. --Juan

Sept. 20, 2007

Re: [PATCH] fix small spelling mistakes

by Peter Oberndorfer

On Thursday 20 September 2007, Juan Lang wrote: > Hi Peter, > > > correct spelling of iff -> if > > Just so you know, iff is a common abbreviation meaning "if and only > if." I don't think that's any clearer in the cases you modified, so I > think your patch is fine, I just thought you (and others) might be > interested. > > --Juan > > > Hi, you are right, i never heard about using it as a abbreviation for "if and only if." Maybe i should have tried a dictionary first before fixing it :-) Thanks telling me Greetings Peter

Sept. 20, 2007

Re: Will ROS and WINE still be steady be synchronized ?

by Shachar Shemesh

Alexander Nicolaysen Sørnes wrote: > > ReactOS has been known for disassembling Microsoft binaries, which is illegal > in some countries, notably the US. As far as I understand this, if I disassemble Microsoft binaries (it is legal in Israel), then the resulting knowledge is legal to use - anywhere (in other words - the only one who can be sued is me, and the jurisdiction is Israel, where the action is legal). Do RoS people disassemble binaries in countries in which it is illegal to disassemble them? Shachar

Sept. 20, 2007

Re: WineD3D: better video memory reporting

by John Klehm

On 9/20/07, Juan Lang <juan.lang(a)gmail.com> wrote: > Your patch appears empty: > Came through ok here. --John Klehm

Sept. 20, 2007

[PATCH] fix small spelling mistakes

by Juan Lang

Hi Peter, > correct spelling of iff -> if Just so you know, iff is a common abbreviation meaning "if and only if." I don't think that's any clearer in the cases you modified, so I think your patch is fine, I just thought you (and others) might be interested. --Juan

Sept. 20, 2007

Re: WineD3D: better video memory reporting

by Juan Lang

Your patch appears empty: > Type: text/x-diff > Size: 0 bytes --Juan

Sept. 20, 2007

Re: setupapi(8/22): Implement SetupDiCreateDeviceInfoW

by Juan Lang

> This won't do what you want. You'd need *= 10 if the id is really > stored in decimal, though 0000 looks more like it should be in hex. You're right. It is decimal. Thanks. --Juan

Sept. 20, 2007

Re: winex11.drv/wintab32[3/5]: make wintab strings in winex11.drv Unicode, implement WTInfoW

by Alexandre Julliard

"Mikolaj Zalewski" <mikolajz(a)google.com> writes: > @@ -305,8 +312,8 @@ void X11DRV_LoadTabletInfo(HWND hwnddefault) > hwndTabletDefault = hwnddefault; > > /* Do base initializaion */ > - strcpy(gSysContext.lcName, "Wine Tablet Context"); > - strcpy(gSysDevice.NAME,"Wine Tablet Device"); > + MultiByteToWideChar(CP_UTF8, 0, "Wine Tablet Context", -1, gSysContext.lcName, WT_MAX_NAME_LEN); > + MultiByteToWideChar(CP_UTF8, 0, "Wine Tablet Device", -1, gSysDevice.NAME, WT_MAX_NAME_LEN); You should define Unicode constants directly, there's no need to convert at run-time. > @@ -695,10 +716,12 @@ int X11DRV_AttachEventQueueToTablet(HWND hOwner) > X11DRV_expect_error(data->display,Tablet_ErrorHandler,NULL); > for (cur_loop=0; cur_loop < gNumCursors; cur_loop++) > { > + char cursorNameA[WT_MAX_NAME_LEN]; > int event_number=0; > > + WideCharToMultiByte(CP_UNIXCP, 0, gSysCursor[cur_loop].NAME, -1, cursorNameA, WT_MAX_NAME_LEN, NULL, NULL); > for (loop=0; loop < num_devices; loop ++) > - if (strcmp(devices[loop].name,gSysCursor[cur_loop].NAME)==0) > + if (strncmp(devices[loop].name, cursorNameA, WT_MAX_NAME_LEN)==0) This doesn't look right; if the name overflows the buffer we shouldn't simply match against a truncated string. -- Alexandre Julliard julliard(a)winehq.org

Sept. 20, 2007

Re: Alexandre Julliard : wordpad.exe: We can now store binary files in the repository.

by Alexandre Julliard

Markus Amsler <markus.amsler(a)oribi.org> writes: > So we can also add ttf fonts and get rid of the fontforge dependency? No, I don't think ttf is an appropriate source format. -- Alexandre Julliard julliard(a)winehq.org

Sept. 20, 2007

Re: setupapi(8/22): Implement SetupDiCreateDeviceInfoW

by Alexandre Julliard

"Juan Lang" <juan.lang(a)gmail.com> writes: > +static DWORD SETUPDI_DevNameToDevID(LPCWSTR devName) > +{ > + LPCWSTR ptr; > + DWORD devNameLen = lstrlenW(devName), devInst = 0; > + BOOL valid = TRUE; > + > + TRACE("%s\n", debugstr_w(devName)); > + for (ptr = devName; valid && *ptr && ptr - devName < devNameLen; ) > + { > + if (isdigitW(*ptr)) > + { > + devInst <<= 8; This won't do what you want. You'd need *= 10 if the id is really stored in decimal, though 0000 looks more like it should be in hex. -- Alexandre Julliard julliard(a)winehq.org

Sept. 20, 2007

Re: kernel32: Don't crash accessing an invalid handle in GlobalSize [try3]

by James Hawkins

On 9/20/07, Dmitry Timoshkov <dmitry(a)codeweavers.com> wrote: > "James Hawkins" <truiken(a)gmail.com> wrote: > > > This version compares the pointer value to 64000 instead of using a > > try/catch block, as suggested by Alexandre. > > > - if(pintern->Magic==MAGIC_GLOBAL_USED) > > + if(pintern > (PGLOBAL32_INTERN)64000 && pintern->Magic==MAGIC_GLOBAL_USED) > > { > > if (!pintern->Pointer) /* handle case of GlobalAlloc( ??,0) */ > > It should be either 65535 (0xffff), or a test for HIWORD(pintern) being 0. > Also the cast should be to (char *) on both sides rather than (PGLOBAL32_INTERN). > I'll go with HIWORD(pintern), thanks for the suggestion. I knew it wasn't quite right :) -- James Hawkins

Sept. 20, 2007

Re: kernel32: Don't crash accessing an invalid handle in GlobalSize [try3]

by Dmitry Timoshkov

"James Hawkins" <truiken(a)gmail.com> wrote: > This version compares the pointer value to 64000 instead of using a > try/catch block, as suggested by Alexandre. > - if(pintern->Magic==MAGIC_GLOBAL_USED) > + if(pintern > (PGLOBAL32_INTERN)64000 && pintern->Magic==MAGIC_GLOBAL_USED) > { > if (!pintern->Pointer) /* handle case of GlobalAlloc( ??,0) */ It should be either 65535 (0xffff), or a test for HIWORD(pintern) being 0. Also the cast should be to (char *) on both sides rather than (PGLOBAL32_INTERN). -- Dmitry.

Sept. 20, 2007

Re: Handling New IDLs with widl

by Robert Shearman

Roy wrote: > I'm looking for advice on how to integrate a new idl into wine. > > I'm implementing bits and using the widl compiler on the IDL I wrote. My > current setup places the IDL in include/ with the build system generating a > header file that it posits there. So far, so good. > > I now need to generate bits_i.c and link that into my dll located in dlls/bits. > Adding: > > ---- > IDL_I_SRCS = $(TOPSRCDIR)/include/bits.idl > ---- > > to dlls/bits/Makefile.in solves this problem but drops the bits_i.c into > include/. This is rather messy. > Hi Roy, The way it's done for shared IDL files, like include/wine/irot.idl is to create a dummy IDL file in the directory you want to create the generated files in (see dlls/ole32/irot.idl) that simply includes the shared IDL file. Also, you might want to use your full name when submitting patches. -- Rob Shearman

Sept. 19, 2007

Handling New IDLs with widl

by Roy

Howdy, I'm looking for advice on how to integrate a new idl into wine. I'm implementing bits and using the widl compiler on the IDL I wrote. My current setup places the IDL in include/ with the build system generating a header file that it posits there. So far, so good. I now need to generate bits_i.c and link that into my dll located in dlls/bits. Adding: ---- IDL_I_SRCS = $(TOPSRCDIR)/include/bits.idl ---- to dlls/bits/Makefile.in solves this problem but drops the bits_i.c into include/. This is rather messy. Any advice on if I should (1) moving my IDL into dlls/bits/ and do everything locally or (2) dig into Make.rules.in and update the rule for *_i.c generation to allow the generated file to exist locally with the .idl sitting in include. The second idea is a little tricky since the current rule is careful to rename the generated .i file with the -U option: ---- .idl_i.c: $(WIDL) $(IDLFLAGS) -u -U $@ $< ---- which is a great idea, but uses the full path to the original idl when renaming the file. Probably good for the general case, but not for me. Thanks for the advice! -Roy

Sept. 19, 2007

Re: wineboot: Pass non-zero timeouts to SendMessageTimeout, otherwise the call won't timeout.

by Alexandre Julliard

Robert Shearman <rob(a)codeweavers.com> writes: > The original code seems to have been designed to not wait at all, but > this got changed by a user32 commit that made a timeout of zero behave > as INFINITE. I don't think so, AFAICT it has always been supposed to wait (except with the SMTO_ABORTIFHUNG flag). > So which should it be - no timeout or timeout followed by a dialog to > carry on waiting or kill the process? Either is OK, though a timeout followed by a dialog is better. What's not OK is a timeout and then a kill while the user might be trying to save the file. -- Alexandre Julliard julliard(a)winehq.org

Sept. 19, 2007

Re: Minimal Background Intelligent Transfer Service (bits) implementation

by Roy

Howdy, Resubmission of bits skeleton with a DLL_WINE_PREATTACH case added to DllMain. Thanks for the advice! Peace, -Roy On Wed, Sep 19, 2007 at 09:57:45AM +0200, Michael Stefaniuc wrote: > Hello Roy, > > Roy wrote: > > This patch provides a first skeleton version of bits.dll. > thanks for the patch but could you please add also a DLL_WINE_PREATTACH > case in DllMain()? That way Wine can prefer to load a native bits DLL > (if one is available) until the Wine one gets usable. See for example > the mscoree DLL for an example (dlls/mscoree/mscoree_main.c). > > thanks > bye > michael > > > From 95f4728e706861f201fb2b399e7ca5db128e71b3 Mon Sep 17 00:00:00 2001 > > From: Roy Shea <roy(a)cs.ucla.edu> > > Date: Tue, 18 Sep 2007 14:29:14 -0700 > > Subject: Minimal Background Intelligent Transfer Service (bits) implementation > > > > --- > > dlls/bits/Makefile.in | 16 ++++++++++++++++ > > dlls/bits/bits.spec | 4 ++++ > > dlls/bits/bits_main.c | 43 +++++++++++++++++++++++++++++++++++++++++++ > > 3 files changed, 63 insertions(+), 0 deletions(-) > > create mode 100644 dlls/bits/Makefile.in > > create mode 100644 dlls/bits/bits.spec > > create mode 100644 dlls/bits/bits_main.c > > > > diff --git a/dlls/bits/Makefile.in b/dlls/bits/Makefile.in > > new file mode 100644 > > index 0000000..2c9e293 > > --- /dev/null > > +++ b/dlls/bits/Makefile.in > > @@ -0,0 +1,16 @@ > > +TOPSRCDIR = @top_srcdir@ > > +TOPOBJDIR = ../.. > > +SRCDIR = @srcdir@ > > +VPATH = @srcdir@ > > +MODULE = bits.dll > > +IMPORTS = kernel32 > > +EXTRALIBS = > > + > > +C_SRCS = \ > > + bits_main.c > > + > > +#RC_SRCS = version.rc > > + > > +(a)MAKE_DLL_RULES@ > > + > > +(a)DEPENDENCIES@ # everything below this line is overwritten by make depend > > diff --git a/dlls/bits/bits.spec b/dlls/bits/bits.spec > > new file mode 100644 > > index 0000000..c5fc87a > > --- /dev/null > > +++ b/dlls/bits/bits.spec > > @@ -0,0 +1,4 @@ > > +@ stub DllCanUnloadNow > > +@ stub DllGetClassObject > > +@ stub DllRegisterServer > > +@ stub DllUnregisterServer > > diff --git a/dlls/bits/bits_main.c b/dlls/bits/bits_main.c > > new file mode 100644 > > index 0000000..37be887 > > --- /dev/null > > +++ b/dlls/bits/bits_main.c > > @@ -0,0 +1,43 @@ > > +/* > > + * Implementation of bits.dll > > + * Background Intelligent Transfer Service (bits) interface > > + * > > + * Copyright 2007 Google (Roy Shea) > > + * > > + * This library is free software; you can redistribute it and/or > > + * modify it under the terms of the GNU Lesser General Public > > + * License as published by the Free Software Foundation; either > > + * version 2.1 of the License, or (at your option) any later version. > > + * > > + * This library is distributed in the hope that it will be useful, > > + * but WITHOUT ANY WARRANTY; without even the implied warranty of > > + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU > > + * Lesser General Public License for more details. > > + * > > + * You should have received a copy of the GNU Lesser General Public > > + * License along with this library; if not, write to the Free Software > > + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA > > + */ > > + > > +#include <stdarg.h> > > + > > +#include "windef.h" > > +#include "wine/debug.h" > > + > > +WINE_DEFAULT_DEBUG_CHANNEL( bits ); > > + > > +BOOL WINAPI DllMain(HINSTANCE hinstDLL, DWORD fdwReason, LPVOID lpvReserved) > > +{ > > + TRACE("(%p, %d, %p)\n", hinstDLL, fdwReason, lpvReserved); > > + > > + switch (fdwReason) > > + { > > + case DLL_PROCESS_ATTACH: > > + DisableThreadLibraryCalls(hinstDLL); > > + break; > > + case DLL_PROCESS_DETACH: > > + break; > > + } > > + > > + return TRUE; > > +} > > > > > > ------------------------------------------------------------------------ > > > > > > > -- > Michael Stefaniuc Tel.: +49-711-96437-199 > Sr. Network Engineer Fax.: +49-711-96437-111 > -------------------------------------------------------------------- > Reg. Adresse: Red Hat GmbH, Hauptst??tter Strasse 58, 70178 Stuttgart > Handelsregister: Amtsgericht Stuttgart HRB 153243 > Gesch??ftsf??hrer: Brendan Lane, Charlie Peters, Michael Cunningham, > Werner Knoblich

Sept. 19, 2007

Re: wineboot: Pass non-zero timeouts to SendMessageTimeout, otherwise the call won't timeout.

by Robert Shearman

Alexandre Julliard wrote: > Robert Shearman <rob(a)codeweavers.com> writes: > > >> Set the default for WM_QUERYENDSESSION to be 5s which should be enough >> time for users to respond to any message boxes and WM_ENDSESSION to a >> lower 1s. >> > > If you are going to timeout on the user interaction you need to popup > a confirmation dialog of some kind to allow users to force killing the > app; you can't kill it just because the user didn't click on a "do you > want to save your file?" message box fast enough. The original code seems to have been designed to not wait at all, but this got changed by a user32 commit that made a timeout of zero behave as INFINITE. So which should it be - no timeout or timeout followed by a dialog to carry on waiting or kill the process? -- Rob Shearman

Sept. 19, 2007

Re: Alexandre Julliard : wordpad.exe: We can now store binary files in the repository.

by Markus Amsler

So we can also add ttf fonts and get rid of the fontforge dependency? Markus

Sept. 19, 2007

Re: widl [2/2]: Implement pointer descriptions for complex structures

by Robert Shearman

Dan Hipschman wrote: > On Tue, Sep 18, 2007 at 11:15:26PM +0100, Robert Shearman wrote: > >>> Sorry, I missed your last three patches somehow. My patch not only uses >>> RPC_FC_POINTER instead of RPC_FC_LONG here, but it also implements the >>> pointer description section for complex structures. >>> >> Right, but I was really talking about the approach of passing a flag >> into write_member_types or passing the containing type in. I think it is >> slightly cleaner to do the former, but your patch works very well for >> the case I was trying to fix. >> > > Well, it's such a small section of code it probably doesn't matter much > either way. I tend to avoid boolean arguments because when you look at > the call to the function and just see a bunch of TRUEs and FALSEs, it's > hard to remember what they mean. Sure, it doesn't really matter and your set of patches implements more functionality. I'm happy to drop my patches in favour of yours. -- Rob Shearman

Sept. 19, 2007

Re: winsock: Increase amount of iovec's to 32

by Alexandre Julliard

Maarten Lankhorst <m.b.lankhorst(a)gmail.com> writes: > I don't know why WS_MSG_MAXIOVLEN is used here, that seems to be a > message identifier rather then the actual size of MAXIOVLEN, so I added > a define for MAXIOVLEN and made it 32. MSG_MAXIOVLEN is not a message identifier, it's really supposed to be an iovec size. Do you have a test case showing that Windows uses 32 instead? -- Alexandre Julliard julliard(a)winehq.org

Sept. 19, 2007

Re: include: add a macro to declare critical sections as we declare them in Wine - initialized and with a debug

by Alexandre Julliard

"Mikolaj Zalewski" <mikolajz(a)google.com> writes: > Would such a patch be accepted? No, debug.h is about the libwine debugging API, critical sections have nothing to do with that. -- Alexandre Julliard julliard(a)winehq.org

Sept. 19, 2007

Re: winex11.drv: When minimizing a top level window move it outside of the visible desktop area

by Alexandre Julliard

Dmitry Timoshkov <dmitry(a)codeweavers.com> writes: > Changelog: > winex11.drv: When minimizing a top level window move it outside of > the visible desktop area. This will break apps that expect to paint their icon window, and probably also break the positioning of the icon title. It also doesn't work even for standard icons in desktop mode. -- Alexandre Julliard julliard(a)winehq.org

Sept. 19, 2007

Re: winex11.drv[1/4]: fix two bugs in WTInfoA

by Alexandre Julliard

"Mikolaj Zalewski" <mikolajz(a)google.com> writes: > Two bugs I've found working on previous implentations > From b233532f7adc140c4e36fb2288581fe0ff181935 Mon Sep 17 00:00:00 2001 > From: =?utf-8?q?Miko=C5=82aj_Zalewski?= <mikolaj(a)zalewski.pl> > Date: Mon, 17 Sep 2007 11:54:41 -0700 > Subject: [PATCH] winex11.drv: fix two bugs in WTInfoA Please write a better description of the changes, "fix bug" can apply to just about any commit (and in general, if you have two bugs to fix there should be two patches...) -- Alexandre Julliard julliard(a)winehq.org

Sept. 19, 2007

Re: wineboot: Pass non-zero timeouts to SendMessageTimeout, otherwise the call won't timeout.

by Alexandre Julliard

Robert Shearman <rob(a)codeweavers.com> writes: > Set the default for WM_QUERYENDSESSION to be 5s which should be enough > time for users to respond to any message boxes and WM_ENDSESSION to a > lower 1s. If you are going to timeout on the user interaction you need to popup a confirmation dialog of some kind to allow users to force killing the app; you can't kill it just because the user didn't click on a "do you want to save your file?" message box fast enough. -- Alexandre Julliard julliard(a)winehq.org

Sept. 19, 2007

Re: Minimal Background Intelligent Transfer Service (bits) implementation

by Michael Stefaniuc

Hello Roy, Roy wrote: > This patch provides a first skeleton version of bits.dll. thanks for the patch but could you please add also a DLL_WINE_PREATTACH case in DllMain()? That way Wine can prefer to load a native bits DLL (if one is available) until the Wine one gets usable. See for example the mscoree DLL for an example (dlls/mscoree/mscoree_main.c). thanks bye michael > From 95f4728e706861f201fb2b399e7ca5db128e71b3 Mon Sep 17 00:00:00 2001 > From: Roy Shea <roy(a)cs.ucla.edu> > Date: Tue, 18 Sep 2007 14:29:14 -0700 > Subject: Minimal Background Intelligent Transfer Service (bits) implementation > > --- > dlls/bits/Makefile.in | 16 ++++++++++++++++ > dlls/bits/bits.spec | 4 ++++ > dlls/bits/bits_main.c | 43 +++++++++++++++++++++++++++++++++++++++++++ > 3 files changed, 63 insertions(+), 0 deletions(-) > create mode 100644 dlls/bits/Makefile.in > create mode 100644 dlls/bits/bits.spec > create mode 100644 dlls/bits/bits_main.c > > diff --git a/dlls/bits/Makefile.in b/dlls/bits/Makefile.in > new file mode 100644 > index 0000000..2c9e293 > --- /dev/null > +++ b/dlls/bits/Makefile.in > @@ -0,0 +1,16 @@ > +TOPSRCDIR = @top_srcdir@ > +TOPOBJDIR = ../.. > +SRCDIR = @srcdir@ > +VPATH = @srcdir@ > +MODULE = bits.dll > +IMPORTS = kernel32 > +EXTRALIBS = > + > +C_SRCS = \ > + bits_main.c > + > +#RC_SRCS = version.rc > + > +(a)MAKE_DLL_RULES@ > + > +(a)DEPENDENCIES@ # everything below this line is overwritten by make depend > diff --git a/dlls/bits/bits.spec b/dlls/bits/bits.spec > new file mode 100644 > index 0000000..c5fc87a > --- /dev/null > +++ b/dlls/bits/bits.spec > @@ -0,0 +1,4 @@ > +@ stub DllCanUnloadNow > +@ stub DllGetClassObject > +@ stub DllRegisterServer > +@ stub DllUnregisterServer > diff --git a/dlls/bits/bits_main.c b/dlls/bits/bits_main.c > new file mode 100644 > index 0000000..37be887 > --- /dev/null > +++ b/dlls/bits/bits_main.c > @@ -0,0 +1,43 @@ > +/* > + * Implementation of bits.dll > + * Background Intelligent Transfer Service (bits) interface > + * > + * Copyright 2007 Google (Roy Shea) > + * > + * This library is free software; you can redistribute it and/or > + * modify it under the terms of the GNU Lesser General Public > + * License as published by the Free Software Foundation; either > + * version 2.1 of the License, or (at your option) any later version. > + * > + * This library is distributed in the hope that it will be useful, > + * but WITHOUT ANY WARRANTY; without even the implied warranty of > + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU > + * Lesser General Public License for more details. > + * > + * You should have received a copy of the GNU Lesser General Public > + * License along with this library; if not, write to the Free Software > + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA > + */ > + > +#include <stdarg.h> > + > +#include "windef.h" > +#include "wine/debug.h" > + > +WINE_DEFAULT_DEBUG_CHANNEL( bits ); > + > +BOOL WINAPI DllMain(HINSTANCE hinstDLL, DWORD fdwReason, LPVOID lpvReserved) > +{ > + TRACE("(%p, %d, %p)\n", hinstDLL, fdwReason, lpvReserved); > + > + switch (fdwReason) > + { > + case DLL_PROCESS_ATTACH: > + DisableThreadLibraryCalls(hinstDLL); > + break; > + case DLL_PROCESS_DETACH: > + break; > + } > + > + return TRUE; > +} > > > ------------------------------------------------------------------------ > > -- Michael Stefaniuc Tel.: +49-711-96437-199 Sr. Network Engineer Fax.: +49-711-96437-111 -------------------------------------------------------------------- Reg. Adresse: Red Hat GmbH, Hauptstätter Strasse 58, 70178 Stuttgart Handelsregister: Amtsgericht Stuttgart HRB 153243 Geschäftsführer: Brendan Lane, Charlie Peters, Michael Cunningham, Werner Knoblich

Sept. 19, 2007

Re: widl [2/2]: Implement pointer descriptions for complex structures

by Dan Hipschman

On Tue, Sep 18, 2007 at 11:15:26PM +0100, Robert Shearman wrote: > >Sorry, I missed your last three patches somehow. My patch not only uses > >RPC_FC_POINTER instead of RPC_FC_LONG here, but it also implements the > >pointer description section for complex structures. > > Right, but I was really talking about the approach of passing a flag > into write_member_types or passing the containing type in. I think it is > slightly cleaner to do the former, but your patch works very well for > the case I was trying to fix. Well, it's such a small section of code it probably doesn't matter much either way. I tend to avoid boolean arguments because when you look at the call to the function and just see a bunch of TRUEs and FALSEs, it's hard to remember what they mean. > >I actually left > >these as RPC_FC_LONG on purpose, until the pointer description part was > >finished. Isn't using RPC_FC_POINTER with a pointer description offset > >of zero just going to crash? It's not really any more correct in that > >case. There are no tests, so I'm not sure what it would do. > > Sure, but it was slightly better than the existing code, which would > have appeared to correctly unmarshal data, but fail in strange ways. If > you want to always output correct format strings then you have to detect > unimplemented cases and output an error! I agree. It's much better to crash than to do the wrong thing without warning. In both cases, though, it was impossible to write tests since the output was wrong, so there wasn't much point worrying about it. If I had gone to the effort of doing anything about it, I would have just implemented what was needed. I just didn't get around to it until now. Thanks for the comments.

Sept. 18, 2007

Re: widl [2/2]: Implement pointer descriptions for complex structures

by Robert Shearman

Dan Hipschman wrote: > On Tue, Sep 18, 2007 at 09:23:37AM +0100, Robert Shearman wrote: > >> Dan Hipschman wrote: >> >>> var_t *field, >>> } >>> else if (is_ptr(type)) >>> { >>> - print_file(file, 2, "0x8,\t/* FC_LONG */\n"); >>> + unsigned char fc = (cont->type == RPC_FC_BOGUS_STRUCT >>> + ? RPC_FC_POINTER >>> + : RPC_FC_LONG); >>> + print_file(file, 2, "0x%x,\t/* %s */\n", fc, string_of_type(fc)); >>> *tfsoff += 1; >>> } >>> else if (!write_base_type(file, type, tfsoff)) >>> >>> >> I sent a patch yesterday which does a similar thing to this chunk. I >> think we need to agree on which approach we're going to take so we can >> let Alexandre know which patch to commit. >> > > Sorry, I missed your last three patches somehow. My patch not only uses > RPC_FC_POINTER instead of RPC_FC_LONG here, but it also implements the > pointer description section for complex structures. Right, but I was really talking about the approach of passing a flag into write_member_types or passing the containing type in. I think it is slightly cleaner to do the former, but your patch works very well for the case I was trying to fix. > I actually left > these as RPC_FC_LONG on purpose, until the pointer description part was > finished. Isn't using RPC_FC_POINTER with a pointer description offset > of zero just going to crash? It's not really any more correct in that > case. There are no tests, so I'm not sure what it would do. > Sure, but it was slightly better than the existing code, which would have appeared to correctly unmarshal data, but fail in strange ways. If you want to always output correct format strings then you have to detect unimplemented cases and output an error! -- Rob Shearman

Sept. 18, 2007

Re: A first implentation of services.exe

by Robert Shearman

Mikolaj Zalewski wrote: > I don't want this patch to get committed but to maybe get some > comments. If the patch will be considered to be ok I can send it as > three smaller patches. It implements the basic infrastructure for > services.exe - a process that maintains the services database and > states. I think this can be useful in the long run. Absolutely. > This patch > currently doesn't implement anything useful - it connects to > services.exe in OpenSCManager and disconnect in CloseServiceHandle and > validating the database name is now done in services.exe but if this > approach is good I will add more functionality. > It looks good to me, but I'm a bit worried that you'll come across bugs/unimplemented features in widl when you start adding more calls. > Is there a way to check if a RPC connection over a named pipe was > closed by the client (probably we should close all it's handles in > such a case)? Yes, you can do this using context handles, but unfortunately we're lacking support for these on the RPC server side and in widl. > Also this program show a bug in our rpcrt4 - if we pass > NULL as a [in,unique,string] parameter rpcrt4 will crash because of a > strlen on it in NdrConformantStringBufferSize. To see it one can > remove the workaround in OpenSCManager and do a `make test`. That sounds like a bug in widl to me, and I'll take a look. > Macros > like RET_IF_ERROR should be removed from the final version or can they > stay? Definitely should be removed. > diff --git a/include/wine/svcctl_shared.h b/include/wine/svcctl_shared.h > new file mode 100644 > index 0000000..5687ddc > --- /dev/null > +++ b/include/wine/svcctl_shared.h > @@ -0,0 +1,35 @@ > +/* > + * SVCCTL (services.exe RPC interface) shared constants > + * > + * Copyright (C) 2007 Google (Mikolaj Zalewski) > + * > + * This library is free software; you can redistribute it and/or > + * modify it under the terms of the GNU Lesser General Public > + * License as published by the Free Software Foundation; either > + * version 2.1 of the License, or (at your option) any later version. > + * > + * This library is distributed in the hope that it will be useful, > + * but WITHOUT ANY WARRANTY; without even the implied warranty of > + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU > + * Lesser General Public License for more details. > + * > + * You should have received a copy of the GNU Lesser General Public > + * License along with this library; if not, write to the Free Software > + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA > + */ > + > +#ifndef __WINE_SVCCTL_SHARED_H > +#define __WINE_SVCCTL_SHARED_H > + > +#include <stdarg.h> > +#include <windef.h> > +#include <winbase.h> > + > +#define SVCCTL_TRANSPORT {'n','c','a','c','n','_','n','p',0} > +#define SVCCTL_ENDPOINT {'\\','p','i','p','e','\\','s','v','c','c','t','l',0} > + > +/* Not the Windows event name - if needed the true one can be found in Inside Windows */ > +#define SVCCTL_STARTED_EVENT (const WCHAR[]){'S','v','c','c','t','l','S','t','a','r','t','e','d',0} > + > + > +#endif /* __WINE_RPCSS_SHARED_H */ > We shouldn't need another file for this. You can put this same stuff in svcctl.idl using cpp_quote. > +DWORD RPC_MainLoop() Make sure to fully specify the parameters for functions with no parameters by declaring them as "(void)". -- Rob Shearman

Sept. 18, 2007

Re: widl [2/2]: Implement pointer descriptions for complex structures

by Dan Hipschman

On Tue, Sep 18, 2007 at 09:23:37AM +0100, Robert Shearman wrote: > Dan Hipschman wrote: > >var_t *field, > > } > > else if (is_ptr(type)) > > { > >- print_file(file, 2, "0x8,\t/* FC_LONG */\n"); > >+ unsigned char fc = (cont->type == RPC_FC_BOGUS_STRUCT > >+ ? RPC_FC_POINTER > >+ : RPC_FC_LONG); > >+ print_file(file, 2, "0x%x,\t/* %s */\n", fc, string_of_type(fc)); > > *tfsoff += 1; > > } > > else if (!write_base_type(file, type, tfsoff)) > > > > I sent a patch yesterday which does a similar thing to this chunk. I > think we need to agree on which approach we're going to take so we can > let Alexandre know which patch to commit. Sorry, I missed your last three patches somehow. My patch not only uses RPC_FC_POINTER instead of RPC_FC_LONG here, but it also implements the pointer description section for complex structures. I actually left these as RPC_FC_LONG on purpose, until the pointer description part was finished. Isn't using RPC_FC_POINTER with a pointer description offset of zero just going to crash? It's not really any more correct in that case. There are no tests, so I'm not sure what it would do.

Sept. 18, 2007

Re: Want to contact Mike McCormack

by Robert Shearman

Stefan Ensslen wrote: > I wanted to contact Mike McCormack from the WinE developer team, but my email to mike(a)codeweavers.com bounced. If you could send me his new email address, I'd be very happy. > > The question I wanted to ask Mr. McCormack is the following: > Dear Mike, > > I have found your code for generating proxy/stub-code at www.winehq.org. I wanted to ask where you have found a documentation of the functions used in this code (for example NdrProxyInitialize, NdrProxySendReceive, and so on). The MSDN doesn't provide any _valuable_ documentation for this function, so you surely have another source of documentation. > I'm not Mike, but I can probably answer this. Some of the knowledge about what NdrProxy* functions are supposed to do come from their use in MIDL-generated code. The rest of the knowledge comes from writing tests for the functions. The code for generating proxy/stub code was probably arrived at by comparing the output against MIDL and trying to get it as close as possible. -- Rob Shearman

Sept. 18, 2007

Re: AppDB down - too many SQL connections?

by Chris Morgan

This has happened in the past a handful of times. I suspect that our own code is taking the server down via retries. I'm thinking we should detect a too many connections error and cause the client to wait for a period of time before retrying. I'm not sure how best to do this or if this really is the cause of the problem. I tried to connect to the server to see what queries are running but I can't get connected to issue the request. Chris On 9/18/07, Alexander Nicolaysen Sørnes <alex(a)thehandofagony.com> wrote: > Hello, > > I am unable to access the database now because of too many MySQL connections. > Does anyone have a log file to indicate whether this is some sort of attack, > or whether we just have too many legitimate users? > > > Alexander N. Sørnes > > >

Sept. 18, 2007

Re: make crosstest is missing -lcabinet

by Pedro Araújo

On 9/17/07, John Klehm <xixsimplicityxix(a)gmail.com> wrote: > Sounds like you need to patch your mingw w32api (not surprising as the > way it comes is way out of sync with wine). > > These pages should help you: > [1] http://www.winehq.org/site/docs/winedev-guide/cross-compiling-tests > [2] http://www.winehq.org/site/docs/winedev-guide/testing-windows > > You will actually need more patches than are included in Hans > Leidekker's RPM (linked from [1]). You'll get different missing lib > errors otherwise. I believe Stefan Liechter has sent the needed > patches since Han's has updated his rpm. They are attached to this > mail. > > If you just want a binary winetest to run on windows its much easier > to grab paul millars: > Linked to from [2]. > > I haven't tested this lately and the w32api may need more patches > still in order to get it to compile the latest wine git tests. Paul > millars latest winetest is usually a good indication of when the last > time wine's tests were cross compilable. Thank you! I'll have to give this a shot, since I need to do some custom tests here. Should anything go wrong again, I'll shout back. Cheers, Pedro. -- "Compatibility means deliberately repeating other people's mistakes." - David John Wheeler

Sept. 18, 2007

AppDB down - too many SQL connections?

by Alexander Nicolaysen Sørnes

Hello, I am unable to access the database now because of too many MySQL connections. Does anyone have a log file to indicate whether this is some sort of attack, or whether we just have too many legitimate users? Alexander N. Sørnes

Sept. 18, 2007

Re: Problem with TCP Option 0x00000008

by Dmitry Timoshkov

"Juan Lang" <juan.lang(a)gmail.com> wrote: >> As I wrote in my other email, I obviously need to update the PSDK I have >> installed in Wine to get the header file. I'll look into that a bit later. I >> assume that TCP_ATMARK could be similar to SO_OOBINLINE, but that's a wild >> guess. I will need to write some test setup for this. > > I also don't have this header file or definition on my (latest, I believe) PSDK. I have here PSDK edition that is called "Microsoft Windows Software Development Kit Update for Windows Vista" and it contains ws2ipdef.h with '#define TCP_ATMARK 8'. -- Dmitry.

Sept. 18, 2007

Re: Problem with TCP Option 0x00000008

by Juan Lang

> I don't think so. SIOCATMARK is an WSAIoctl/ioctlsocket option in the winsock > api, too. We don't seem to handle that in WSAIoctl, it seems, patch coming > up. Fair enough, I didn't investigate it further than with grep. > As I wrote in my other email, I obviously need to update the PSDK I have > installed in Wine to get the header file. I'll look into that a bit later. I > assume that TCP_ATMARK could be similar to SO_OOBINLINE, but that's a wild > guess. I will need to write some test setup for this. I also don't have this header file or definition on my (latest, I believe) PSDK. --Juan

Sept. 18, 2007

Re: compile wine from sources - strange messages

by Robert Shearman

Andrey Gavrilin wrote: > I wish to compile wine from sources. Has received such message. The > distribution: Ubuntu 7.04. What headers files does not suffice? PS I > used the program the translator from Russian to English. See http://wiki.winehq.org/Recommended_Packages -- Rob Shearman

Sept. 18, 2007

Want to contact Mike McCormack

by Stefan Ensslen

Dear Sirs, I wanted to contact Mike McCormack from the WinE developer team, but my email to mike(a)codeweavers.com bounced. If you could send me his new email address, I'd be very happy. The question I wanted to ask Mr. McCormack is the following: Dear Mike, I have found your code for generating proxy/stub-code at www.winehq.org. I wanted to ask where you have found a documentation of the functions used in this code (for example NdrProxyInitialize, NdrProxySendReceive, and so on). The MSDN doesn't provide any _valuable_ documentation for this function, so you surely have another source of documentation. Thanks in advance, Stefan Ensslen -- GMX FreeMail: 1 GB Postfach, 5 E-Mail-Adressen, 10 Free SMS. Alle Infos und kostenlose Anmeldung: http://www.gmx.net/de/go/freemail

Sept. 18, 2007

Re: widl [2/2]: Implement pointer descriptions for complex structures

by Robert Shearman

Dan Hipschman wrote: > var_t *field, > } > else if (is_ptr(type)) > { > - print_file(file, 2, "0x8,\t/* FC_LONG */\n"); > + unsigned char fc = (cont->type == RPC_FC_BOGUS_STRUCT > + ? RPC_FC_POINTER > + : RPC_FC_LONG); > + print_file(file, 2, "0x%x,\t/* %s */\n", fc, string_of_type(fc)); > *tfsoff += 1; > } > else if (!write_base_type(file, type, tfsoff)) > I sent a patch yesterday which does a similar thing to this chunk. I think we need to agree on which approach we're going to take so we can let Alexandre know which patch to commit. -- Rob Shearman

Sept. 18, 2007

Re: kernel32: Don't crash accessing an invalid handle in GlobalSize

by Robert Shearman

James Hawkins wrote: > + if (IsBadReadPtr(pintern, 1)) > + { > + SetLastError(ERROR_INVALID_HANDLE); > + return 0; > + } > + > if(pintern->Magic==MAGIC_GLOBAL_USED) > { > if (!pintern->Pointer) /* handle case of GlobalAlloc( ??,0) */ It would be better to wrap the Magic access with an exception handler rather than use IsBadReadPtr. -- Rob Shearman

Sept. 18, 2007

Re: Problem with TCP Option 0x00000008

by Kai Blin

On Monday 17 September 2007 22:44:05 Juan Lang wrote: > > The TCP Option 0x00000008 is defined as TCP_ATMARK in ws2_ipdef.h in > > the Windows SDK. Does anyone know how this should be mapped ? > > Is this SIOCATMARK? I don't think so. SIOCATMARK is an WSAIoctl/ioctlsocket option in the winsock api, too. We don't seem to handle that in WSAIoctl, it seems, patch coming up. As I wrote in my other email, I obviously need to update the PSDK I have installed in Wine to get the header file. I'll look into that a bit later. I assume that TCP_ATMARK could be similar to SO_OOBINLINE, but that's a wild guess. I will need to write some test setup for this. Cheers, Kai -- Kai Blin WorldForge developer http://www.worldforge.org/ Wine developer http://wiki.winehq.org/KaiBlin Samba team member http://www.samba.org/samba/team/ -- Will code for cotton.

Sept. 18, 2007

Re: Problem with TCP Option 0x00000008

by Kai Blin

On Friday 14 September 2007 12:09:25 Christian Mahnke wrote: > The TCP Option 0x00000008 is defined as TCP_ATMARK in ws2_ipdef.h in > the Windows SDK. Does anyone know how this should be mapped ? What version of the SDK is that? It's not in the one I currently have installed. Arguably, my version is a bit older. I don't seem to have that header file at all. Cheers, Kai -- Kai Blin WorldForge developer http://www.worldforge.org/ Wine developer http://wiki.winehq.org/KaiBlin Samba team member http://www.samba.org/samba/team/ -- Will code for cotton.

Sept. 18, 2007

Re: fix for bug 9665

by Robert Shearman

James Hawkins wrote: > On 9/17/07, Martin <martin.marsan(a)gmail.com> wrote: > >> 2007-09-17 Martin Sand <deadlock(a)start.no> >> >> * dlls/ddraw/direct3d.c: >> ddraw: fixes a bug which prevents Kings Quest 8 from starting(bug 9665) >> >> > > A good changelog entry details what you changed, not what apps are > fixed by a change. Actually, a good changelog entry also describes why something was changed, if not obvious from the patch such as fixing a todo_wine in the tests. Although as James says, you'll need a better reason than just fixing a program though, as that hints at programming by trial-and-error rather than by reason. -- Rob Shearman

Sept. 17, 2007

Re: make crosstest is missing -lcabinet

by John Klehm

On 9/17/07, Pedro Araújo <inckie(a)gmail.com> wrote: > Hi all, > > I've been trying to cross-compile the Wine tests, but I keep getting this > error: > > $ ./configure && make crosstest > [SNIP] > i586-mingw32msvc-gcc advpack.cross.o files.cross.o install.cross.o > testlist.cross.o -o advpack_crosstest.exe -lcabinet -luser32 -ladvapi32 > -lkernel32 > /usr/lib/gcc/i586-mingw32msvc/3.4.5/../../../../i586-mingw32msvc/bin/ld: > cannot find -lcabinet > collect2: ld returned 1 exit status > make[2]: *** [advpack_crosstest.exe] Error 1 > make[2]: Leaving directory `/.../dlls/advpack/tests' > make[1]: *** [advpack/tests/__crosstest__] Error 2 > make[1]: Leaving directory `/.../dlls' > make: *** [dlls/__crosstest__] Error 2 > > Am I missing anything? Thanks in advance. > Sounds like you need to patch your mingw w32api (not surprising as the way it comes is way out of sync with wine). These pages should help you: [1] http://www.winehq.org/site/docs/winedev-guide/cross-compiling-tests [2] http://www.winehq.org/site/docs/winedev-guide/testing-windows You will actually need more patches than are included in Hans Leidekker's RPM (linked from [1]). You'll get different missing lib errors otherwise. I believe Stefan Liechter has sent the needed patches since Han's has updated his rpm. They are attached to this mail. If you just want a binary winetest to run on windows its much easier to grab paul millars: Linked to from [2]. I haven't tested this lately and the w32api may need more patches still in order to get it to compile the latest wine git tests. Paul millars latest winetest is usually a good indication of when the last time wine's tests were cross compilable. Regards, John Klehm

Sept. 17, 2007

Re: fix for bug 9665

by Stefan Dösinger

Am Montag, 17. September 2007 20:45:04 schrieb Martin: I doubt this fix is correct, as explained in bug 9665 a few minutes ago. It should fail the d3d test too

Sept. 17, 2007

Re: Problem with TCP Option 0x00000008

by Juan Lang

> The TCP Option 0x00000008 is defined as TCP_ATMARK in ws2_ipdef.h in > the Windows SDK. Does anyone know how this should be mapped ? Is this SIOCATMARK? --Juan

Sept. 17, 2007

Re: Problem with TCP Option 0x00000008

by Kai Blin

On Friday 14 September 2007 12:09:25 Christian Mahnke wrote: > The TCP Option 0x00000008 is defined as TCP_ATMARK in ws2_ipdef.h in > the Windows SDK. Does anyone know how this should be mapped ? Hm, I'll need to dig a bit, I guess. I've just returned from my vacation, so I'll need some time to catch up with other things, but I'll put this on my list. Cheers, Kai -- Kai Blin WorldForge developer http://www.worldforge.org/ Wine developer http://wiki.winehq.org/KaiBlin Samba team member http://www.samba.org/samba/team/ -- Will code for cotton.

Sept. 17, 2007

Re: fix for bug 9665

by James Hawkins

On 9/17/07, Martin <martin.marsan(a)gmail.com> wrote: > > > 2007-09-17 Martin Sand <deadlock(a)start.no> > > * dlls/ddraw/direct3d.c: > ddraw: fixes a bug which prevents Kings Quest 8 from starting(bug 9665) > A good changelog entry details what you changed, not what apps are fixed by a change. -- James Hawkins

Sept. 17, 2007