August 2013 - Wine-Devel - list.winehq.org

Re: [PATCH 01/11] widl: New option --rt for enabling rt's specific language extensions.
by André Hentschel Aug. 9, 2013

Aug. 9, 2013

Am 08.08.2013 17:51, schrieb Kai Tietz: > Hi, > > this patch adds new --rt option to widl. By it you can control, if > RT's IDL-language-extension(?s) getting active. > This initial patch just adds the option-scanning and the global > flag-variable to widl. > > Ok for apply? I'd say no, you are adding dead code, so you might want to merge it with 2/11 to avoid that.

3 2

Re: [2/2] wineserver: perform access check when creating event objects (try 2)
by Andrew Cook Aug. 9, 2013

Aug. 9, 2013

Actually ignore this patch, the problem was introduced by 92db6d2c back in 2007, server: Don't do access checks on the security descriptors of newly created objects which fixes sync.c:363 "CreateEventW with blank sd failed" i'll need to take another look at this, calling CreateEventEx with generic flags fails to correctly set permissions, the patch as it stands though will likely cause regressions. On 02/08/13 21:58, Andrew Cook wrote: > --- > dlls/ntdll/tests/om.c | 4 ++-- > server/event.c | 5 +---- > 2 files changed, 3 insertions(+), 6 deletions(-) > >

1 0

Re: configure.ac: allow disabling winemac.drv
by Ken Thomases Aug. 8, 2013

Aug. 8, 2013

On Aug 8, 2013, at 5:56 PM, Austin English wrote: > +AC_ARG_WITH(winemac, AS_HELP_STRING([--without-winemac],[do not build native Mac (Cocoa) driver]), > + [if test "x$withval" = "xno"; then ac_cv_header_ApplicationServices_ApplicationServices_h=no; fi]) A switch to disable the Mac driver is fine with me, but you can't pretend that the ApplicationServices framework header isn't present. There are other places within Wine which check that. http://source.winehq.org/source/dlls/windowscodecs/icnsformat.c#L24 http://source.winehq.org/source/dlls/winspool.drv/info.c#L48 -Ken

1 0

Re: [PATCH 04/11] widl: Handle default value for public vs hidden attribute.
by André Hentschel Aug. 8, 2013

Aug. 8, 2013

Am 08.08.2013 17:57, schrieb Kai Tietz: > - if (! is_attr(attrs, ATTR_PUBLIC)) > + if (! is_attr(attrs, ATTR_PUBLIC) && ! is_attr (attrs, ATTR_HIDDEN)) style issue: space after !

2 1

Re: [PATCH 09/11] widl: Handle LLP64 target same as LP64 one.
by André Hentschel Aug. 8, 2013

Aug. 8, 2013

Am 08.08.2013 18:07, schrieb Kai Tietz: > Hi, > > this thing here puzzles me most. I reject this patch, but would like > to get here some discussion to find the actual intended meaning of > this xstrtoul function. Then you should remove it from the patchset and send it as RFC to wine-devel

1 0

Re: Synchronization-patches of widl
by André Hentschel Aug. 8, 2013

Aug. 8, 2013

Am 08.08.2013 10:56, schrieb Kai Tietz: > All these changes were necessary to support the VLC-people to begin > with their WinRT port. By this we began on mingw-w64 to improve - > thanks to Jacek - our IDL-generated header-sources. So you are on a good way with WinRT, nice. What about Windows RT, was there already some ARM work? > A general note I would like to make about the use of sizeof(void*) in > widl to make out what type-library-variant to be choosen. That seems > to be a bit wacky variant, as it makes widl non-useable as cross-tool > on host with different pointer-size. isn't that controled by command line arguments?

1 0

winetestbot question
by Stefan Leichter Aug. 8, 2013

Aug. 8, 2013

Hello, i like to know why the WXPX64 VM of winetestbot does not have the 32 bit vc runtime 2008. The other 64 bit VMs have the runtime installed. Because of the missing runtime each patch for the unit tests of the dlls msvcr90 and msvcp90 is marked as "Failed" at http://source.winehq.org/patches -- Stefan

2 1

re: msvcrt: strncpy doesn't compliant C standard (try 3)
by Dan Kegel Aug. 8, 2013

Aug. 8, 2013

Minor problem: +static void test_strncpy(void) +{ + size_t len = 10; + char *ret; + char dst[len + 1]; Hmm. That last line is a VLA, and might not compile in all C compilers because it's not allowed in C89. http://stackoverflow.com/questions/448844/variable-sized-arrays-in-c Wine seems to want to build in c89 compilers: http://www.winehq.org/pipermail/wine-cvs/2011-April/076721.html Try using a #define with a more unique name for the length instead, or better yet, replace len=10 with len=sizeof(dst)-1 or something. Larger problem: ret = strncpy(dst, not_null_less, len); This is ill-defined. You're copying 10 bytes from a 5 byte buffer. Don't do that. Also, passing a not-null-terminated string in as a special test case doesn't seem interesting... just pass in the longer null-terminated string, and specify a shorter length. Fewer test strings needed, easier to read code. Please run 'make check', and test using http://testbot.winehq.org, before submitting patches.

2 1

Re: [PATCH 1/1] msvcrt: Set ?: environment variable in _(w)chdir functions.
by Marvin Aug. 8, 2013

Aug. 8, 2013

Hi, While running your changed tests on Windows, I think I found new failures. Being a bot and all I'm not very good at pattern recognition, so I might be wrong, but could you please double-check? Full results can be found at http://testbot.winehq.org/JobDetails.pl?Key=26639 Your paranoid android. === WINEBUILD (build) === Make failed

1 0

Re: msvcrt: strncpy doesn't compliant C standard (try 3)
by Marvin Aug. 7, 2013

Aug. 7, 2013

Hi, While running your changed tests on Windows, I think I found new failures. Being a bot and all I'm not very good at pattern recognition, so I might be wrong, but could you please double-check? Full results can be found at http://testbot.winehq.org/JobDetails.pl?Key=26638 Your paranoid android. === W2KPROSP4 (32 bit string) === string.c:2585: Test failed: dst != 012345 === WXPPROSP2 (32 bit string) === string.c:2585: Test failed: dst != 012345 === WXPX64 (32 bit string) === string.c:2585: Test failed: dst != 012345 === W2K3R2SESP2 (32 bit string) === string.c:2585: Test failed: dst != 012345 === WVISTAX64 (32 bit string) === string.c:2585: Test failed: dst != 012345 === W2K8SE (32 bit string) === string.c:2585: Test failed: dst != 012345 === W7PROX64 (32 bit string) === string.c:2585: Test failed: dst != 012345 === TEST64_W7SP1 (32 bit string) === string.c:2585: Test failed: dst != 012345

1 0

Re: msvcp90/tests: Add dynamically loaded functions for ARM
by Marvin Aug. 7, 2013

Aug. 7, 2013

Hi, While running your changed tests on Windows, I think I found new failures. Being a bot and all I'm not very good at pattern recognition, so I might be wrong, but could you please double-check? Full results can be found at http://testbot.winehq.org/JobDetails.pl?Key=26637 Your paranoid android. === WXPX64 (32 bit misc) === No test summary line found === WXPX64 (32 bit ios) === No test summary line found

1 0

Re: Request to add wow64 keyword to bugzilla
by Ken Sharp Aug. 7, 2013

Aug. 7, 2013

+1 from me. On 07/08/13 15:03, Rosanne DiMesio wrote: > On Wed, 07 Aug 2013 14:23:53 +0100 > Ken Sharp <kennybobs(a)o2.co.uk> wrote: > > >> Would I be right in assuming you would like to see bugs in 32-bit >> applications that are only present in a wow64 WINEPREFIX? Are there many? > > Yes. As to how many there are, I've been able to find 7 by doing searches of summaries/comments for things like "wow64" and "64" and then reading the bugs to see if they fit, but since people have such wildly varying ways of phrasing things, I don't know if that's all of them. I have seen it often enough on the forum that I now routinely tell 64 bit users to try a 32 bit wineprefix and have added instructions on how to create one to the FAQ. That's why I'd like an easier way to track these bugs. I see. A lot of people say this over and over again. +1 from me for a wow64 keyword. They may even be fairly easy to solve (picking the right paths / registry entries). > > As to keyword usage, some of those bugs were tagged with the win64 keyword, some weren't. I know there is also a win32 keyword, and perhaps that is meant to be used for the kinds of bugs I am looking for, but if so, that is not at all clear from the keyword definition, and it is not how people are using the keywords. > There's no win32 keyword?

3 2

Re: msvcp90/tests: Skip tests on missing functions
by Piotr Caban Aug. 7, 2013

Aug. 7, 2013

On 08/06/13 23:00, André Hentschel wrote: > @@ -749,6 +749,13 @@ static BOOL init(void) > > SET(p_basic_istream_char_read_uint64, > "??5?$basic_istream(a)DU?$char_traits(a)D@std@@@std@@QAEAAV01(a)AA_K@Z"); > + > + if (!p_basic_istream_char_read_uint64) > + { > + skip("_basic_istream_char_read_uint64 not found, skipping tests\n"); > + return FALSE; > + } Why are you skipping the tests instead of properly initializing functions pointers on ARM?

2 1

Request to add wow64 keyword to bugzilla
by Rosanne DiMesio Aug. 7, 2013

Aug. 7, 2013

Can wow64 be added as a keyword to bugzilla? I know we already have win64 as a keyword, but that's being used for both 64 bit apps and 32 bit apps in a 64 bit wineprefix. I'm interested in being able to track the latter, as it's hitting increasing numbers of users. -- Rosanne DiMesio <dimesio(a)earthlink.net>

2 1

Re: [PATCH] typo fixes (https://github.com/vlajos/misspell_fixer)
by Frédéric Delanoy Aug. 7, 2013

Aug. 7, 2013

On Mon, Aug 5, 2013 at 9:41 PM, Veres Lajos <vlajos(a)gmail.com> wrote: > > Signed-off-by: Veres Lajos <vlajos(a)gmail.com> > --- > dlls/comctl32/tests/trackbar.c | 2 +- > documentation/ChangeLog.ALPHA | 80 ++++++++++++++++++++-------------------- > documentation/ChangeLog.BETA | 44 +++++++++++----------- > documentation/ChangeLog.OLD | 24 ++++++------ > tools/winebuild/utils.c | 2 +- > 5 files changed, 76 insertions(+), 76 deletions(-) The "documentation/ChangeLog.*" files have been removed so the patch won't apply. Frédéric Delanoy

1 0

Re: msvcrt: strncpy doesn't compliant C standard (try 2)
by Frédéric Delanoy Aug. 7, 2013

Aug. 7, 2013

On Tue, Aug 6, 2013 at 10:44 PM, Álvaro Nieto <alvaro.nieto(a)gmail.com> wrote: > Remove strlen and implement some tests You should probably add tests for when the source string is not null-terminated. Frédéric Delanoy

1 0

Re: msvcp90/tests: Skip tests on missing functions
by Marvin Aug. 6, 2013

Aug. 6, 2013

Hi, While running your changed tests on Windows, I think I found new failures. Being a bot and all I'm not very good at pattern recognition, so I might be wrong, but could you please double-check? Full results can be found at http://testbot.winehq.org/JobDetails.pl?Key=26628 Your paranoid android. === WXPX64 (32 bit misc) === No test summary line found === WXPX64 (32 bit ios) === No test summary line found

1 0

Re: msvcr90/tests: Skip when _type_info_name_internal_method is not found
by Marvin Aug. 6, 2013

Aug. 6, 2013

Hi, While running your changed tests on Windows, I think I found new failures. Being a bot and all I'm not very good at pattern recognition, so I might be wrong, but could you please double-check? Full results can be found at http://testbot.winehq.org/JobDetails.pl?Key=26627 Your paranoid android. === WXPX64 (32 bit msvcr90) === No test summary line found

1 0

Re: ws2_32/tests: Fix a comment
by Juan Lang Aug. 6, 2013

Aug. 6, 2013

Hi Andre, - /* this is a io heavy test, do it at the end so the kernel doesn't start dropping packets */ + /* this is a heavy io test, do it at the end so the kernel doesn't start dropping packets */ To my eyes, this isn't an improvement. A slight improvement might be "this is a io-heavy test", or "this is an io-heavy test", but I'm not sure it's enough of an improvement to bother. --Juan

1 0

Re: msvcrt: strncpy doesn't compliant C standard
by Nikolay Sivov Aug. 6, 2013

Aug. 6, 2013

On 8/6/2013 11:09, Álvaro Nieto wrote: > This patch solves [Bug 34211]. The implementation of strncpy function > doesn't compliant with C standard [1]. Also Microsoft Visual Studio > C/C++ compiler is ok with the standard [2]. > > Extract from msdn; > > "The strncpy function copies the initial count characters of strSource > to strDest and returns strDest. If count is less than or equal to the > length of strSource, a null character is not appended automatically to > the copied string. If count is greater than the length of strSource, > the destination string is padded with null characters up to length > count. The behavior of strncpy is undefined if the source and > destination strings overlap." > > [1] http://pubs.opengroup.org/onlinepubs/009695399/functions/strncpy.html > [2] http://msdn.microsoft.com/en-us/library/vstudio/xdsywd25.aspx This needs tests.

1 0

Fwd: [AppDB] Submitted test data deleted
by Ken Sharp Aug. 5, 2013

Aug. 5, 2013

Deleting test results because the Wine version is old? What the Hell is the plan there? -------- Original Message -------- Subject: [AppDB] Submitted test data deleted Date: Mon, 05 Aug 2013 15:48:43 -0500 From: AppDB <appdb-noreply(a)winehq.org> Reply-To: AppDB <appdb-noreply(a)winehq.org> To: appdb-noreply(a)winehq.org Submitted test data deleted ------------------------------------------------------- The test report you submitted for 'Microsoft Access 2007' has been deleted. The action was performed by Joerg Schiermeier Reasons given Old wine - deprecated. Best regards. The AppDB team http://appdb.winehq.org/ If you don't want to receive any other e-mail, please change your preferences: http://appdb.winehq.org/preferences.php

1 0

wine mannual is wrong?
by 中川祥 Aug. 5, 2013

Aug. 5, 2013

I'm interpretering wine.man. I found it says"this requires X11 to run". But the newest ver. does not require X11,does it?

3 2

Draft: improve IMC/IMCC related functions
by Qian Hong Aug. 5, 2013

Aug. 5, 2013

Hello, This is a follow up of http://www.winehq.org/pipermail/wine-devel/2013-July/100459.html ( [PATCH] imm32: Fixed crashing in ImmGetIMCCSize. ) I believe the attached patchset is the correct fix for the Office crashing bug, how ever, I'm looking for a more graceful way to improve ImmCreateContext and other IMC functions, I appreciate to any feedback on the patches especially Patch 6/6. My tests show that Windows return a HGLOBAL hmem for ImmCreateIMCC() as the returned HIMCC, so it is easy to re-implement ImmCreateIMCC in a graceful way. However, Windows return a magic number for ImmCreateContext, (for instance, have a look at 000B01D9 in the test show in https://testbot.winehq.org/JobDetails.pl?Key=26619&log_207=1#k207) I have no idea how that magic number comes, while I understand there is no need to re-implement ImmCreateContext in the same way as Windows does, I'm still interesting in a more beautiful implementation like my ImmCreateIMCC reached, get rid of casting InputContextData to HIMC if possible. Thanks in advance for any suggestions! If there is no better way, I'll send this series to wine-patches. -- Regards, Qian Hong - http://www.winehq.org

1 0

Re: inputscope.idl: Imported from mingw-w64.
by Ken Sharp Aug. 5, 2013

Aug. 5, 2013

On 05/08/13 12:14, Jacek Caban wrote: > + * No warranty is given; refer to the file DISCLAIMER.PD within this package. It's a minor point but this may be a bit confusing given the file doesn't exist.

1 0

Re: msvcrt: invoke invalid_handler in _wcscpy_s
by Piotr Caban Aug. 5, 2013

Aug. 5, 2013

On 08/02/13 20:08, Daniel Lehman wrote: > + if(!MSVCRT_CHECK_PMT(wcDest)) > + { > + *MSVCRT__errno() = MSVCRT_EINVAL; > return MSVCRT_EINVAL; > + } The MSVCRT_CHECK_PMT macro sets the errno value. There's also MSVCRT_CHECK_PMT_ERR macro if you need to set different error.

1 0

po: Add English (Philippines) resource
by Ken Sharp Aug. 5, 2013

Aug. 5, 2013

Hi everyone, Would anyone mind looking at the attached patch? It adds the English (Philippines) resource by linking (or copying) the English (US) resource. See http://bugs.winehq.org/show_bug.cgi?id=23124 Actually, the attached version links to fr.po (French) to make testing much easier. The final won't do that. It's explained in the bug. I have tested this successfully in Linux only. I wanted to test others but: Cygwin 1.7.22: creates its own "symlink" but no translations are built at all. Compilation is basically broken. PC-BSD 9.1: complains about gettext although the config.log seems to complain about gm4. I gave up. OpenIndiana: can't seem to find its own packages. Gave up. Debian kFreeBSD: all sorts of problems in VBox. Might try again later. So clearly I've had no luck on other platforms. If anyone else could quickly test it, that would be great! Thanks to Austin and François for the help thus far. Of course, all feedback welcome. TIA, Ken

2 3

Wiki RFC: Redirects, swarm tactics, etc.
by Kyle Auble Aug. 5, 2013

Aug. 5, 2013

So I've finished with pretty much all of the edits I had in mind for the wiki, but before I ride off into the sunset for a while, I wanted to toss out a few ideas. 1. Do we want some kind of guideline on redirects for the wiki? Some stable "interface" pages to the main site might be good, but beyond that, I think minimizing redirects makes sense in this use case. 2. There's still a lot of old/missing content on the wiki, and much of it requres a good sense of where the code is at. Also, it might be too overwhelming for one person to work in depth on more than a few pages at this point. I feel like a semi-coordinated swarm of editing might be the best bet for further improvements. I was picturing a table of all relevant pages, then people could adopt one or two to work on, then strike/delete a record once that page is finished. Any thoughts? 3. There are actually a few more fixes to the theme code at the head of my bitbucket repo (and also branches for 2 different Moinmoin upgrade paths). I'm cool with keeping the theme code for the near future; if and when we move it onto WineHQ's git server though, just let me know so I can note that I'm not upstream anymore. 4. Finally, spammers... they keep coming... and they're getting smarter. We can mostly stalemate them with the regex filter, but it blocks legit edits too sometimes (in very annoying fashion). In fact, I think there are a few of them that have learned to turn the filter against us by completely wiping pages with false positives so we can't revert the page. Are there any relatively easy things we could do to cut the spam? I don't know, but it's possible the newer version of Moinmoin has more potent anti-spam tools. -Kyle

5 5

help [Abount Sql2005 and WMI]
by Jing Li Aug. 4, 2013

Aug. 4, 2013

Hi, all: I have met a problem when installing sql2005, and it seems related to some WMI problems. As WMI has just been implemented in wine-1.6 ( right? ), I want to know in what degree it has been implemented.By the way is there anyone succeed in install of it? Thanks. Appreciate your kind help.

2 1

Vacation
by Alexandre Julliard Aug. 2, 2013

Aug. 2, 2013

Folks, I'll be on vacation for the next 10 days, so you'll have to live without commits for a while... -- Alexandre Julliard julliard(a)winehq.org

2 1

Re: [docs] wineusr: Assorted formatting & tagging fixes (try 2)
by Alexandre Julliard Aug. 2, 2013

Aug. 2, 2013

Frédéric Delanoy <frederic.delanoy(a)gmail.com> writes: > @@ -710,17 +714,18 @@ Path="c:\windows;c:\windows\system;e:\;e:\test;f:\" > choose between several edition modes: > <itemizedlist> > <listitem> > - <para> > - Emacs: the same keybindings as under > - emacs are available. For example, Ctrl-A > - will bring the cursor to the beginning > - of the edition line. See your emacs > - manual for the details of the commands. > - </para> > + <para> > + <literal>Emacs</literal>: the same keybindings as under > + <command>emacs</command> are available. For example, > + <keycombo><keycap>Ctrl</keycap><keycap>A</keycap></keycombo> > + will bring the cursor to the beginning > + of the edition line. See your <command>emacs</command> > + manual for the details of the commands. > + </para> Please avoid messing up the indentation. -- Alexandre Julliard julliard(a)winehq.org

2 2

Re: gdi32: Avoid a crash by loading AppKit before using Core Text if might use the Mac driver.
by Alexandre Julliard Aug. 2, 2013

Aug. 2, 2013

Ken Thomases <ken(a)codeweavers.com> writes: > + /* If the Mac driver might be used, then load AppKit now before using the Core Text API. > + Otherwise, AppKit crashes on Mac OS X 10.7+ because CTFontDescriptor isn't toll-free > + bridged to NSCTFontDescriptor. That bridging is only set up if AppKit has been loaded > + when CTFontDescriptor is first used. */ > + /* @@ Wine registry key: HKCU\Software\Wine\Drivers */ > + if (!RegOpenKeyA(HKEY_CURRENT_USER, "Software\\Wine\\Drivers", &hkey)) > + { > + char buffer[MAX_PATH]; > + DWORD type, count = sizeof(buffer); > + > + if (!RegQueryValueExA(hkey, "Graphics", 0, &type, (LPBYTE) buffer, &count)) > + { > + char *name = buffer; > + > + usingMacDriver = FALSE; > + while (name) > + { > + char *next = strchr(name, ','); > + if (next) *next++ = 0; > + > + if (!strcmp(name, "mac")) > + { > + usingMacDriver = TRUE; > + break; > + } > + > + name = next; > + } > + } > + > + RegCloseKey(hkey); > + } That's ugly. freetype.c has no business knowing about the details of the graphics driver. -- Alexandre Julliard julliard(a)winehq.org

2 5

Re: riched20: Make ITextDocument stub reusable for ITextServices (try 4)
by Alexandre Julliard Aug. 2, 2013

Aug. 2, 2013

Caibin Chen <tigersoldi(a)gmail.com> writes: > diff --git a/dlls/riched20/tomimpl.h b/dlls/riched20/tomimpl.h > new file mode 100644 > index 0000000..886c3a8 > --- /dev/null > +++ b/dlls/riched20/tomimpl.h > @@ -0,0 +1,59 @@ > +/* > + * RichEdit - TOM interfaces implementations > + * > + * Copyright 2013 by Caibin Chen > + * > + * 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 __RICHOLE_H > +#define __RICHOLE_H > + > +#include <tom.h> > + > +#include "editstr.h" > + > +typedef struct ITextDocumentImpl ITextDocumentImpl; > +typedef struct ITextSelectionImpl ITextSelectionImpl; > + > +/* ITextDocument */ > + > +struct ITextDocumentImpl { > + ITextDocument ITextDocument_iface; > + IUnknown *outer_unk; > + ME_TextEditor *editor; > + ITextSelectionImpl *txtSel; > +}; You shouldn't need a new header either. Particularly since implementation details shouldn't be global. > +/** > + * Create an ITextDocumentImpl object, which acts as an inner object of COM aggression. I'm sure writing COM code can make you feel aggressive, but that's not the right term ;-) > + * > + * The ITextDocumentImpl object will delegate all of its IUnknown calls to the specified > + * {@code outer_unk}. > + * > + * @param outer_unk The outer objects that creates and delegates ITextDocument > + * method calls to the created ITextDocumentImpl. > + * @param editor The editor implementation. > + * > + * @return An ITextDocumentImpl object that should be destroyed with > + * {@code ITextDocumentImpl_destroy} > + */ Please don't use that sort of documentation markup. Besides, things like parameters and return should be obvious from the code, you shouldn't have to add documentation for this, especially not for internal functions. -- Alexandre Julliard julliard(a)winehq.org

2 1

Re: xmllite: Improve attribute value parsing
by Alexandre Julliard Aug. 2, 2013

Aug. 2, 2013

Nikolay Sivov <nsivov(a)codeweavers.com> writes: > +static void reader_normalize_space(xmlreader *reader, WCHAR *ptr) > +{ > + encoded_buffer *buffer = &reader->input->buffer->utf16; > + > + if (!is_wchar_space(*ptr)) return; > + > + if (*ptr == '\r' && *ptr == '\n') That's not going to happen very often... -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [PATCH] oleaut32: ERR on 32bit typelibs on 64bit
by Alexandre Julliard Aug. 2, 2013

Aug. 2, 2013

Marcus Meissner <meissner(a)suse.de> writes: > Triggered by the Novell Groupwise installer currently, > from the ISBE64W.exe Installshield Helper program. > > The typelib is contained directly in ISBE64W.exe itself :/ > > Basically if we encounter this case the program will crash > anyway (but with random backtrace), so output an ERR(). That looks like it's just hiding the problem. -- Alexandre Julliard julliard(a)winehq.org

2 1

Re: d3dx9 [patch 1/2]: Implement D3DXCreatePolygon
by Matteo Bruni Aug. 2, 2013

Aug. 2, 2013

2013/8/1 Nozomi Kodama <nozomi.kodama(a)yahoo.com>: > + vertices = HeapAlloc(GetProcessHeap(), 0, 2 * (sides + 1) * sizeof(D3DXVECTOR3)); + if (!vertices) + { + TRACE("Not memory enough for vertex buffer\n"); + polygon->lpVtbl->Release(polygon); + return E_OUTOFMEMORY; + } + + hr = polygon->lpVtbl->LockVertexBuffer(polygon, D3DLOCK_DISCARD, (VOID **)&vertices); Why are you allocating memory and immediately dropping track of it? Recheck how ID3DXMesh::LockVertexBuffer works. Also I'm having a deja-vu here, this patch looks suspiciously similar (up to having this same issue) to one I've already reviewed in the past.

2 1

appwiz: Correct Wine Mono installer title in several languages
by Ken Sharp Aug. 2, 2013

Aug. 2, 2013

Hi Alexandre, http://source.winehq.org/patches/data/97460 should be correct but would I be right in assuming you would rather have the changes made along with the next string for each language?

2 4

Re: [PATCH 4/5] d3dx9/tests: Add ID3DXConstantTable struct test.
by Matteo Bruni Aug. 2, 2013

Aug. 2, 2013

2013/7/24 Rico Schüller <kgbricola(a)web.de>: > --- > dlls/d3dx9_36/tests/shader.c | 308 > +++++++++++++++++++++++++++++++++++++++++++ > 1 file changed, 308 insertions(+) > This is okay, but as a followup can you add some tests with mixed-type structs? Something like: struct { float f; int i; bool b; }; If you have already written tests of this kind, I'd like to know what does the compiler do in this case :)

2 6

Re: [1/2] ntdll: Implement NtQueryEvent
by Alexandre Julliard Aug. 2, 2013

Aug. 2, 2013

Andrew Cook <ariscop(a)gmail.com> writes: > diff --git a/dlls/ntdll/tests/Makefile.in b/dlls/ntdll/tests/Makefile.in > index 10d6674..a44b880 100644 > --- a/dlls/ntdll/tests/Makefile.in > +++ b/dlls/ntdll/tests/Makefile.in > @@ -7,6 +7,7 @@ C_SRCS = \ > directory.c \ > env.c \ > error.c \ > + event.c \ > exception.c \ > file.c \ > generated.c \ Please don't add a new file for just a couple of tests, merge this into an existing file. -- Alexandre Julliard julliard(a)winehq.org

1 0

jscript FTBFS in Cygwin 1.7.22
by Ken Sharp Aug. 2, 2013

Aug. 2, 2013

Probably better to post this here rather than the forums: http://forum.winehq.org/viewtopic.php?f=2&t=19501 "Not sure if this is a Wine bug (as usual) so I'd rather put this here than to open a new bug. On Cygwin 1.7.22 the compilation stops at jscript, apparently a conflict in the declaration of dtoa. $ make ccache cc -c -I/home/ken/wine-git/dlls/jscript -I. -I/home/ken/wine-git/include -I../../include -D__WINESRC__ -D_REENTRANT -Wall -pipe -fno-strict-aliasing -Wdeclaration-after-statement -Wempty-body -Wignored-qualifiers -Wstrict-prototypes -Wtype-limits -Wunused-but-set-parameter -Wwrite-strings -Wpointer-arith -Wlogical-op -gdwarf-2 -gstrict-dwarf -fno-omit-frame-pointer -g -O0 -D_WIN32 -o jscript_main.o /home/ken/wine-git/dlls/jscript/jscript_main.c ccache cc -c -I/home/ken/wine-git/dlls/jscript -I. -I/home/ken/wine-git/include -I../../include -D__WINESRC__ -D_REENTRANT -Wall -pipe -fno-strict-aliasing -Wdeclaration-after-statement -Wempty-body -Wignored-qualifiers -Wstrict-prototypes -Wtype-limits -Wunused-but-set-parameter -Wwrite-strings -Wpointer-arith -Wlogical-op -gdwarf-2 -gstrict-dwarf -fno-omit-frame-pointer -g -O0 -D_WIN32 -o lex.o /home/ken/wine-git/dlls/jscript/lex.c ccache cc -c -I/home/ken/wine-git/dlls/jscript -I. -I/home/ken/wine-git/include -I../../include -D__WINESRC__ -D_REENTRANT -Wall -pipe -fno-strict-aliasing -Wdeclaration-after-statement -Wempty-body -Wignored-qualifiers -Wstrict-prototypes -Wtype-limits -Wunused-but-set-parameter -Wwrite-strings -Wpointer-arith -Wlogical-op -gdwarf-2 -gstrict-dwarf -fno-omit-frame-pointer -g -O0 -D_WIN32 -o number.o /home/ken/wine-git/dlls/jscript/number.c /home/ken/wine-git/dlls/jscript/number.c:57:20: error: conflicting types for ‘dtoa’ In file included from /home/ken/wine-git/include/wine/port.h:47:0, from /home/ken/wine-git/dlls/jscript/number.c:20: /usr/include/stdlib.h:161:35: note: previous declaration of ‘dtoa’ was here Makefile:199: recipe for target `number.o' failed make: *** [number.o] Error 1 dlls/jscript/number.c: static inline void dtoa(double d, WCHAR *buf, int size, int *dec_point) /usr/include/stdlib.h: char * _EXFUN(dtoa,(double, int, int, int *, int*, char**));"

1 0

Re: ntdll: Implement compatible FindActCtxSectionString() for window class section (try3)
by Alexandre Julliard Aug. 2, 2013

Aug. 2, 2013

Nikolay Sivov <nsivov(a)codeweavers.com> writes: > @@ -1031,6 +1105,17 @@ static BOOL parse_window_class_elem(xmlbuf_t* xmlbuf, struct dll_redirect* dll) > > if (!(entity->u.class.name = xmlstrdupW(&content))) return FALSE; > > + /* each class entry needs index, data and string data */ > + acl->actctx->wndclass_section_len += sizeof(struct wndclass_index); > + acl->actctx->wndclass_section_len += sizeof(struct wndclass_redirect_data); > + /* original name is stored separately */ > + acl->actctx->wndclass_section_len += aligned_string_len((content.len+1)*sizeof(WCHAR)); > + /* versioned name and module name are stored one after another */ > + len = get_assembly_version(assembly, NULL) + content.len + 2 /* null terminator and '!' separator */; > + len += strlenW(dll->name) + 1; > + acl->actctx->wndclass_section_len += aligned_string_len(len*sizeof(WCHAR)); > + acl->actctx->wndclass_count++; > + It would be better to do this at the time you build the structure. The parser shouldn't need to worry about this. -- Alexandre Julliard julliard(a)winehq.org

2 2

Re: po: Add English (Canada) resource [Take 2]
by Ken Sharp Aug. 2, 2013

Aug. 2, 2013

Please disregard this patch. I missed a couple of words and with those corrected en_CA.po becomes identical to en.po (except that "serialisation" is fuzzy) and Wine defaults to that anyway for Canadian English. Will re-submit if there are any changes (just noticed -able/-eable). Will keep an eye on the English translations. On 30/07/13 22:37, Ken Sharp wrote: > - Re-based to latest git > - Removed all the fuzzies thanks to http://www.etymonline.com/index.php > > Turns out all the -ise/-ize words used in Wine are of Latin origin anyway. > > > -------- Original Message -------- > Subject: po: Add English (Canada) resource > Date: Mon, 22 Jul 2013 19:28:51 +0100 > From: Ken Sharp <kennybobs(a)o2.co.uk> > To: wine-patches(a)winehq.org > > Canadian English follows some British/French rules but not others. > > I have left some of the -ise/-ize words fuzzy because: > > "the etymological convention that verbs derived from Greek roots are > spelled with -ize and those from Latin with -ise is preserved in that > practice." > http://en.wikipedia.org/wiki/Canadian_english#Spelling_and_dictionaries > > I tried looking for the origins of the words and I couldn't get any. The > non-fuzzies are correct. > > With British, US and Canadian I believe all the English variations (used > in Wine) are covered, except Philippines (follows US rules). > > >

1 1

Re: reg.exe: Add query function
by Nikolay Sivov Aug. 2, 2013

Aug. 2, 2013

On 8/1/2013 17:44, Hugh McMaster wrote: > This patch builds significantly on the query function stub in programs/reg.exe. The patch includes recursion for subkeys. > > Some examples of usage: > > wine reg.exe query "HKCU" > wine reg.exe query "HKCU\Console" > wine reg.exe query "HKCU\Console" /v "ScreenSize" > wine reg.exe query "HKCU\Software" /s > > --- > programs/reg/reg.c | 232 +++++++++++++++++++++++++++++++++++++++++++++++++--- > 1 file changed, 221 insertions(+), 11 deletions(-) > > +static LPVOID WINAPI mem_alloc(size_t size) > +{ > + return HeapAlloc(GetProcessHeap(), 0, size); > +} > + > +static void WINAPI mem_free(LPVOID mem) > +{ > + HeapFree(GetProcessHeap(), 0, mem); > +} Why WINAPI? It's better to follow naming from existing module, there's plenty of them with similar helpers. > + case REG_SZ: > + case REG_MULTI_SZ: > + case REG_EXPAND_SZ: No tabs please. > +static void reg_output_subkey(LPWSTR key_name, LPWSTR subkeyName) > +{ > + reg_printfW(formatSW, key_name); > + > + if (subkeyName > 0) > + { > + reg_printfW(formatSW, slashW); > + reg_printfW(formatSW, subkeyName); > + reg_printfW(formatSW, newlineW); > + } > +} Comparison looks a bit odd. Also why do you need to break it down to three printf calls instead of one? > + p = strchrW(key_name,'\\'); > + if (!p) > + { > + p = 0; > + } > + else p++; I'm not sure what this is supposed to do. > + case REG_BINARY: > + case REG_NONE: > + pValue = value; > + for (i=0; i<valueSize; i++, pValue++) > + reg_printfW(formatXW, *pValue); > + break; There's no need for separate variable here nor for incrementing pointer.

2 3

Clang warns on "wrong" header guard in "config.h"
by Charles Davis Aug. 2, 2013

Aug. 2, 2013

Hi, I've been building Wine with Clang, and it (apparently, unlike GCC) produces warnings like this: ../../include/config.h:4:9: warning: 'WINE_CROSSTEST' is used as a header guard here, followed by #define of a different macro [-Wheader-guard] #ifndef WINE_CROSSTEST ^~~~~~~~~~~~~~ ../../include/config.h:5:9: note: '__WINE_CONFIG_H' is defined here; did you mean 'WINE_CROSSTEST'? #define __WINE_CONFIG_H ^~~~~~~~~~~~~~~ WINE_CROSSTEST I tracked this down to these three lines in configure.ac: AH_TOP([#ifndef WINE_CROSSTEST #define __WINE_CONFIG_H]) AH_BOTTOM([#endif /* WINE_CROSSTEST */]) These lines generate what Clang thinks is the header guard for the "config.h" header, but we know that it isn't--that header AFAICT has never had an include guard, and __WINE_CONFIG_H is just there to indicate that it's been included. Should we change the __WINE_CONFIG_H macro to be a real header guard instead, or (as I'll admit it has been in the past) is Clang broken here? I'd imagine you guys don't want to rig configure to add -Wno-header-guard to CFLAGS if you can help it. I will point out that lots of these warnings are generated when building with Clang, because (for obvious reasons) this particular header is included all over the place. It makes it harder to see real warnings (like the ones that caused AJ not to commit my last two patches). And please don't tell me to use GCC instead. I suspect that GCC might pick up a similar warning--there's been lots of cross pollination going on between GCC and Clang, so we might have to deal with this sooner or later--and I'd prefer sooner. (I know the commit that added the WINE_CROSSTEST checks is from over 3 years ago, but this warning was only added recently to Clang, which is why I'm only bringing this up now.) Chip

2 1

Re: [PATCH] d3dx9_36: Implement ID3DXEffect_FindNextValidTechnique + add tests.
by Rico Schüller Aug. 2, 2013

Aug. 2, 2013

On 02.08.2013 00:03, Christian Costa wrote: > + technique = (D3DXHANDLE)0xdeadbeef; > + hr = effect->lpVtbl->FindNextValidTechnique(effect, technique1, &technique); > + ok(hr == D3D_OK, "FindNextValidTechnique failed, got %#x, expected %#x\n", hr, D3D_OK); > + ok(technique == technique2, "Technique returned %p, expected %p\n", technique, technique2); > + > + technique = (D3DXHANDLE)0xdeadbeef; > + hr = effect->lpVtbl->FindNextValidTechnique(effect, technique2, &technique); > + ok(hr == S_FALSE, "FindNextValidTechnique, got %#x, expected %#x\n", hr, S_FALSE); Looks fine to me, but I found one minor issue. I'd write "FindNextValidTechnique failed, ..." as in the line above. Cheers Rico

1 0

Re: [11/11] apphelp: Add tests.
by Marvin Aug. 1, 2013

Aug. 1, 2013

Hi, While running your changed tests on Windows, I think I found new failures. Being a bot and all I'm not very good at pattern recognition, so I might be wrong, but could you please double-check? Full results can be found at http://testbot.winehq.org/JobDetails.pl?Key=26579 Your paranoid android. === WINEBUILD (build) === Patch failed to apply

1 0

gdi32: Don't let the gdi32 driver allocate memory in the dplayx region.
by Alexei Svitkine Aug. 1, 2013

Aug. 1, 2013

Don't let the gdi32 driver allocate memory in the dplayx region. Fixes bug: http://bugs.winehq.org/show_bug.cgi?id=34095 The issue is that the dplayx code in wine needs to allocate some memory at a static address (0x50000000) and loading the Mac native gdi driver ends up using that part of the address space. This patch fixes that by reserving that part of memory while the display driver is initialized and releasing it after. --- dlls/gdi32/driver.c | 10 ++++++++++ 1 file changed, 10 insertions(+)

2 5

Re: [PATCH 1/4] d3dx9: Track ID3DXEffect::Begin/End().
by Bruno Jesus Aug. 1, 2013

Aug. 1, 2013

On Thu, Aug 1, 2013 at 9:14 AM, Rico Schüller <kgbricola(a)web.de> wrote: > --- > dlls/d3dx9_36/effect.c | 13 ++++++++++--- > 1 file changed, 10 insertions(+), 3 deletions(-) > > + if (!This->started) > + return D3D_OK; > + > + This->started = FALSE; > + > + return D3D_OK; Sorry to bother but why not simply force false and using a single return? Best wishes, Bruno

2 1

Re: [PATCH 1/11] apphelp: Add definitions.
by Alexandre Julliard Aug. 1, 2013

Aug. 1, 2013

Mislav Blazevic <krofnica996(a)gmail.com> writes: > @@ -25,6 +26,209 @@ extern "C" { > > BOOL WINAPI ApphelpCheckShellObject(REFCLSID, BOOL, ULONGLONG *); > > +typedef enum _PATH_TYPE { > + DOS_PATH, > + NT_PATH > +} PATH_TYPE; > + > +typedef struct _DB { > + HANDLE file; > + DWORD size; > + PBYTE data; > +} DB, *PDB; > + > +typedef DWORD TAGID; > +typedef WORD TAG; > + > +#define TAGID_NULL 0x0 > +#define TAGID_ROOT 0x0 > + > +/* The above definition of TAGID_ROOT is used in winapi but doesn't make > + * sense, especially internally, because TAGID represents offset into > + * database data and there is a header at offset 0, NOT a tag. > + * Therfore, this definition should only be used internally. */ > +#define _TAGID_ROOT 0xC If it's internal it shouldn't be in the public header. In fact I don't see any of these definitions in the SDK header. -- Alexandre Julliard julliard(a)winehq.org

1 0

reg.exe: Add query function
by Hugh McMaster Aug. 1, 2013

Aug. 1, 2013

This patch significantly builds on the query function stub in programs/reg.exe. The patch includes recursion for subkeys. Some examples of usage: wine reg.exe query "HKCU" wine reg.exe query "HKCU\Console" wine reg.exe query "HKCU\Console" /v "ScreenSize" wine reg.exe query "HKCU\Software" /s --- programs/reg/reg.c | 232 +++++++++++++++++++++++++++++++++++++++++++++++++--- 1 file changed, 221 insertions(+), 11 deletions(-)

2 2

Re: [PATCH 2/2] ddraw: Fix double to float conversion warnings
by Henri Verbeet Aug. 1, 2013

Aug. 1, 2013

On 1 August 2013 12:25, Stefan Dösinger <stefan(a)codeweavers.com> wrote: > diff --git a/dlls/ddraw/executebuffer.c b/dlls/ddraw/executebuffer.c > index 2925cfa..5c12272 100644 > --- a/dlls/ddraw/executebuffer.c > +++ b/dlls/ddraw/executebuffer.c > @@ -382,10 +382,10 @@ HRESULT d3d_execute_buffer_execute(struct d3d_execute_buffer *buffer, > dst->u8.tv = src->u8.tv; > > /* Now, the matrix multiplication */ > - dst->u1.sx = (src->u1.x * mat._11) + (src->u2.y * mat._21) + (src->u3.z * mat._31) + (1.0 * mat._41); > - dst->u2.sy = (src->u1.x * mat._12) + (src->u2.y * mat._22) + (src->u3.z * mat._32) + (1.0 * mat._42); > - dst->u3.sz = (src->u1.x * mat._13) + (src->u2.y * mat._23) + (src->u3.z * mat._33) + (1.0 * mat._43); > - dst->u4.rhw = (src->u1.x * mat._14) + (src->u2.y * mat._24) + (src->u3.z * mat._34) + (1.0 * mat._44); > + dst->u1.sx = (src->u1.x * mat._11) + (src->u2.y * mat._21) + (src->u3.z * mat._31) + (1.0f * mat._41); > + dst->u2.sy = (src->u1.x * mat._12) + (src->u2.y * mat._22) + (src->u3.z * mat._32) + (1.0f * mat._42); > + dst->u3.sz = (src->u1.x * mat._13) + (src->u2.y * mat._23) + (src->u3.z * mat._33) + (1.0f * mat._43); > + dst->u4.rhw = (src->u1.x * mat._14) + (src->u2.y * mat._24) + (src->u3.z * mat._34) + (1.0f * mat._44); > I get where the 1.0f comes from, but I think you might as well just remove the multiplication completely. And please get rid of the tabs if you're resending anyway.

1 0