August 2009 - Wine-Devel - list.winehq.org

AppDB isn't working
by Igor Tarasov Aug. 10, 2009

Aug. 10, 2009

AppDB displays only decorations and navigation - no content on all pages. Maybe this is due to recent commits? -- Igor

4 4

D3D shader assembler (round 2)
by Matteo Bruni Aug. 10, 2009

Aug. 10, 2009

I'm sending a new version of the shader assembler, which tries to address the caveats and suggestions previously given by Henri Verbeet (http://www.winehq.org/pipermail/wine-devel/2009-July/077428.html for reference) and has a few other changes, such as making the parser reentrant and matching more faithfully native behavior in handling errors in shader programs. One point of discussion could be the internal transformations done to shaders version 1.x and 2.x, towards a sm 3.0-type representation. This is not really necessary for the assembler (which could naively keep the original instruction opcode and register type), but I think it can be useful looking at a reuse of the bytecode writer in the shader compiler: this way the compiler could be somewhat simplified, having to deal just with sm 3.0 registers and conventions, while it would be writer's responsibility to convert to the destination format (e.g. transforming some mov to texcoord in <= ps 1.3). Anyway, if this is not considered useful, it could be removed. As usual, reviews, comments and suggestions are much appreciated. Regards, Matteo Bruni.

3 2

backward binary compatibility of winelib
by Андрей Пономаренко Aug. 9, 2009

Aug. 9, 2009

Hello, I'm from Institute for System Programming of Russian Academy of Sciences and we are developing a free lightweight tool for checking backward/forward binary compatibility of shared C/C++ libraries in OS Linux. It checks interface signatures and data type definitions in two library versions (headers and shared objects) and searches ABI changes that may lead to incompatibility. We have released 1.2 version of this tool and we'd like you to consider its usefulness for your project. The wiki-page with the latest release of binary compatibility checker is http://ispras.linux-foundation.org/index.php/ABI_compliance_checker Andrey Ponomarenko

2 1

re: cppcheck run against 1.27
by Dan Kegel Aug. 8, 2009

Aug. 8, 2009

Chris wrote: >... >[../wine-git/tools/widl/write_msft.c:2540]: (error) Deallocating a >deallocated pointer: fd - This just means your freeing an already freed >resource... This is more of a warning than anything Double frees are serious errors... double closes not so much. I don't see either on that line, though...? - Dan

1 1

.net 3.5 installer happy already?!
by Dan Kegel Aug. 8, 2009

Aug. 8, 2009

After the recent success with the .net 3.0 installer, I decided to try .net 3.5: $ wget http://download.microsoft.com/download/7/0/3/703455ee-a747-4cc8-bd3e-98a615… $ sha1sum dotNetFx35setup.exe b0ca05c12ebb9a3610206bad7f219e02b7873cbd dotNetFx35setup.exe $ rm -rf ~/.wine $ sh winetricks dotnet20 $ wine dotNetFx35setup.exe And surprise, although there was a crash dialog about printfilterpipelinesvc, the installer reported that it finished ok! I have no idea if the installed runtime actually works, and there are a ton of log messages after installation finishes. (Presumably the .net optimization service?) But it's nice to have gotten this far. - Dan p.s. A few of the common warnings at random: Could not load type 'System.DateTimeOffset' from assembly 'mscorlib, Version=2.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089'. while compiling method token 0x6000723 fixme:shell:URL_ParseUrl failed to parse L"System.ServiceModel" Could not load file or assembly 'System.ServiceModel, Version=3.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089' or one of its dependencies. Exception from HRESULT: 0x80070002 while compiling method token 0x60001bc Error compiling System.ServiceModel.Web, Version=3.5.0.0, Culture=neutral, PublicKeyToken=31bf3856ad364e35: Could not find or load a type. (Exception from HRESULT: 0x80131522) fixme:shell:URL_ParseUrl failed to parse L"System.ServiceModel" Warning: System.IO.FileNotFoundException: Could not load file or assembly 'System.ServiceModel, Version=3.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089' or one of its dependencies. Exception from HRESULT: 0x80070002 File name: 'System.ServiceModel, Version=3.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089' WRN: Assembly binding logging is turned OFF. To enable assembly bind failure logging, set the registry value [HKLM\Software\Microsoft\Fusion!EnableLog] (DWORD) to 1. Note: There is some performance penalty associated with assembly bind failure logging. To turn this feature off, remove the registry value [HKLM\Software\Microsoft\Fusion!EnableLog].

1 0

Re: wine-devel Digest, Vol 49, Issue 21
by chris ahrendt Aug. 8, 2009

Aug. 8, 2009

> > Subject: > re: cppcheck run against 1.27 > From: > Dan Kegel <dank(a)kegel.com> > Date: > Sat, 8 Aug 2009 15:53:25 -0700 > To: > wine-devel(a)winehq.org > > To: > wine-devel(a)winehq.org > > > Chris wrote: > >> ... >> [../wine-git/tools/widl/write_msft.c:2540]: (error) Deallocating a >> deallocated pointer: fd - This just means your freeing an already freed >> resource... This is more of a warning than anything >> > > Double frees are serious errors... double closes not so much. > I don't see either on that line, though...? > - Dan > I can run it with the --all option to see if it gives any clarification Chris

1 0

Building inkscape for Windows using mingw under wine
by Dan Kegel Aug. 8, 2009

Aug. 8, 2009

http://bugs.winehq.org/show_bug.cgi?id=19615 shows that wine is very nearly in good enough shape to run the fetch-all-the-libs-and-sources script with cygwin, and run the inkscape win32 build system, and pop out a working inkscape.exe. We're just one little leak away from full success, I think. If you ask me, that's pretty cool. - Dan

1 2

dotnet30 and winetricks
by Dan Kegel Aug. 8, 2009

Aug. 8, 2009

Thanks to AF and Hans (and Codeweavers), there's now a short recipe for installing .net 30. I've added it to winetricks. Give it a shot and let me know if it works for you...

3 3

cppcheck run against 1.27
by chris ahrendt Aug. 8, 2009

Aug. 8, 2009

[../wine-git/dlls/msvcrt/tests/file.c:997]: (error) Resource leak: stream1 [../wine-git/dlls/msvcrt/tests/file.c:997]: (error) Resource leak: stream4 [../wine-git/dlls/msvcrt/tests/heap.c:424]: (error) Deallocating a deallocated pointer: mem [../wine-git/dlls/ntdll/server.c:802]: (error) Resource leak: fd [../wine-git/tools/fnt2bdf.c:219]: (error) Resource leak: fp [../wine-git/tools/fnt2bdf.c:653]: (error) Resource leak: fd [../wine-git/tools/fnt2fon.c:304]: (error) Memory leak: file_lens [../wine-git/tools/widl/write_msft.c:2540]: (error) Deallocating a deallocated pointer: fd [../wine-git/tools/winebuild/res16.c:187]: (error) Resource leak: fd [../wine-git/tools/winedump/pe.c:1549]: (error) Memory leak: map Latest Check against the 1.27 Tree Chris

3 2

Re: cppcheck run against 1.27
by chris ahrendt Aug. 8, 2009

Aug. 8, 2009

Marcus Meissner wrote: > On Fri, Aug 07, 2009 at 06:36:36PM -0700, chris ahrendt wrote: > >> [../wine-git/dlls/ntdll/server.c:802]: (error) Resource leak: fd >> > > I do not fully understand the code (keep fd open to have to lock > existing?), but it is in a fatal exit path. > >> [../wine-git/tools/fnt2bdf.c:219]: (error) Resource leak: fp >> [../wine-git/tools/fnt2bdf.c:653]: (error) Resource leak: fd >> > > Fix submitted, seem necessary. > > >> [../wine-git/tools/fnt2fon.c:304]: (error) Memory leak: file_lens >> > > Fix submitted, but program is shortlived. > > >> [../wine-git/tools/widl/write_msft.c:2540]: (error) Deallocating a >> deallocated pointer: fd >> > > I do not understand what it mislikes about it. I think the detection > is wrong. > > >> [../wine-git/tools/winebuild/res16.c:187]: (error) Resource leak: fd >> > > Fix submitted. Small issue though. > > >> [../wine-git/tools/winedump/pe.c:1549]: (error) Memory leak: map >> > > Fix submitted. Small issue though. > > Ciao, Marcus > > > here is what its checking for : Standard checks that are always used * out of bounds * Using 'memfunc' on class * Using 'memfunc' on struct that contains a 'std::classname' * Class Base which is inherited by class Derived does not have a virtual destructor * Mismatching allocation and deallocation: varname * Memory leak <http://en.wikipedia.org/wiki/Memory_leak>: varname * Resource leak <http://en.wikipedia.org/wiki/Resource_leak>: varname * Deallocating a deallocated pointer: varname * Using 'varname' after it is deallocated / released * The given size sz is mismatching * Invalid radix in call to strtol <http://en.wikipedia.org/wiki/Strtol> or strtoul. Must be 0 or 2-36 * Overlapping data buffer varname * Unsigned division. The result will be wrong. * Unusual pointer arithmetic * Returning pointer to local array variable * Same iterator is used with both container1 and container2 * Dangerous usage of erase * After push_back or push_front, the iterator 'iterator_name' may be invalid * Wrong assignment of an auto-variable to an effective parameter of a function * Return of the address of an auto-variable * Division by zero * STL <http://en.wikipedia.org/wiki/Standard_Template_Library>: check usage of iterators after erase, as erase invalidates the iterator * STL: check usage of iterators and pointers after push_back and push_front. If using a vector the iterator or pointer may become invalid. * STL: Range checks with iterators should use != instead of < I have not run it with the following : Extra checks that you enable with "--all" These checks are not part of the standard checking because they produce false positives. * Array index out of bounds * Buffer overrun * Dangerous usage of strncat <http://en.wikipedia.org/wiki/Strcat>, possible buffer overrun * Memory leak: varname * The size argument is given as a char constant Extra checks that you enable with "--style" * The class 'classname' has no constructor * Member variable not initialized in the constructor 'classname::varname' * Unused private function 'classname::funcname' * 'operator=' should return something * C-style pointer casting * Redundant condition. It is safe to deallocate a NULL pointer * Redundant condition. The remove function in the STL will not do anything if element doesn't exist * Found redundant if condition - 'if (condition);' * struct or union member 'structname::varname' is never used * Function parameter 'parname' is passed by value. It could be passed by reference instead. * Redundant code: Found a statement that begins with type constant * Warning - using char variable as array index * Warning - using char variable in bit operation * Condition is always truefalse Checks that are enabled when both "--all" and "--style" is given * Warning: Division with signed and unsigned operators When giving "--unused-functions" * The function 'funcname' is never used I am going to start running with the --all flag going forward. So the leak with fd cppcheck thinks that you have an exit path that will leave the fd allocated still without freeing the resource. [../wine-git/dlls/ntdll/server.c:802]: (error) Resource leak: fd so this exits without freeing the resource. [../wine-git/tools/widl/write_msft.c:2540]: (error) Deallocating a deallocated pointer: fd - This just means your freeing an already freed resource... This is more of a warning than anything Chris

1 0