September 2012 - Wine-Devel

Re: d3dx9_36 [patch 1/2]: Implement D3DXSHEvalDirectionalLight
by Rico Schüller Sept. 14, 2012

Sept. 14, 2012

On 14.09.2012 11:07, Nozomi Kodama wrote: > Hello > > one remark about this patch. > > Test in real windows shows that D3DXSHEvalDirectionalLight accepts order > > D3DXSH_MAXORDER or order < D3DXSH_MINORDER. > But in these cases, the colour outputs are unpredictable. For the same > calling arguments, D3DSXHEvalDirectionalLight returns different colours. > > That explains the simplicity of the implementation of this function. > > Nozomi. > + FLOAT s, temp[max( 1, order * order )]; Is this allowed in c and does it really work the way one would expect? + for (order = D3DXSH_MINORDER; order <= D3DXSH_MAXORDER; order++) ... + if ( ( order == 0 ) || ( j >= order * order ) ) Why would you check for order == 0? That's imho never the case that this is true. At least if you only test from D3DXSH_MINORDER onward. + unsigned int j, order, start[] = {0, 4, 13, 29, 54}; Using the start[] looks a bit ugly ... why don't you calculate the index? + expected = table[start[order - 2] + j]; ! You've again trailing white spaces. What happens in the following test case? Which result is returned? D3DXSHEvalDirectionalLight(order, &dir, 1.7f, 2.6f, 3.5f, rout, rout, rout); Well the implementation seems to be correct, but it may be nice to test for this as well. Maybe in another patch. Cheers Rico

2 1

Re: user32/tests: Add test for SetFocus() (with few todo)(Try 2).
by Sergey Guralnik Sept. 14, 2012

Sept. 14, 2012

I've sent next patch more than month ago, and have no comments about it. From 57adc6991431cd765dcdb97082263db834e4b533 Mon Sep 17 00:00:00 2001 From: Sergey Guralnik <serhio(a)etersoft.ru> Date: Thu, 9 Aug 2012 17:12:56 +0400 Subject: user32/tests: Add test for SetFocus() (with few todo)(Try 2). Is anything wrong with this one? -- Sergey

2 1

Re: [PATCH] advapi32: Eliminate dead stores (Clang).
by Alexandre Julliard Sept. 14, 2012

Sept. 14, 2012

Charles Davis <cdavis5x(a)gmail.com> writes: > @@ -615,9 +613,9 @@ static DWORD mac_write_credential(const CREDENTIALW *credential, BOOL preserve_b > attr_list.count = 0; > attr_list.attr = NULL; > } > - status = SecKeychainItemModifyAttributesAndData(keychain_item, &attr_list, > - preserve_blob ? 0 : strlen(password), > - preserve_blob ? NULL : password); > + SecKeychainItemModifyAttributesAndData(keychain_item, &attr_list, > + preserve_blob ? 0 : strlen(password), > + preserve_blob ? NULL : password); I expect you'd want to handle that error instead. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [4/4] windowscodecs: Implement Graphic Control Extension metadata reader.
by Marvin Sept. 14, 2012

Sept. 14, 2012

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=21585 Your paranoid android. === WINEBUILD (build) === Patch failed to apply

1 0

Re: [3/4] windowscodecs: Implement Image Descriptor metadata reader.
by Marvin Sept. 14, 2012

Sept. 14, 2012

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=21584 Your paranoid android. === WINEBUILD (build) === Patch failed to apply

1 0

Re: vcomp: better stub for _vcomp_fork
by Alexandre Julliard Sept. 13, 2012

Sept. 13, 2012

Dan Kegel <dank(a)kegel.com> writes: > @@ -41,6 +42,80 @@ double CDECL omp_get_wtime(void) > return GetTickCount() / 1000.0; > } > > +/* Call helper with the ncount args after ncount */ > +extern void WINAPIV _vcomp_fork(DWORD parallel, int ncount, void *helper, ...); > +#ifdef __i386__ > +__ASM_GLOBAL_FUNC( _vcomp_fork, > + "push %ebp\n\t" > + "movl %esp,%ebp\n\t" > + > + /* Push varargs */ > + "movl 12(%ebp),%ecx\n\t" /* if (nargs > 0) */ > + "test %ecx,%ecx\n\t" > + "jz copyend\n\t" > + "leal 20(%ebp),%edx\n\t" /* dx -> first vararg */ > + "leal (%edx,%ecx,4),%edx\n\t" /* dx += nargs * sizeof arg */ > + "copyloop:\n\t" /* while (nargs-- > 0) */ > + "sub $4,%edx\n\t" > + "pushl (%edx)\n\t" /* push next vararg */ > + "dec %ecx\n\t" > + "jnz copyloop\n\t" > + "copyend:\n\t" > + > + /* Call helper and return */ > + "call *16(%ebp)\n\t" > + "movl %ebp,%esp\n\t" > + "pop %ebp\n\t" > + "ret\n\t" > + ) > +#endif There are several places that already implement that sort of thing the right way, you should copy one of them instead of reinventing the wheel. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: d3dx9_36 [patch 1/2: try 2]: Implement D3DXSHEvalDirectionalLight
by Alexandre Julliard Sept. 13, 2012

Sept. 13, 2012

Nozomi Kodama <nozomi.kodama(a)yahoo.com> writes: > @@ -2231,6 +2231,40 @@ FLOAT* WINAPI D3DXSHEvalDirection(FLOAT *out, UINT order, CONST D3DXVECTOR3 *dir > return out; > } > > +HRESULT WINAPI D3DXSHEvalDirectionalLight(UINT order, CONST D3DXVECTOR3 *dir, FLOAT Rintensity, FLOAT Gintensity, FLOAT Bintensity, FLOAT *Rout, FLOAT *Gout, FLOAT *Bout) > +{ > + FLOAT s, temp[36]; > + UINT j; > + > + TRACE("Order %u, Vector %p, Red %f, Green %f, Blue %f, Rout %p, Gout %p, Bout %p\n", order, dir, Rintensity, Gintensity, Bintensity, Rout, Gout, Bout); > + > + if ( order > D3DXSH_MAXORDER ) > + { > + FIXME("This case is not implemented yet. Anyway, following to MSDN, this case should not happen\n"); > + > + return D3D_OK; > + } You should test the expected results for that case too, the implementation should be easy either way. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [PATCH 4/4] ntdll: Add more complete implementation of NtPowerInformation
by Alexandre Julliard Sept. 13, 2012

Sept. 13, 2012

James Eder <jimportal(a)gmail.com> writes: > @@ -2261,7 +2242,19 @@ NTSTATUS WINAPI NtInitiatePowerAction( > SystemAction,MinSystemState,Flags,Asynchronous); > return STATUS_NOT_IMPLEMENTED; > } > - > + > +#ifdef linux > +static int have_cpufreq(void) > +{ > + FILE* f = fopen("/sys/devices/system/cpu/cpu0/cpufreq/scaling_cur_freq", "r"); > + if(f) > + { > + fclose(f); > + return 1; > + } > + return 0; > +} > +#endif You don't want to open a file just to see if you will be able to open it for real later on. You'll need to check the subsequent open anyway, so you can do the fallback at that point. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [PATCH 1/2] jscript: Don't leave current scope in return statement before evaluating the expression
by Marvin Sept. 13, 2012

Sept. 13, 2012

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=21565 Your paranoid android. === WNT4WSSP6 (32 bit) === No test summary line found === W2KPROSP4 (32 bit) === No test summary line found === WXPPROSP3 (32 bit) === No test summary line found === W2K3R2SESP2 (32 bit) === No test summary line found === WVISTAADM (32 bit) === No test summary line found === W2K8SE (32 bit) === No test summary line found === W7PRO (32 bit) === No test summary line found === W7PROX64 (32 bit) === No test summary line found === TEST64_W7SP1 (32 bit) === No test summary line found === W7PROX64 (64 bit) === No test summary line found === TEST64_W7SP1 (64 bit) === No test summary line found

1 0

Re: [PATCH 5/6] mshtml: Added IHTMLDocument4::fireEvent implementation
by Marvin Sept. 13, 2012

Sept. 13, 2012

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=21559 Your paranoid android. === W2KPROSP4 (32 bit) === No test summary line found === WXPPROSP3 (32 bit) === No test summary line found === W2K3R2SESP2 (32 bit) === No test summary line found === WVISTAADM (32 bit) === No test summary line found === W2K8SE (32 bit) === No test summary line found === W7PRO (32 bit) === No test summary line found === W7PROX64 (32 bit) === No test summary line found === TEST64_W7SP1 (32 bit) === No test summary line found === W7PROX64 (64 bit) === No test summary line found === TEST64_W7SP1 (64 bit) === No test summary line found

1 0

Translations of the About Pages
by Kyle Auble Sept. 13, 2012

Sept. 13, 2012

I just submitted a draft rewrite of WineHQ's About page in English that integrates some things from the wiki page, but there are still Hebrew and Spanish about pages on the wiki. I started working on translating the Spanish one, but realized someone else should probably do it after one paragraph, and my Hebrew is even weaker. I think the goal is still to move all central pages to WineHQ so I'd like to remove those versions from the wiki soon too. If any Hebrew or Spanish-speaking contributors would prefer to work from the wiki content though, I can leave it up for a while. If you would rather translate from the new English version to keep the different translations in sync though, I'll go ahead and clean up the wiki some more. - Kyle

1 0

Re: [2/2] windowscodecs: Create Graphic Control Extension metadata block for a GIF frame.
by Marvin Sept. 12, 2012

Sept. 12, 2012

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=21552 Your paranoid android. === WINEBUILD (build) === Patch failed to apply

1 0

Re: [1/2] windowscodecs: Create Image Descriptor metadata block for a GIF frame.
by Vincent Povirk Sept. 12, 2012

Sept. 12, 2012

> + stream = create_stream(&IMD_data, sizeof(IMD_data)); Er, where is create_stream defined? I don't see it in git or a previous patch.

2 1

dssenh: implementing the dssenh cryptographic service provider
by Marek Chmiel Sept. 12, 2012

Sept. 12, 2012

Hello all, over the summer I had a chance to complete the regression test suite for the DSSENH cryptographic provider; thank you to those that helped along the way. I am currently researching how to implement the actual provider, and any suggestions would be greatly appreciated. A question I have is: RSAENH AND DSSENH both use the same encryption algorithms (I am not 100% sure about that yet), so for example could I use one that was written for RSAENH or shall I write those functions on my own using Libtomcrypt? Thank you. --Marek

2 1

Re: ddraw: Don't print a fixme for the flags in DirectDrawEnumerateExA; they don't change the outcome
by Michael Stefaniuc Sept. 12, 2012

Sept. 12, 2012

After talking to Henri on irc, please ignore this patch. There are patches out there to enumerate more devices than the primary one and then the flags do play a role. On 09/12/2012 11:08 PM, Michael Stefaniuc wrote: > --- > dlls/ddraw/main.c | 5 +---- > 1 files changed, 1 insertions(+), 4 deletions(-) > > diff --git a/dlls/ddraw/main.c b/dlls/ddraw/main.c > index 61e8ec9..b79b12e 100644 > --- a/dlls/ddraw/main.c > +++ b/dlls/ddraw/main.c > @@ -377,12 +377,9 @@ HRESULT WINAPI DirectDrawEnumerateExA(LPDDENUMCALLBACKEXA Callback, void *Contex > DDENUM_NONDISPLAYDEVICES)) > return DDERR_INVALIDPARAMS; > > - if (Flags) > - FIXME("flags 0x%08x not handled\n", Flags); > - > TRACE("Enumerating default DirectDraw HAL interface\n"); > > - /* We only have one driver by now */ > + /* We have only one driver (primary) which is always enumerated regardless of the flags */ > __TRY > { > static CHAR driver_desc[] = "DirectDraw HAL", > bye michael

1 0

Re: [PATCH 4/4] ntdll: Add more complete implementation of NtPowerInformation
by Ken Thomases Sept. 12, 2012

Sept. 12, 2012

On Sep 12, 2012, at 2:17 PM, James Eder wrote: > - valSize = sizeof(longVal); > - if (!sysctlbyname("hw.cpufrequency", &longVal, &valSize, NULL, 0)) > - cpuHz = longVal; I haven't tried building, but this would still leave the warning about longVal being declared but never used, right? I don't see you removing that declaration. Regards, Ken

1 0

Re: [2/2] windowscodecs: Add a bunch of GIF metadata tests.
by Marvin Sept. 12, 2012

Sept. 12, 2012

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=21528 Your paranoid android. === WINEBUILD (build) === Patch failed to apply

1 0

Re: msls31: Add stub dll.
by Austin English Sept. 11, 2012

Sept. 11, 2012

On Tue, Sep 11, 2012 at 9:32 AM, Josh DuBois <duboisj(a)codeweavers.com> wrote: > Presence of a stub of msls31.dll allows Visio 2003 to open .svg files, which > it fails to do without the stub. > Visio 2003 does complain that some values in the files are missing or of the > incorrect data type when opening these files, but it will nonetheless open > them and can edit them successfully when the stubbed .dll is present. @@ -0,0 +1,79 @@ +70 stub LssbFDoneDisplay +65 stub LsCompressSubline +1 stub LsCreateContext +3 stub LsCreateLine ..etc. Does it really need them by ordinal? The spec files are typically sorted alphabetically by function name, e.g.,: +@ stub LsCompressSubline +@ stub LsCreateContext +@ stub LsCreateLine +@ stub LssbFDoneDisplay ..etc. Some other minor things: A) you don't need to include the changes to configure, just configure.ac. B) Your comments are formatted a bit strange /* --------------------------------------------------------------- ** msls31.dll ** Copyright CodeWeavers, Inc. ** Created: 9/10/2012 by Josh DuBois for CodeWeavers. most files have: /* * msls31.dll * * Copyright 2012 Josh DuBois for CodeWeavers, Inc. * */ similar for your comment about preferring the native version. Welcome to Wine! -- -Austin

2 1

Re: [PATCH 5/5] ntdll: Add more complete implementation of NtPowerInformation
by James Eder Sept. 11, 2012

Sept. 11, 2012

I've sent a patch set that I need some help testing. The first four in the series should be fine as they don't change anything specific to a particular platform. The last one however makes some changes which are platform specific (affects Linux, Apple, and various BSDs). I've tested the Linux code but I do not have an Apple, for example, to test that code. While the code looks good to me and is is largely based off of code removed from fill_cpu_info(), it is still worth some concern that it hasn't seen any real testing. If you have a Mac or BSD, I would be grateful if you could try the patches out to make sure things are working as expected. You'll find a few more details in the commit message http://source.winehq.org/patches/data/89698 . Also of concern is that the testbot seems to have only one processor. This results in most of the tests added in http://source.winehq.org/patches/data/89697 being skipped. So testbot doesn't seem that helpful for this test. As for the implementation of the test itself, would it be better to use broken() and just test Wine for the Win7 behavior? From my testing, the test as is works for XP, Win7, and Wine with my patch set. If we can write the test to work everywhere, don't use broken()? There are more details in the patch itself so please have a look and let me know what you think. -- Jim

2 2

Re: wininet: disable TLSv1.1/1.2 by default
by Hiroshi Miura Sept. 11, 2012

Sept. 11, 2012

Hi, Are there any comment for it? It is a motivation that Evernote windows client cannot connect its server on Wine/OpenSSL 1.0.1. it is caused by server is only support TLS1.0/SSL3.0 , apache2/openssl 0.9.8 thing, but client ask TLS1.1/1.2 and then server returns that session is fails. I know wininet(in Microsoft implementation) has a behavior to re-try by TLS1.0 after TLS1.2 fails. It is discussed in IETF TLS maillist http://www.ietf.org/mail-archive/web/tls/current/msg08218.html How do you think whether we should implement similar way or not? If we adopt similar fall back mechanism, compatibility may become well, but is it good behavior? Hiroshi On 2012.08.26 11:51, Hiroshi Miura wrote: > > OpenSSL 1.0.x now support TLSv1.1 and TLSv1.2 > but Windows7 is disabled by default. > > Schannel registry indicate TLSv1.1/1.2 config. > > It fixes the error of Evernote client connection. > http://bugs.winehq.org/show_bug.cgi?id=30598 > > Signed-off-by: Hiroshi Miura <miurahr(a)linux.com> > --- > dlls/wininet/netconnection.c | 68 ++++++++++++++++++++++++++++++++++++++++++ > 1 file changed, 68 insertions(+) > > > > > >

2 1

Re: [PATCH] secur32/schannel: diabled TLS1.1/1.2 by Default
by Marcus Meissner Sept. 10, 2012

Sept. 10, 2012

On Mon, Sep 10, 2012 at 12:14:51AM +0900, Hiroshi Miura wrote: > -Set TLS1.1/1.2 disabled by Default that is > same as Windows 7 default. > > See registry entry for schannel and control > enable/disable tls versions. > > It also see grbitEnabledProtocols defined in > credentials that take precedence over registry. I think the regression issue with TLS 1.1/1.2 is the "empty fragments" sending, right? Perhaps we can just disable that and not all of TLS 1.1/1.2? Ciao, Marcus

2 1

Re: [PATCH 3/5] gdi32: Fixed localized font style name.
by Qian Hong Sept. 10, 2012

Sept. 10, 2012

Hello, On Tue, Sep 11, 2012 at 1:05 AM, Qian Hong <fracting(a)gmail.com> wrote: > and the symbol font bug (with todo_wine because it need to be fixed in [PATCH 4/5].) Sorry for a typo, should be: with todo_wine because it need to be fixed in [PATCH 5/5]. Thanks! -- Regards, Qian Hong - Sent from Ubuntu http://www.ubuntu.com/

1 0

Re: [PATCH 2/3] jscript: Added support for no new line between continue and identifier rule
by Marvin Sept. 10, 2012

Sept. 10, 2012

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=21489 Your paranoid android. === WNT4WSSP6 (32 bit) === No test summary line found === W2KPROSP4 (32 bit) === No test summary line found === WXPPROSP3 (32 bit) === No test summary line found === W2K3R2SESP2 (32 bit) === No test summary line found === WVISTAADM (32 bit) === No test summary line found === W2K8SE (32 bit) === No test summary line found === W7PRO (32 bit) === No test summary line found === W7PROX64 (32 bit) === No test summary line found === TEST64_W7SP1 (32 bit) === No test summary line found === W7PROX64 (64 bit) === No test summary line found === TEST64_W7SP1 (64 bit) === No test summary line found

1 0

Re: [PATCH 1/3] jscript: Added support for no new line between break and identifier rule
by Marvin Sept. 10, 2012

Sept. 10, 2012

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=21488 Your paranoid android. === WNT4WSSP6 (32 bit) === No test summary line found === W2KPROSP4 (32 bit) === No test summary line found === WXPPROSP3 (32 bit) === No test summary line found === W2K3R2SESP2 (32 bit) === No test summary line found === WVISTAADM (32 bit) === No test summary line found === W2K8SE (32 bit) === No test summary line found === W7PRO (32 bit) === No test summary line found === W7PROX64 (32 bit) === No test summary line found === TEST64_W7SP1 (32 bit) === No test summary line found === W7PROX64 (64 bit) === No test summary line found === TEST64_W7SP1 (64 bit) === No test summary line found

1 0

Re: mshtml: Added support for navigating to anchors with IDs containing '#'
by Alexandre Julliard Sept. 10, 2012

Sept. 10, 2012

Jacek Caban <jacek(a)codeweavers.com> writes: > --- > dlls/mshtml/navigate.c | 42 +++++++++++++++++++++++++++++++++++++++--- > dlls/mshtml/nsiface.idl | 11 +++++++++++ > 2 files changed, 50 insertions(+), 3 deletions(-) It doesn't work here: ../../../tools/runtest -q -P wine -M mshtml.dll -T ../../.. -p mshtml_test.exe.so htmldoc.c && touch htmldoc.ok htmldoc.c:4650: Test failed: unexpected riid {5C8B21BC-EF6E-4599-A26F-FACC05B4ADBE} make: *** [htmldoc.ok] Error 1 -- Alexandre Julliard julliard(a)winehq.org

2 1