November 2007 - Wine-Devel - list.winehq.org

Re: USER32: Patch 3 of 6: DlgDirList converts path specification to uppercase.
by Alexandre Julliard Nov. 14, 2007

Nov. 14, 2007

Alex Villacís Lasso <a_villacis(a)palosanto.com> writes: > Observed while preparing conformance tests. The filepath (if valid) gets > converted to uppercase on return. > > Changelog: > * DlgDirList must convert path specification to uppercase. You should do that after the SetCurrentDirectory, uppercasing it first means we are much more likely to have to do a slow case-insensitive search on the whole path. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [4/4] msxml: Test for removeAttribute
by Robert Shearman Nov. 14, 2007

Nov. 14, 2007

Alistair Leslie-Hughes wrote: > @@ -589,6 +589,10 @@ static void test_domnode( void ) > ok( !lstrcmpW(V_BSTR(&var), szstr1), "wrong attr value\n"); > VariantClear( &var ); > SysFreeString( str ); > + > + > + str = SysAllocString( szvr ); > + SysFreeString( str ); > > r = IXMLDOMElement_get_attributes( element, &map ); > ok( r == S_OK, "get_attributes returned wrong code\n"); > @@ -1077,7 +1081,7 @@ static void test_create(void) > ok( r == S_OK, "returns %08x\n", r ); > r = IXMLDOMNamedNodeMap_get_length( attr_map, &num ); > ok( r == S_OK, "returns %08x\n", r ); > - ok( num == 1, "num %ld\n", num ); > + ok( num == 1, "num %ld\n", num ); > IXMLDOMNamedNodeMap_Release( attr_map ); > VariantClear(&var); > r = IXMLDOMElement_getAttribute( element, name, &var ); > These two changes don't appear to be relevant to the test you're adding. -- Rob Shearman

1 0

Re: msxml: Implement cloneNode and correct insertBefore
by James Hawkins Nov. 14, 2007

Nov. 14, 2007

On 11/13/07, Alistair Leslie-Hughes <leslie_alistair(a)hotmail.com> wrote: > > Hi James, > Thanks for the pointers. > > I resend the patches shortly. Is there an easy way to split a Diff up? via > git? More to the point changes to a single file (without have to backup it > up every change) > > I just get frustrated with try to keep lots of diffs around. > mailing list etiquette: make sure to bottom-post on this ML and CC wine-devel so others can share in the information exchange. To answer your question, you have to get into the git mindset. That is, you make one fix, commit the fix, write the next fix, etc. It's really a development change. You should know what to fix before you go in to fix it, not hack around until an app works. -- James Hawkins

3 2

Re: msxml: Implement Element removeAttribute
by Steven Edwards Nov. 14, 2007

Nov. 14, 2007

On Nov 12, 2007 3:57 AM, Alistair Leslie-Hughes <leslie_alistair(a)hotmail.com> wrote: > > + > + TRACE("(%p)->%s\n", iface, debugstr_w(p)); > + // TODO Replace if Attribute has a Default value. Also another thing. No C++ style comments in Wine source code. Thanks Steven -- Steven Edwards "There is one thing stronger than all the armies in the world, and that is an idea whose time has come." - Victor Hugo

1 0

Re: msxml: Implement cloneNode and correct insertBefore
by James Hawkins Nov. 14, 2007

Nov. 14, 2007

On 11/12/07, Alistair Leslie-Hughes <leslie_alistair(a)hotmail.com> wrote: > Hi, > > Changelog: > InsertBefore is allowed to have outNewChild as NULL > Implement cloneNode > One fix per patch please. Also, please add tests for these fixes. > Best Regards > Alistair Leslie-Hughes > > diff --git a/dlls/msxml3/node.c b/dlls/msxml3/node.c > index 61d5ab2..86d7614 100644 > --- a/dlls/msxml3/node.c > +++ b/dlls/msxml3/node.c > @@ -373,7 +373,7 @@ static HRESULT WINAPI xmlnode_insertBefore( > > TRACE("(%p)->(%p,var,%p)\n",This,newChild,outNewChild); > > - if (!(newChild && outNewChild)) > + if (!newChild) > return E_INVALIDARG; > > switch(V_VT(&refChild)) > @@ -412,9 +412,13 @@ static HRESULT WINAPI xmlnode_insertBefore( > xmlAddChild(This->node, new_child_node); > } > > - IXMLDOMNode_Release(new); > IXMLDOMNode_AddRef(newChild); > - *outNewChild = newChild; > + IXMLDOMNode_Release(new); > + if(outNewChild) > + { > + *outNewChild = newChild; > + } > + > TRACE("ret S_OK\n"); > return S_OK; > } > @@ -517,8 +521,36 @@ static HRESULT WINAPI xmlnode_cloneNode( > VARIANT_BOOL deep, > IXMLDOMNode** cloneRoot) > { > - FIXME("\n"); > - return E_NOTIMPL; > + xmlnode *This = impl_from_IXMLDOMNode( iface ); > + xmlNodePtr pClone = NULL; > + IXMLDOMNode *pNode = NULL; > + > + TRACE("%p (%d)\n", This, deep); > + > + if(!cloneRoot) > + return E_INVALIDARG; > + > + pClone = xmlCopyNode(This->node, deep ? 1 : 0); > + if(pClone) > + { > + pClone->doc = This->node->doc; > + > + pNode = create_node(pClone); > + if(!pNode) > + { > + ERR("Copy failed\n"); > + return E_FAIL; > + } > + > + *cloneRoot = pNode; > + } > + else > + { > + ERR("Copy failed\n"); > + return E_FAIL; > + } > + > + return S_OK; > } > > static HRESULT WINAPI xmlnode_get_nodeTypeString( > @@ -832,7 +864,7 @@ static ULONG WINAPI Internal_Release( > if ( ref == 0 ) > { > if( This->node ) > - xmldoc_release( This->node->doc ); > + xmldoc_release( This->node->doc ); > Please don't make random whitespace changes. -- James Hawkins

1 0

svchost Patch Question
by Roy Shea Nov. 13, 2007

Nov. 13, 2007

Howdy, I'm looking for advice on how to clean up the implementation of svchost. The latest version is available at: http://www.winehq.org/pipermail/wine-patches/2007-November/046467.html If the stopper for this patch is the lack of a DLL using it, I can resend svchost later this week when I have a stub implementation of BITS using svchost ready. Thanks for any pointers! -Roy

2 1

xrun handling in winealsa.drv
by Timofei V. Bondarenko Nov. 13, 2007

Nov. 13, 2007

Hi. I've noticed the winealsa does not handle alsa XRUNs. In the case of XRUN application hangs. That's happened because return value of snd_pcm_avail_update() is not checked. For example: wavein.c 175 DWORD bytes; 180 /* read all the fragments accumulated so far */ 181 frames = snd_pcm_avail_update(wwi->pcm); 182 bytes = snd_pcm_frames_to_bytes(wwi->pcm, frames); waveout.c 476 DWORD availInQ; 479 availInQ = snd_pcm_avail_update(wwo->pcm); 482 if (availInQ > 0) { The snd_pcm_avail_update() returns a _signed_ type and it really does return negative number -EPIPE in the case of XRUN. More over, any return value of less than period_size may indicate XRUN state too. The snd_pcm_state() may help then. Th dsoutput.c and dscapture.c do exactly the same wrong thigs. BTW, using DWORD type causes problems. Following fragment is from wavein.c too: 177 DWORD read; 181 frames = snd_pcm_avail_update(wwi->pcm); 182 bytes = snd_pcm_frames_to_bytes(wwi->pcm, frames); 193 read = wwi->read(wwi->pcm, lpWaveHdr->lpData + 194 bytesRead = snd_pcm_frames_to_bytes(wwi->pcm, 197 if (bytesRead != (DWORD) -1) This is obviously will not work as expected. I hope someone will fix these issues :) Regards, Tim.

3 5

Grumble. Lost the Valgrind magic.
by Dan Kegel Nov. 13, 2007

Nov. 13, 2007

I had to reinstall my main valgrind testing machine (I foolishly ran the vmware server install script, and it nuked my wireless), and since then, valgrind isn't giving me stack traces. I guess something's wrong in my recipe, but I'm not sure what, the vg-3.2.3 patch is being applied. Sigh. Is anybody else having luck building a valgrind that gives good stack traces for wine?

2 1

Re: rpcrt4: Implement RpcCancelThread for the ncacn_ip_tcp protocol sequence.
by Hans Leidekker Nov. 12, 2007

Nov. 12, 2007

On Monday 12 November 2007 21:10:19 Robert Shearman wrote: > +static CRITICAL_SECTION_DEBUG threaddata_cs_debug = > +{ > + 0, 0, &uuid_cs, I smell a copy/paste artifact ;) -Hans

1 0

Too many WM_MOUSEACTIVATE
by Peter Åstrand Nov. 12, 2007

Nov. 12, 2007

I have lot's of problems with WM_MOUSEACTIVATE messages in "Magna Cura". The effekt is that certain popup dialog doesn't work: After selecting something, say a date, the whole "form" is cleared. After much debugging, I've found out that WM_MOUSEACTIVATE is the problem. If I explicitly stop WM_MOUSEACTIVATE from propagating from dialogs to parents, then things works OK. I do this by addding: case WM_MOUSEACTIVATE: return MA_ACTIVATE; in DefDlgProc{A,W}. I wonder how many things that will brake with a change like this.... Just a long shot, but has anyone else seen anything similiar to this? (It turns out that there are multiple differences compared to Windows; one with two many messages generated when the popup is created, another with too few messages when changing tab in a SysTabControl32 (but zero messages works fine as well.) Rgds, --- Peter Åstrand ThinLinc Chief Developer Cendio AB http://www.cendio.se Wallenbergs gata 4 583 30 Linköping Phone: +46-13-21 46 00 ThinLinc User Group 2007 Höstkonferens --- 15-16 november, Linköping. Program och anmälan: http://www.thinlincusergroup.se/aktiviteter

1 0

Problem with wine + Dreamweaver + VNC
by Francisco José Gallego Durán Nov. 12, 2007

Nov. 12, 2007

Hello, I am trying to run Dreamweaver.exe and Photoshop.exe in a remote gnome-session through vnc4server. When I try to run it I get the following message: Xlib: extension "XFree86-DRI" missing on display ": 3.0". X Error of failed request: BadValue (integer parameter out of range for operation) Major opcode of failed request: 148 (GLX) Minor opcode of failed request: 3 (X_GLXCreateContext) Value in failed request: 0x24 Serial number of failed request: 11 Current serial number in output stream: 12 Might it be a kind of a bug? I am running wine 0.9.49, Debian etch with latest updates, and vnc4server. Thank you for your help. Francisco Gallego.

1 0

Re: wininet: update test to pass on newer versions of windows
by Alexandre Julliard Nov. 12, 2007

Nov. 12, 2007

"Lei Zhang" <thestig(a)google.com> writes: > + > + /* sz is 14 on XP SP1 and beyond, 30 on XP and before */ > + todo_wine ok( sz == 14, "sz wrong\n"); The tests have to pass on all Windows versions. If some return 30 and some return 14 the test must accept both. -- Alexandre Julliard julliard(a)winehq.org

2 1

Re: widl [2/9]: Add an option to use Wine exception handlers.
by Alexandre Julliard Nov. 12, 2007

Nov. 12, 2007

Dan Hipschman <dsh(a)linux.ucla.edu> writes: > This patch adds a --wine-seh option to widl to use Wine's SEH instead of > the Windows one. This should not be a widl option. widl should generate a single version of the C code that works on all compilers, using appropriate macros and #ifdefs. Also you shouldn't use the wine/exception.h macros, since you are generating the code you can output the code directly, in a much more straightforward way than the convoluted __TRY macros. -- Alexandre Julliard julliard(a)winehq.org

1 0

How can we handle undcomented shell functionsAW, that MS declared now (WSTR version only)
by Detlef Riekenberg Nov. 12, 2007

Nov. 12, 2007

Hi There are a some undocumented Functions in shell32, which have one entrypoint that use STR in win9x, but WSTR in NT: - dlls/shell32/undocshell.h - http://www.geocities.com/SiliconValley/4942/paths.html Some of them are now declared without A/W suffix (include/shobj.h from SDK 2003), but since win9x is no longer supported by MS, it's only for the Wide version. Examples: PathProcessCommend, PathResolve, PathQualify Question for include/shlobj.h: What is the correct way to declare such functions? In undocshell.h, we use functionAW with LPVOID for the first two, but for PathQualify, we use: // BOOL WINAPI PathQualifyA(LPCSTR path); BOOL WINAPI PathQualifyW(LPCWSTR path); #define PathQualify WINELIB_NAME_AW(PathQualify) BOOL WINAPI PathQualifyAW(LPCVOID path); // And PathResolve in exported by name (see shell32.spec) ############################################################# Next question for shell32.spec: Are we fixed to "ptr" for such functions? ("str", "wstr" and "long" are all used and need to be fixed) Is it possible / useful to add support for "awstr" (or a similar name)? What about support for UNICODE-Strings ("ustr")? -- By by ... Detlef

2 1

Re: cmdlgtst: Rename to dialogtest
by Dmitry Timoshkov Nov. 12, 2007

Nov. 12, 2007

"Detlef Riekenberg" <wine.dev(a)web.de> wrote: > Windows has a lot more Dialogs in shell32, user32, kernel32 and various > other dlls. > > This is only a rename, the only code that changed is the name of the > debug-channel. IMHO it's better to remove programs/cmdlgtst altogether, it doesn't really test anything, and this is not a place for tests anyway. -- Dmitry.

2 1

Patch Follow up
by Alistair Leslie-Hughes Nov. 12, 2007

Nov. 12, 2007

Hi, Are there any reason they following patches where missed? Patch 1 of 2 - Return E_INVALIDARG on bad parameter Patch 2 of 2 - Enable Test for Bad Argument / Fix Memory leak clusapi: Corrected value in GetNodeClusterState Best regards Alistair Leslie-Hughes

2 1

Valgrind results for Oct 9: still lots of hangs
by Dan Kegel Nov. 12, 2007

Nov. 12, 2007

Results for today's git (plus my dde patch) are online at http://kegel.com/wine/valgrind/20071109/ ole32/tests/marshall.c has always been hanging when run under valgrind. In the last two days, the following additional tests hang under valgrind: mshtml/tests/dom.c mshtml/tests/htmldoc.c ole32/tests/marshal.c shdocvw/tests/webbrowser.c I can reproduce one hang with cd dlls/mshtml/tests make testclean wineserver -k RUNTEST_USE_VALGRIND=1 make test (with the wine-vg-utils.patch I provide. Make by itself does not behave well under valgrind, so you can't just do "valgrind make test"). Sadly, I don't have time to do a regression test, but it's pretty easy to repeat my results. See http://wiki.winehq.org/Wine_and_Valgrind It'd be nice if somebody else verified that these hangs occur for them too, that they didn't occur with wine-0.9.48, and maybe did a regression test to find out when they started... - Dan

3 4

Re: OLEDLG: use wide strings
by Dmitry Timoshkov Nov. 12, 2007

Nov. 12, 2007

"Vitaly Lipatov" <lav(a)etersoft.ru> wrote: > Changelog: > - fix buffers size for registry functions > - use unicode registry functions > - do not requires ntdll anymore (remove RtlMultiByteToUnicodeN using) This patch is extremely buggy, it passes wrong buffer sizes to various APIs in many places. -- Dmitry.

1 0

%fs, %gs on AROS hosted
by Staf Verhaegen Nov. 11, 2007

Nov. 11, 2007

Hello wine developers, I am an AROS developer (http://www.aros.org) that is an open source implementation of the old AmigaOS. It has also the possiblity to run hosted on other OSes and is thus very similar to wine. I know that this group is for wine posts but I think I could reuse much of the knowledge gained by the wine developers for AROS. Recently there has been discussion about the use of the %fs and %gs segment for AROS purposes and thus also for the hosted version. As wine also has to handle these registers, I am interested to know how wine handles these registers. Or doesn't wine have to take these registers into account ? I did download the 0.9.48 source code of wine and I only did find some places where a proper saving and restoring of these registers was done. Is there something I am overlooking ? Currently wine seems to run on the following platforms: Linux version 2.0.36 or above FreeBSD 6.2 or later Solaris x86 2.5 or later NetBSD-current Mac OS X 10.4 or later I found almost no OS specific code for each of the architectures in the source. Are there special precautions to be taken in one of these OSes to be able to use the segment register in user land code ? My last question is related to an AROS version hosted on Windows. Would Windows put any restrictions on the use of the segment registers for this purpose. Is there a difference between 32bit and 64bit ? thanks, Staf.

2 5

What the hell is up with searching google for wine-devel and wine-patches posts?
by Steven Edwards Nov. 10, 2007

Nov. 10, 2007

Hi, I guess this issue is more directed at Dan and the rest of the google guys we have lurking here but has anyone else noticed how hard it is to do a google search for discussions on wine-devel or look for patches that have been submitted? I used to just be able to search as follows wine-patches Steven Edwards and see every patch I ever submitted to wine. Ditto for most threads on wine-devel. This is really getting to be a pain for me because I need to be able to search for patches that might not have been merged in but still have merit for some of my customers. Thanks -- Steven Edwards "There is one thing stronger than all the armies in the world, and that is an idea whose time has come." - Victor Hugo

3 3

Linux/Windows app connection, how?
by Joshua Scholar Nov. 10, 2007

Nov. 10, 2007

I really want to have some way to communicate between a Linux program and a Windows program running under WINE. The connection doesn't need to be high speed, a stream is fine - I'm just sending some unicode text. I'm writing both programs myself, so I can implement this in the easiest way, though if there was some way to create and call a Windows COM object from Linux, that would be the most direct solution... But reading what I can about WINE, I'm guessing that the only easy thing to do is to write a Windows server application with winegcc and either connect to it through a socket - or to have it spawn a Linux application that it has a stream to. But I need some sample code to know how to do this, since I don't have too much time to experiment. Just to make it clear why I'm doing this - I have some wonderful TTS voices (Neospeech's SAPI5 Japanese voices) installed in WINE and I need to be able to invoke them from a Linux program. Thanks in advance for any pointers, Joshua Scholar

6 5

alsa->pulseaudio->alsa directsound hang
by Al Tobey Nov. 10, 2007

Nov. 10, 2007

I spent some time this evening playing around with trying to get wine sound output to go through pulseaudio on my fc8/rawhide (x86_64) box. I tried wine 0.9.43 (built), 0.9.47 (fc8 rpm), and Crossover 6.2.0. I was testing with Steam/TF2, but found that the same issue comes up with a install of Winamp Lite using its default DirectSound output plugin. If I configure nothing and let wine grab the soundcard (I have an Audigy, so it allows multiple streams), it works fine. If I configure it to use pulseaudio through the alsa plugin, it hangs. I see the stream show up in the pulseaudio volume control, but nothing plays and winamp's meter and progress bar are inactive. The "Test Sound" button in winecfg works and shows up in pavucontrol as a mono stream. I've tried numerous variations on emulation and bitrates with no change. I didn't have any luck getting the ESD support (which would go straight to PA) to even show up in winecfg, even after forcing it in regedit. I probably missed something obvious. I'll do some more debugging tomorrow, but I thought I'd throw my config out there for others to experiment with and ridicule. There isn't much out on Google about this kind of config yet, but I expect there will be more requests as the popularity of pulseaudio goes up, especially with the release of fc8. I think the latest Ubuntu has it enabled by default, too. Did anybody actually write a patch to directly support pulseaudio or was the earlier discussion devoid of code? My registry mods and .asoundrc for using the alsa/pulseaudio plugin are attached. Getting pulseaudio is left as an exercise to the reader ;) Thanks, -Al

3 4

Re: USER32: Patch 1 of 6: Fix returned value of LB_DIR
by James Hawkins Nov. 10, 2007

Nov. 10, 2007

On 11/10/07, Alex Villacís Lasso <a_villacis(a)palosanto.com> wrote: > Tests reveal that (on an unsorted listbox), LB_DIR returns the value of the > highest index inserted, as if it had sorted the list of files before starting > the insertion. This patch attempts to replicate this behavior. > > Changelog: > * Make LB_DIR return an index value as if it had inserted filenames in sorted > order. > Tests? Tests that aren't in the test suite that should be to verify the patch is correct? -- James Hawkins

1 1

CryptAcquireContext Failure, default/Type 024 requested
by Rob Seger Nov. 9, 2007

Nov. 9, 2007

Wine currently supports types 001 and 012, but I have a new game attempting to use type 024 (Software\\Microsoft\\Cryptography\\Defaults \\Provider Types\\Type 024). I've googled, as well as searched MSDN, for a reference as to what the crypt provider type 024 even is with no success. Does anyone know? Looking through the advapi32.dll code, this looks like it might be a serious undertaking to implement. Is that estimation accurate? Does anyone have any tips or pointers about implementing an entire CSP in wine? I appreciate any help, keep up the good work all! Rob

3 7

Re: [14/15] WineD3D: Get rid of the conditionals in shader_glsl_compare
by Mirek Slugeň Nov. 9, 2007

Nov. 9, 2007

2 2