June 2011 - Wine-Devel - list.winehq.org

shouldn't the 1.3 dependencies include binfmt-support and lib32nss

by Susan Cragin

Just wondering. apt-get build-dep wine1.3 does not load binfmt-support and lib32nss... Installation of natspeak requires them. The installation script tries to grab them but failed for me.

June 24, 2011

Ars Technica article

by Brian Vincent

Here's some light reading for your weekend enjoyment. I found this on Ars Technica and thought it was a pretty good summary of some stuff from Microsoft's past and a pretty good guess at how some future technology pieces fit together. http://arstechnica.com/microsoft/news/2011/06/windows-8-for-software-develo… -Brian

June 24, 2011

Re: How to do single-step debugging while testing?

by Michael Mc Donnell

On Fri, Jun 24, 2011 at 3:24 PM, Michael Mc Donnell <michael(a)mcdonnell.dk> wrote: > On Fri, Jun 24, 2011 at 2:56 PM, Stefan Dösinger <stefandoesinger(a)gmx.at> wrote: >> On Friday 24 June 2011 14:44:10 Michael Mc Donnell wrote: >>> Now I want to connect my graphical debugger. I enter "target remote >>> localhost:57380" in ddd's command line and it connects. But how do I >>> attach the source code and so on (I'm use to debugging in Eclipse and >>> VS)? >> I'm not sure about that, I never used any graphical debuggers with Wine. But >> gdb(and winedbg) always found the Wine source, my assumption was that the path >> to it is stored in the binaries, so ddd should automatically find it. > > Well I guess I can go old school with the command line version instead :-) So much for going old school. It does work, but it's next to useless for the stuff I'm doing. I really miss having a list of all the variables in the stack frame with their values, and seeing the code as I step through it. I tried again to follow the instructions from http://www.winehq.org/docs/winedev-guide/dbg-others but I could not get ddd nor kdbg to work. ddd comes up with "A syntax error in expression, near `0x681f2fbd'." and an empty source view. kdbg comes up and it does not find the sources either. The following error message is displayed "Unknown or malformed query Attached" when the test is run. After a few exceptions the test runs to the end. Does anyone use a graphical debugger?

June 24, 2011

Re: Behavior of SHChangeNotify

by Andrew Eikum

On 06/24/2011 12:27 PM, Jay Yang wrote: > To try to implement copy-paste for shell folders, I used SHFileOperation for the actual copy. SHFileOperation calls SHChangeNotify to notify listeners about changes to the file system. But I noticed that the shell view would not update correctly if it was at 'C:\' but it would if it was at '/home/user_name/.wine/drive_c/'. In the end I traced this to the fact that in the first case the shell view calls SHChangeNotifyRegister with a pidl of the form<Desktop|My Computer|C:> while SHFileOperation is using SHChangeNotify to send a notification to those listeners registered on pidls of the form<Desktop|/|home|user_name|.wine|drive_c>. This makes it such that the shell view doesn't always get the notifications it expects. I can think of a couple of ways to fix this. > > 1. Have SHChangeNotifyRegister normalize all file system pidls to ones rooted in the unixfs. > 2. Have the shell view register on a unixfs pidl. > > I would like to know if either of these is a good solution, or if there are better solutions. I can also write a quick test to demonstrate the problem if necessary. I ran into a similar problem with IShellFolder::AddFolder. That's commit 0a55ec2c9e1dc5fd0b99ad1202fd7849b0187808 which was for Bug 18606. Maybe that will help you? Just popped into my head when I saw your email... Andrew

June 24, 2011

Behavior of SHChangeNotify

by Jay Yang

To try to implement copy-paste for shell folders, I used SHFileOperation for the actual copy. SHFileOperation calls SHChangeNotify to notify listeners about changes to the file system. But I noticed that the shell view would not update correctly if it was at 'C:\' but it would if it was at '/home/user_name/.wine/drive_c/'. In the end I traced this to the fact that in the first case the shell view calls SHChangeNotifyRegister with a pidl of the form <Desktop|My Computer|C:> while SHFileOperation is using SHChangeNotify to send a notification to those listeners registered on pidls of the form <Desktop|/|home|user_name|.wine|drive_c>. This makes it such that the shell view doesn't always get the notifications it expects. I can think of a couple of ways to fix this. 1. Have SHChangeNotifyRegister normalize all file system pidls to ones rooted in the unixfs. 2. Have the shell view register on a unixfs pidl. I would like to know if either of these is a good solution, or if there are better solutions. I can also write a quick test to demonstrate the problem if necessary. -- Jay Yang

June 24, 2011

Re: GSoC: dinput8 Action Mapping

by Ricardo Filipe

there was no commit round today yet. usually allow for 2 days to pass if your patches are still "new". if then there is no feedback you can try asking in #winehackers 2011/6/24 Lucas Zawacki <lfzawacki(a)gmail.com> > Is there any problem with the latest patches I sent? They're sitting > there with the 'New' status but I have no feedback (good or bad) about > them. > > Cheers > > 2011/6/20 Lucas Zawacki <lfzawacki(a)gmail.com>: > >> Because callback doesn't have any string parameters it would be an easy > >> thing to do. > > > > I've made an EnumDevicesBySemantics version using A -> W calls, but it > > turned out to be trickier than I thought because the callback function > > is passed the device interfaces and then the A callback gets W > > interfaces. I'll send the patches with corrections later today but I > > really think the implementation for EnumDevicesBySemantics should > > remain separate right now. > > > > >

June 24, 2011

Re: [1/3] cmd/tests: Use plain strings instead of char arrays for keywords

by Dan Kegel

2011/6/24 Frédéric Delanoy <frederic.delanoy(a)gmail.com>: > On Fri, Jun 24, 2011 at 17:16, Dan Kegel <dank(a)kegel.com> wrote: >> Frédéric Delanoy wrote: >> - static const char escaped_space[] = {'@','s','p','a','c','e','@','\0'}; >> + static const char *escaped_space = "@space@"; >> >> I'm not sure it's worth the effort, since eventually we want >> cmd to use unicode internally, at which point we'll have >> to go back to using arrays. > > Well it seems you changed your mind since you were the one suggesting > this (on IRC) ;) I think we discussed it in email. I was wondering about it: > The current code is a bit ugly. If you're going to add a > nice is_todo_wine_line() helper, it might make sense > to do a cleanup patch that cleans up the @keyword@ > recognition in general first. (Why the verbose array > declaration instead of a "constant"? etc.) ... >Nah, looking at the code, it was so that sizeof() would not include >the trailing NUL. and it's true, I didn't mention the wchar issue then. Sorry. - Dan

June 24, 2011

Re: GSoC: dinput8 Action Mapping

by Lucas Zawacki

Is there any problem with the latest patches I sent? They're sitting there with the 'New' status but I have no feedback (good or bad) about them. Cheers 2011/6/20 Lucas Zawacki <lfzawacki(a)gmail.com>: >> Because callback doesn't have any string parameters it would be an easy >> thing to do. > > I've made an EnumDevicesBySemantics version using A -> W calls, but it > turned out to be trickier than I thought because the callback function > is passed the device interfaces and then the A callback gets W > interfaces. I'll send the patches with corrections later today but I > really think the implementation for EnumDevicesBySemantics should > remain separate right now. >

June 24, 2011

Re: [1/3] cmd/tests: Use plain strings instead of char arrays for keywords

by Frédéric Delanoy

On Fri, Jun 24, 2011 at 17:16, Dan Kegel <dank(a)kegel.com> wrote: > Frédéric Delanoy wrote: > - static const char escaped_space[] = {'@','s','p','a','c','e','@','\0'}; > + static const char *escaped_space = "@space@"; > > I'm not sure it's worth the effort, since eventually we want > cmd to use unicode internally, at which point we'll have > to go back to using arrays. Well it seems you changed your mind since you were the one suggesting this (on IRC) ;) Frédéric

June 24, 2011

re: [1/3] cmd/tests: Use plain strings instead of char arrays for keywords

by Dan Kegel

Frédéric Delanoy wrote: - static const char escaped_space[] = {'@','s','p','a','c','e','@','\0'}; + static const char *escaped_space = "@space@"; I'm not sure it's worth the effort, since eventually we want cmd to use unicode internally, at which point we'll have to go back to using arrays.

June 24, 2011

Re: Dylan Smith : d3dx9/tests: Added tests for D3DXCreateSkinInfo.

by Dylan Smith

On Fri, Jun 24, 2011 at 11:04 AM, Dylan Smith <dylan.ah.smith(a)gmail.com> wrote: > So how about the revised patch that uses memset to set all the bits to 1. Oops, forgot to diff against origin. Attached patch against origin.

June 24, 2011

Re: Dylan Smith : d3dx9/tests: Added tests for D3DXCreateSkinInfo.

by Dylan Smith

On Fri, Jun 24, 2011 at 10:19 AM, Stefan Dösinger <stefandoesinger(a)gmx.at> wrote: > On Friday 24 June 2011 16:01:21 Dylan Smith wrote: >> On Fri, Jun 24, 2011 at 9:31 AM, Michael Stefaniuc <mstefani(a)redhat.com> > wrote: >> > this commit breaks compilation on my old RHEL5 box: >> > dlls/d3dx9_36/tests/mesh.c:4813: error: ‘FP_NAN’ undeclared (first use >> > in this function) >> >> Does the attached patch fix the problem. It uses 0.0f / 0.0f like in >> wined3d_private.h. > Ya, visual studio reports a divide by zero error in this case :-/ > I see. I suppose NAN isn't crucial to the tests (the point is the value is not checked), so I could assume the IEEE 745 standard since it shouldn't break the tests if this assumption is wrong. So how about the revised patch that uses memset to set all the bits to 1. On Fri, Jun 24, 2011 at 10:56 AM, Dan Kegel <dank(a)kegel.com> wrote: > Dylan wrote: > - exp_weights[1] = FP_NAN; > + exp_weights[1] = 0.0f / 0.0f; > > It's a bit strange to see 0 being replaced with infinity. What is > your intent in that line? > FP_NAN probably isn't what you think it is. 0.0f / 0.0f seems to give NAN as a result, although using a comment or an appropriately named variable would make it more readable. The intent is to show that the value isn't checked, since usually weights are between 0.0 and 1.0.

June 24, 2011

re: Dylan Smith : d3dx9/tests: Added tests for D3DXCreateSkinInfo.

by Dan Kegel

Dylan wrote: - exp_weights[1] = FP_NAN; + exp_weights[1] = 0.0f / 0.0f; It's a bit strange to see 0 being replaced with infinity. What is your intent in that line? FP_NAN probably isn't what you think it is. On Fri, Jun 24, 2011 at 9:44 AM, Stefan Dösinger <stefandoesinger at gmx.at> wrote: > I'd appreciate a portable solution for generating NaN and Inf values as well, > I have the same problem with wined3d in visual studio. libxml2 has a nifty function trio_nan() which wraps all the portable ways of doing it. See e.g. http://www.opensource.apple.com/source/libxml2/libxml2-4/libxml2/trionan.c - Dan

June 24, 2011

Re: Dylan Smith : d3dx9/tests: Added tests for D3DXCreateSkinInfo.

by Michael Stefaniuc

Thanks Dylan, Dylan Smith wrote: > On Fri, Jun 24, 2011 at 9:31 AM, Michael Stefaniuc <mstefani(a)redhat.com> wrote: >> this commit breaks compilation on my old RHEL5 box: >> dlls/d3dx9_36/tests/mesh.c:4813: error: ‘FP_NAN’ undeclared (first use >> in this function) > > Does the attached patch fix the problem. It uses 0.0f / 0.0f like in > wined3d_private.h. WORKSFORME. > > On Fri, Jun 24, 2011 at 9:44 AM, Stefan Dösinger <stefandoesinger(a)gmx.at> wrote: >> I'd appreciate a portable solution for generating NaN and Inf values as well, >> I have the same problem with wined3d in visual studio. > > I don't see Inf values specified using a macro there, but good for > future reference. bye michael

June 24, 2011

Re: Dylan Smith : d3dx9/tests: Added tests for D3DXCreateSkinInfo.

by Stefan Dösinger

On Friday 24 June 2011 16:01:21 Dylan Smith wrote: > On Fri, Jun 24, 2011 at 9:31 AM, Michael Stefaniuc <mstefani(a)redhat.com> wrote: > > this commit breaks compilation on my old RHEL5 box: > > dlls/d3dx9_36/tests/mesh.c:4813: error: ‘FP_NAN’ undeclared (first use > > in this function) > > Does the attached patch fix the problem. It uses 0.0f / 0.0f like in > wined3d_private.h. Ya, visual studio reports a divide by zero error in this case :-/

June 24, 2011

Re: How to do single-step debugging while testing?

by Michael Mc Donnell

On Fri, Jun 24, 2011 at 3:55 PM, Henri Verbeet <hverbeet(a)gmail.com> wrote: > On 24 June 2011 15:24, Michael Mc Donnell <michael(a)mcdonnell.dk> wrote: >> I can set a break point at test_convert_point_reps_to_adjacency (my >> new test) and it stops there. I can also set a break point for >> ID3DXMeshImpl_ConvertPointRepsToAdjacency but it just goes past it. >> >> Here are the commands I used: >> b test_convert_point_reps_to_adjacency >> cont >> b ID3DXMeshImpl_ConvertAdjacencyToPointReps >> cont >> >> Any idea what I'm doing wrong? >> > Is that a typo, or did you really mean > "ID3DXMeshImpl_ConvertAdjacencyToPointReps" there? Doh! It should have been ID3DXMeshImpl_ConvertPointRepsToAdjacency. Thanks Henri for spotting that one. It now works as expected :-)

June 24, 2011

Re: Dylan Smith : d3dx9/tests: Added tests for D3DXCreateSkinInfo.

by Dylan Smith

On Fri, Jun 24, 2011 at 9:31 AM, Michael Stefaniuc <mstefani(a)redhat.com> wrote: > this commit breaks compilation on my old RHEL5 box: > dlls/d3dx9_36/tests/mesh.c:4813: error: ‘FP_NAN’ undeclared (first use > in this function) Does the attached patch fix the problem. It uses 0.0f / 0.0f like in wined3d_private.h. On Fri, Jun 24, 2011 at 9:44 AM, Stefan Dösinger <stefandoesinger(a)gmx.at> wrote: > I'd appreciate a portable solution for generating NaN and Inf values as well, > I have the same problem with wined3d in visual studio. I don't see Inf values specified using a macro there, but good for future reference.

June 24, 2011

Re: How to do single-step debugging while testing?

by Henri Verbeet

On 24 June 2011 15:24, Michael Mc Donnell <michael(a)mcdonnell.dk> wrote: > I can set a break point at test_convert_point_reps_to_adjacency (my > new test) and it stops there. I can also set a break point for > ID3DXMeshImpl_ConvertPointRepsToAdjacency but it just goes past it. > > Here are the commands I used: > b test_convert_point_reps_to_adjacency > cont > b ID3DXMeshImpl_ConvertAdjacencyToPointReps > cont > > Any idea what I'm doing wrong? > Is that a typo, or did you really mean "ID3DXMeshImpl_ConvertAdjacencyToPointReps" there?

June 24, 2011

Re: Dylan Smith : d3dx9/tests: Added tests for D3DXCreateSkinInfo.

by Stefan Dösinger

On Friday 24 June 2011 15:38:41 Alex Bradbury wrote: > On 24 June 2011 14:31, Michael Stefaniuc <mstefani(a)redhat.com> wrote: > >> Diff: > >> http://source.winehq.org/git/wine.git/?a=commitdiff;h=01ff07e2e1d2e7757 > >> 5ddbd88a705ef12ea6f9073 > > > > this commit breaks compilation on my old RHEL5 box: > > dlls/d3dx9_36/tests/mesh.c:4813: error: ‘FP_NAN’ undeclared (first use > > in this function) > > Austin English ran in to the same problem: > http://bugs.winehq.org/show_bug.cgi?id=27583 I'd appreciate a portable solution for generating NaN and Inf values as well, I have the same problem with wined3d in visual studio.

June 24, 2011

Re: Dylan Smith : d3dx9/tests: Added tests for D3DXCreateSkinInfo.

by Alex Bradbury

On 24 June 2011 14:31, Michael Stefaniuc <mstefani(a)redhat.com> wrote: >> Diff: http://source.winehq.org/git/wine.git/?a=commitdiff;h=01ff07e2e1d2e77575ddb… > this commit breaks compilation on my old RHEL5 box: > dlls/d3dx9_36/tests/mesh.c:4813: error: ‘FP_NAN’ undeclared (first use > in this function) Austin English ran in to the same problem: http://bugs.winehq.org/show_bug.cgi?id=27583 Alex

June 24, 2011

Re: Dylan Smith : d3dx9/tests: Added tests for D3DXCreateSkinInfo.

by Michael Stefaniuc

Hello Dylan, Alexandre Julliard wrote: > Module: wine > Branch: master > Commit: 01ff07e2e1d2e77575ddbd88a705ef12ea6f9073 > URL: http://source.winehq.org/git/wine.git/?a=commit;h=01ff07e2e1d2e77575ddbd88a… > > Author: Dylan Smith <dylan.ah.smith(a)gmail.com> > Date: Wed Jun 22 14:24:41 2011 -0400 > > d3dx9/tests: Added tests for D3DXCreateSkinInfo. > > --- > > dlls/d3dx9_36/tests/mesh.c | 229 ++++++++++++++++++++++++++++++++++++++++++++ > 1 files changed, 229 insertions(+), 0 deletions(-) > > Diff: http://source.winehq.org/git/wine.git/?a=commitdiff;h=01ff07e2e1d2e77575ddb… this commit breaks compilation on my old RHEL5 box: dlls/d3dx9_36/tests/mesh.c:4813: error: ‘FP_NAN’ undeclared (first use in this function) bye michael

June 24, 2011

Re: How to do single-step debugging while testing?

by Michael Mc Donnell

On Fri, Jun 24, 2011 at 2:56 PM, Stefan Dösinger <stefandoesinger(a)gmx.at> wrote: > On Friday 24 June 2011 14:44:10 Michael Mc Donnell wrote: >> Now I want to connect my graphical debugger. I enter "target remote >> localhost:57380" in ddd's command line and it connects. But how do I >> attach the source code and so on (I'm use to debugging in Eclipse and >> VS)? > I'm not sure about that, I never used any graphical debuggers with Wine. But > gdb(and winedbg) always found the Wine source, my assumption was that the path > to it is stored in the binaries, so ddd should automatically find it. Well I guess I can go old school with the command line version instead :-) I can set a break point at test_convert_point_reps_to_adjacency (my new test) and it stops there. I can also set a break point for ID3DXMeshImpl_ConvertPointRepsToAdjacency but it just goes past it. Here are the commands I used: b test_convert_point_reps_to_adjacency cont b ID3DXMeshImpl_ConvertAdjacencyToPointReps cont Any idea what I'm doing wrong?

June 24, 2011

Re: How to do single-step debugging while testing?

by Stefan Dösinger

On Friday 24 June 2011 14:44:10 Michael Mc Donnell wrote: > Now I want to connect my graphical debugger. I enter "target remote > localhost:57380" in ddd's command line and it connects. But how do I > attach the source code and so on (I'm use to debugging in Eclipse and > VS)? I'm not sure about that, I never used any graphical debuggers with Wine. But gdb(and winedbg) always found the Wine source, my assumption was that the path to it is stored in the binaries, so ddd should automatically find it. Also note that you don't have to use winedbg, you can also use gdb directly(ie, "gdb wine", then "start foo.exe.so"). This only works with a Wine binary created by make install, because the wine executable in the build tree is actually a shell script. You can also attach gdb to a running Wine process.

June 24, 2011

Re: How to do single-step debugging while testing?

by Michael Mc Donnell

On Fri, Jun 24, 2011 at 1:30 PM, Stefan Dösinger <stefandoesinger(a)gmx.at> wrote: > On Friday 24 June 2011 11:53:22 Michael Mc Donnell wrote: >> $ ../../../tools/runtest name_of_test >> >> Is there a way to run the test in a debugger, so that I can single >> step through the functions that the test calls? > ../../../wine winedbg foo_test.exe.so should work. Thanks. My next question is now how I use a graphical debugger. I'm trying to run d3dx9_36_test.exe.so, so I use the command: $ ../../../wine winedbg --gdb --no-start d3dx9_36_test.exe.so mesh It then starts and prints out what port to connect on. Something like: target remote localhost:57380 Now I want to connect my graphical debugger. I enter "target remote localhost:57380" in ddd's command line and it connects. But how do I attach the source code and so on (I'm use to debugging in Eclipse and VS)?

June 24, 2011

Re: How to do single-step debugging while testing?

by Stefan Dösinger

On Friday 24 June 2011 11:53:22 Michael Mc Donnell wrote: > $ ../../../tools/runtest name_of_test > > Is there a way to run the test in a debugger, so that I can single > step through the functions that the test calls? ../../../wine winedbg foo_test.exe.so should work.

June 24, 2011