March 2001 - Wine-Devel - list.winehq.org

Re: Handling overlapping buffers in wcstombs_sbcs

by Alexandre Julliard

Fran=E7ois Gouget <fgouget(a)codeweavers.com> writes: > Actually the documentation specifies that WideCharToMultiByte (cal= led > by ConvertVersionInfo32To16) fails if the source and destination > pointers are the same. But it does not specify what happens if the > buffers merelyt overlap. And of course in the ConvertVersionInfo32To16 > case there's a 2 byte difference between them! Then ConvertVersionInfo32To16 is broken; WideCharToMultiByte does not support overlapping buffers. > I'm not really convinced by the loop unrolling but I tried to keep= it > anyway. I just modified it a bit. If someone's interested in doing > actual performance tests, wants to look at the assembly or knows how > manual loop unrolling must be done, go ahead. You know, I did exactly that (looking at the assembly and running performance tests) when implementing it; it may be possible to improve it more by optimizing CPU pipelines usage, which is something I don't know too much about. But I strongly doubt you can improve it by only looking at the C source... --=20 Alexandre Julliard julliard(a)winehq.com

March 23, 2001

Re: [PATCH] Add Callback for wavemapper -> widOpen

by James Hatheway

> But these other functions are not using the WINE_MLD internal winmm > structure, so why do you need it in widOpen()? Couldn't you instead do > exactly what wodOpen() does? Hi Alexandre, Ok, I see what you are saying now. Sorry about that. Humm.. I guess the short answer is because it doesn't work if I do it the same way wodOpen does. :-) I'll look into this and send another patch when I have a bit of time (maybe today or on Monday), it appears that it is breaking in MMDRV_WaveIn_Callback, but I'm not 100% sure yet... -James -- James Hatheway Software Designer - Macadamian Technologies, Inc. james(a)macadamian.com ~ http://www.macadamian.com "Man könnte froh sein, wenn die Luft so rein wäre wie das Bier" "One could be happy if the air were as pure as the beer"

March 23, 2001

Re: [PATCH] Add Callback for wavemapper -> widOpen

by Alexandre Julliard

"James Hatheway" <james(a)macadamian.com> writes: > I implemented the callback in a very similar way as the Wave out callback > is implemented. You can see this in the functions wodOpen(), wodOpenHelper(), > and WAVEMAP_DstCallback() , look at the data passed in the wom structure > (in dlls/winmm/wavemap/wavemap.c). I don't see any other way that I will > be able to callback into the app when the WineOSS layer notifies the wavemap > layer. The way it is now (wineoss directly calling the app) > causes my app to crash, and doesn't work that way on waveout. But these other functions are not using the WINE_MLD internal winmm structure, so why do you need it in widOpen()? Couldn't you instead do exactly what wodOpen() does? -- Alexandre Julliard julliard(a)winehq.com

March 22, 2001

Re: [PATCH] Add Callback for wavemapper -> widOpen

by James Hatheway

> It doesn't seem right to use winmm internal structures here. Are you > sure this is the right way to do it? Hi Alexandre, I implemented the callback in a very similar way as the Wave out callback is implemented. You can see this in the functions wodOpen(), wodOpenHelper(), and WAVEMAP_DstCallback() , look at the data passed in the wom structure (in dlls/winmm/wavemap/wavemap.c). I don't see any other way that I will be able to callback into the app when the WineOSS layer notifies the wavemap layer. The way it is now (wineoss directly calling the app) causes my app to crash, and doesn't work that way on waveout. -James -- James Hatheway Software Designer - Macadamian Technologies, Inc. james(a)macadamian.com ~ http://www.macadamian.com "Man könnte froh sein, wenn die Luft so rein wäre wie das Bier" "One could be happy if the air were as pure as the beer"

March 22, 2001

Re: Nasty Wine/kernel bug, anyone?

by Alexandre Julliard

lawson_whitney(a)juno.com writes: > Or is this a different glibc bug? Yes, it's apparently a bug in reference counting when freeing libraries; since we never free our libraries at the moment, we haven't run into this one yet. -- Alexandre Julliard julliard(a)winehq.com

March 22, 2001

Re: [PATCH] Add Callback for wavemapper -> widOpen

by Alexandre Julliard

"James Hatheway" <james(a)macadamian.com> writes: > Changelog: > James Hatheway - james(a)macadamian.com > widOpen in the wavemapper should use a callback. It doesn't seem right to use winmm internal structures here. Are you sure this is the right way to do it? -- Alexandre Julliard julliard(a)winehq.com

March 22, 2001

Re: problems compiling mfc with winelib

by Francois Gouget

On 21 Mar 2001, Doug Alcorn wrote: [...] > So, my > assumption based on reading winehq.com is that I need to compile the > mfc using winelib. Correct. > I got all the code over to my linux box and converted all the > files/directories to lowercase only. winemaker can do that for you. [...] > I had to define both CFLAGS and CPPFLAGS "-D_WIN32 -DWIN32 -D_WINDOWS > -D_MBSC" in the environment. Here's my configure line: > > $ ./configure --with-win=/path/to/wine/sources > --with-atl-includes=/path/to/atl/includes > --with-mfc-includes=/path/to/parent/of/mfc/includes '--with-atl-includes' and '--with-mfc-includes' are only necessary when compiling an MFC application, not the MFC themselves. They are used to specify where to find the MFC includes. But since you're compiling the MFC you already know where they are: it's coded into your Makefile.in. > It goes allong pretty well until it checks for afx.h. That's when it > bombs majorly. Yes. I guess you let winemaker do its stuff and it guessed that the MFC sources are MFC-based because of the stdafx.h file. But they are obviously not :-)! I suggest you rerun winemaker in interactive mode 'winemaker --interactive .' and specify --nomfc in the mfc/src directory (and the atl/src directory if you're going to really use it which you may very well not). This will also give you a chance to specify the exact include path for each directory. > I get pages and pages of compile time errors in the > config.log. Lots of parse errors, syntax errors, scoping errors, > redifinitions... lots of stuff. Once you get past the configure step above, which should be simple once you've done the above, you'll be able to take care of these. It will take some time, see my recent post in the newsgroup (titled 'Re: winelib + MFC'). > Am I right that I have to compile mfc? Yes, you have to. There's at least two good reasons for this: - winebuild cannot yet import dlls directly, but even if it could, it would not work with C++ dlls because VC++ and g++ use different binary layouts for classes, virtual tables, name mangling, etc. - the headers don't compile as is. So you'll have to modify them before you can use them anyway. > Is this something that most > developers using winelib to compile windows programs do? I don't really know. I guess if you have a large application it would be much more work to get rid of the MFC than to compile the MFC with Winelib. You may also end up having to support two versions unless you go for a cross-platform toolkit. > Are there instructions for doing so? Not much. See my post to the newsgroup. -- Francois Gouget fgouget(a)free.fr http://fgouget.free.fr/ In a world without fences who needs Gates?

March 22, 2001

Re: Nasty Wine/kernel bug, anyone?

by lawson_whitney＠juno.com

On Thu, 22 Mar 2001, gerard patel wrote: > Btw, since Borland released Kylix they talk of a glibc bug in dlopen : > > http://community.borland.com/article/0,1410,26998,00.html > > do you know if it's a concern for Wine too ? > Didn't we code around one of those not too long ago? One in glibc-2.1, and a different one in glibc-2.2 that was provoked by our solution to the first? > Gerard Or is this a different glibc bug? Lawson ---cut here ________________________________________________________________ GET INTERNET ACCESS FROM JUNO! Juno offers FREE or PREMIUM Internet access for less! Join Juno today! For your FREE software, visit: http://dl.www.juno.com/get/tagj.

March 22, 2001

Re: Port to OS/390 mainframe USS?

by James Hatheway

> > > Is there someone at Ulrich that I should be talking to? > > > > Ulrich is a person: Ulrich Weigand. He subscribes to the mailing list > > so I am sure he saw your message. You can get his email address from the > > mail list archives. > > ROTFL ! > > This really fits nicely into the old suspicion that he's not simply one > person, but in fact a myriad of people happily hacking away :) > (due to the amount of patches he used to submit) Hmm, I didn't see that angle. That is quite funny. I can picture a team of developers getting up on stage at a trade show and saying "We Are Ulrich!!" ^_^ -James -- James Hatheway Software Designer - Macadamian Technologies, Inc. james(a)macadamian.com ~ http://www.macadamian.com "Man könnte froh sein, wenn die Luft so rein wäre wie das Bier" "One could be happy if the air were as pure as the beer"

March 22, 2001

Re: Nasty Wine/kernel bug, anyone?

by gerard patel

At 12:14 PM 20/03/2001 -0600, you wrote: <snip> >I saw at least other user report on this (Steve Fox reported this >with Lotus Notes). > >My best guess is that this is a bug in the vfat driver, >but that's just a hunch. Where did this Notes bug turn up ? I never saw it on the news group, and it's not in your bug database. I'm curious because I would like to know how to reproduce it (I don't want not install Win98 and Outlook to test this :-/) Btw, since Borland released Kylix they talk of a glibc bug in dlopen : http://community.borland.com/article/0,1410,26998,00.html do you know if it's a concern for Wine too ? Gerard

March 22, 2001