October 2003 - Wine-Devel - list.winehq.org

Re: NTFS driver (was: ReactOS GPL vs. proprietary drivers)
by Fabian Cenedese Oct. 22, 2003

Oct. 22, 2003

Hi I don't know much about NTFS (only using it) and even less about Linux-NTFS. But it was funny following this conversation. >> Filesystem must be the rock solid data storage structure. You must >> know the meaning of each byte (*) for such reliable and interoperable >> filesystem. > >Exactly. Every needed byte is known. Wow! Really? >> (*) You do not need to know the journalling metadata as long as you do not >> support journalling and/or its recovery. > >One of the unknown issues is journaling :) AHA! And not only "The unknown issue" but "One of the unknown issues". Reminds me of Asterix: Whole Europe is part of the Roman Empire. Whole...? Sorry, I couldn't resist, no offense :) bye Fabi

2 1

Xlib and wine and XGrabButton
by Gerhard W. Gruber Oct. 22, 2003

Oct. 22, 2003

I currently write a small desktop app (in case you are curious http://sourceforge.net/projects/launchmenu) and for this I have to research on how to grab mouse events and eventually keyboard as well. Now I thought that this might be usefull for wine when playing games under the DirectX emulation, so I checked the sources for XGrabButton()/XGrabKey() and I couldn't find any references to it. When I tried to play Thief and Thief 2 I always had the problem that the keys are going to the shell where I started it from, instead of going into the game. I wrote a small test app for my project in order to get this feature straight and there I could capture all mouse events (haven't tried keyboard, but it should work equally well) without any interference with the remaining X system. Is this intenttionally that these functions are not used and if so what is the reason for it? Or is there a layer in between wine and X so that wine is not using Xlib API directly? I think that using this should work at least in programs where the entire screen is taken over, like in games, probably other applications well. -- Gerhard Gruber Für jedes menschliche Problem gibt es immer eine einfache Lösung: Klar, einleuchtend und falsch. (Henry Louis Mencken)

3 4

Re: Changes to kernel messages lang
by Alexandre Julliard Oct. 22, 2003

Oct. 22, 2003

Vincent Béron <vberon(a)mecano.gme.usherb.ca> writes: > W2k kernel messagetable is LANG_NEUTRAL, SUBLANG_NEUTRAL only. > > Changelog: > Make the kernel messagetable LANG_NEUTRAL, as native. Are you sure about that one? All the dlls I have checked (Win95, NT4, XP) have the language set to 0x0409. In what case is that causing trouble? -- Alexandre Julliard julliard(a)winehq.com

2 3

Re: More dutch translations
by Alexandre Julliard Oct. 21, 2003

Oct. 21, 2003

Hans Leidekker <hans(a)it.vu.nl> writes: > Changelog: > - More resources translated to dutch. > - Make an example consistent with it's description (wcmd). The wcmd change is wrong, the default prompt really is $p$g. Also your patch got wrapped, you'll have to resubmit. -- Alexandre Julliard julliard(a)winehq.com

1 0

S3 texture compression patent
by Jonathan Wilson Oct. 21, 2003

Oct. 21, 2003

What specificly does this patent cover and what parts of WINE would it affect? (i.e. which bits of windows, if implemented in WINE, would fall under the patent?)

6 7

Re: Summary of translations
by Vincent Béron Oct. 21, 2003

Oct. 21, 2003

Le mar 21/10/2003 à 14:45, Hans Leidekker a écrit : > This is certainly useful for me, as I was just about to start translating > some dlls to dutch. I checked the page you mention above and it showed > there are some parts missing from dlls that already haven been translated > to dutch. I don't think I would have found those otherwise (I wouldn't have > gone looking for them). So I think this tool can certainly prove valuable in > keeping resources up to date. That's why I did it (after I saw Dimi's output). Although now, all French resources seem to be translated... :) Glad to know it helped you too. > > And now for some criticisms. I can understand that you only check languages > for which at least one resource exists because otherwise the table would grow > large and show lots of empty cells for years to come, but the first thing I need > to set out translating is a list of dlls that have translatable resources, which I > now obtain by doing: > > [wine-src] $ find dlls -name "*En.rc" > > Which is crude and depends on a file naming convention. Maybe your tool > could help here too. If there's an html file, the dll contains some resources. Now, not all resources are translatable (icons, cursors, bitmaps we probably don't care, not sure about versioninfo). That can be integrated in the tool too. For now, if there's only a LANG_NEUTRAL version of a resource, it doesn't complain, else it creates the file. > > Second, those filenames for the html output really aren't fit to be served > by apache. They're truncated in a directory listing page so I have to hover > the links to be able to find a particular file. I think you need to either split > them into directories or provide an index page. As Dimi pointed out, all on a single page would be better. Or a couple of pages, with AAFRIKANS-GERMAN on one, HUNGARIAN-SLOVAK, etc. to prevent the table from being too wide, and marking only the relevant languages (ie, those for which at least one translated resource exists). Another thing is languages for which more than one sublanguage is used, such as Chinese. > > Third, I think you should sort the table columns alphabetically by language > name. But they're sorted! Ok, by language ID :) Good idea. > > Lastly, Konqueror doesn't render those popup/layer thingies well, and in > Mozilla I can't can rid of a popup once I clicked a link (well, I can press > the Back button, but that's non-intuitive to say the least). To be honest I > think they're overkill. Make them a regular link. Remember, you said all > comments were welcome ;-) No problem with that. I tried to get the same feeling as Mozilla's Tinderbox, but I can make them regular links instead. Or "tooltips". Vincent

1 0

Re: NTFS driver (was: ReactOS GPL vs. proprietary drivers)
by Szakacsits Szabolcs Oct. 21, 2003

Oct. 21, 2003

> > Let me tell an example, the (Linux) NTFS driver supports transparent > > compression. Today CPU's are very fast and the disk bandwidth is very > > slow compared. So doing bulk data transfers, the disk bandwidth is the > > bootleneck. Would you be faster using the filesystem's transparent > > compression? In theory yes, CPU can [de]compress when transfering data > > and you must also transfer less data. No main Linux fs supports this, > > hopefully Reiser4 will in the future ... (there was hope for e2compr as > > well for many years but it didn't work out after all). > > ext2 have supported this for ages, check out chattr(1),lsattr(1) This feature is reserved for about 8 years. But it never worked reliable. Google for the story. I wrote above the keyword (e2compr). > In what way are the compression features in NTFS different? It works reliable since about 8 years. Something what none of the main Linux filesystem can do today (ext3, reiserfs, JFS, XFS). Szaka

1 0

Re: [Rebar] Don't store band text if "" is passed into SetBandInfo
by Huw D M Davies Oct. 21, 2003

Oct. 21, 2003

On Tue, Oct 21, 2003 at 07:24:31PM +0100, Robert Shearman wrote: > Hi, > > This makes IE look a bit neater. > > Rob > > Changelog: > - Don't store band text if "" is passed into SetBandInfo > @@ -3447,8 +3450,11 @@ > } > if (lprbbi->lpText) { > INT len = lstrlenW (lprbbi->lpText); > - lpBand->lpText = (LPWSTR)Alloc ((len + 1)*sizeof(WCHAR)); > - strcpyW (lpBand->lpText, lprbbi->lpText); > + if (len > 1) Shouldn't this be len >= 1 ? > + { > + lpBand->lpText = (LPWSTR)Alloc ((len + 1)*sizeof(WCHAR)); > + strcpyW (lpBand->lpText, lprbbi->lpText); > + } > } > } > Huw.

1 0

Re: Fix comctl32 conformance test
by Rolf Kalbermatter Oct. 21, 2003

Oct. 21, 2003

Alexandre Julliard <julliard(a)winehq.org> wrote: <Hans Leidekker <hans(a)it.vu.nl> writes: > >> It appears that the comctl32 conformance test calls functions by name >> that are exported by ordinal only on Windows. Still the test runs (and >> thus run-time links) on Wine. Is this wanted behavior? > >Yes it is, the functions are converted to ordinal imports by the >import library. The existing code works just fine, except when >cross-compiling on Mingw because the import libraries are broken >there. This should be fixed in Mingw not in Wine. Mmh! I have downloaded the PSDK from the MSDN site, and although it says Microsoft Platform SDK February 2003 in the Start menu it really was just downloaded one month ago and my comctl32.lib does not contain symbols for at least DPA_Create() and DPA_GetPtr() while it does contain most of the other DPA symbols. What am I missing here? Rolf Kalbermatter

3 2

Re: Emulate win98 by default
by Andreas Mohr Oct. 21, 2003

Oct. 21, 2003

Hi, On Tue, Oct 21, 2003 at 11:25:55AM +0100, Mike Hearn wrote: > IE setup needs this, and really, there's not much difference between > them. I would change this in the code, but I don't know where this key > is referenced. > > Emulate win98 and DOS 6.11 by default > > Index: documentation/samples/config > =================================================================== > RCS file: /home/wine/wine/documentation/samples/config,v > retrieving revision 1.52 > diff -u -r1.52 config > --- documentation/samples/config 16 Oct 2003 19:09:57 -0000 1.52 > +++ documentation/samples/config 21 Oct 2003 10:19:26 -0000 > @@ -77,9 +77,9 @@ > > [Version] > ; Windows version to imitate (win95,win98,winme,nt351,nt40,win2k,winxp,win2k3,win20,win30,win31) > -;"Windows" = "win98" > +"Windows" = "win98" > ; DOS version to imitate > -;"DOS" = "6.22" > +"DOS" = "6.22" Please DON'T do this. These are FORCED, GLOBAL values, as such they are highly evil (since that way the whole version detection mechanism gets disabled and many programs which need other values cannot have their version guessed properly any more). Instead, the default fallback values in the version code should probably be changed to refer to Win98 to make sure we do use Win98 in case we cannot guess any value (if they are not already, but they might still be at Win95). Note that the version code has some straw man default values which are NOT the real defaults if you analyse the code path properly. And if IE setup still doesn't run after having changed the default guesstimate to Win98, tough luck: you'll have to force version Win98 manually (or implement a smarter version detection mechanism that takes some observations and transforms them into a Win98 version value). And we should also indicate in the sample config file that actually setting these will override any and all intelligent(?) version detection mechanisms... (if only to prevent other people from attempting the same mistake as you ;) -- Open Source Software: because smart people know to prefer a multitude of Support Options to a Single Point of Failure commonly used software represents. (who's left to sue given your average all-denying End User License Agreements, or who to ask for support if a vendor or product decided to silently die away?)

2 1