April 2004 - Wine-Devel - list.winehq.org

Re: Remove debugmsg from guides
by Dimitrie O. Paun April 26, 2004

April 26, 2004

On Mon, 26 Apr 2004, Ivan Leo Murray-Smith wrote: > attaching as text/plain can't be done for various reasons, so I'll both attach > and inline. Please don't do that, we've been over this a thousand times, let's not rehash it. Do either or, otherwise you end up confusing Alexandre's script, etc. Inlining is prefered for a lot of reasons, so if you're using a decent mail client (and you seem to be), just inline it and don't look back :) -- Dimi.

1 0

Re: RFH: RPC patch
by Ivan Leo Murray-Smith April 26, 2004

April 26, 2004

Running server and client with wine CVS current returns Hello Implicit RPC World! but I can't test with your patch because it doesn't cleanly apply against the current CVS code. Ivan.

1 0

service.cgi: script generation
by Dimitrie O. Paun April 26, 2004

April 26, 2004

Currenly we send a script to run and execute all published versions that we have. This is currently problematic, because an error in one causes the abort the entire script. Moreover, releases may no arrive at the server at the same time, and if we're going to receive the poll event in between them, we'll miss some of them. What we need to do is to ask the client to execute one thing at a time. Problem is that we need to keep track of what the client did execute already. Here's a way to do it without storing state on the server side, which is problematic. We need to introduce three new cookies: build_id: the current build ID that we execute. It's going to be something like '200405031000' current: the name of the test that was just executed history: a comma separated list of filenames that where already executed by the client for this current build. It's going to look like so: 'wine-200405031000-paul-mingw.zip,wine-200405031000-kevin-mingw.zip' The algorithm is simple: if $build_id != $current_build_id $build_id = $current_build_id $history = "" else $history = $history . "," . $current $current = pick_next_test($history) # send script to run $current, and to set the values of the # three cookies to the values in our variables where: pick_next_test picks the next test to run from the pool of tests avaiable for our current build, so that it doesn't appear in $history Comments? -- Dimi.

1 0

Re: Font selection - please review
by Shachar Shemesh April 26, 2004

April 26, 2004

Hi Dirk, First of all, thank you for the reference. Regarding the missing flag to TranslateCharsetInfo - My VisualStudio 6 MSDN DVD (Latest that came up for that version - dated October 2001) does not have this flag mentioned. This means either that the flag was not present at that time, or that it was present but they forgot to mention it (has happened to me before). The latest MSDN (http://msdn.microsoft.com/library/default.asp?url=/library/en-us/intl/unico…) does mention it, and claims it's available on Windows 2000 (like you said). I guess this is anyone's guess what this means as far as when you can use it. Fortunetly for Wine, we don't care when it was introduced. If Wine has it, we can use it. At least for code written by us, that rule is fine. So, thank you very much for the good advice. Are you a wine user that needs this flag, or did you just happen to see my email when you were investigating a Win32 problem? Shachar Dirk Luetjens wrote: > Hello Shachar Shemesh, > > sorry for the lost context. Your mail I replyed to, is available at > > http://www.winehq.com/hypermail/wine-patches/2002/11/0221.html > > I know, that the wine project is totally unrelated to any windows dll. > I thought it would be better to mention, that this additonal flag for > TranslateCharsetInfo is not available in a specific windows version, > since the wine project tries to implement the WIN32 interface, and it > should behave as close as possible to the original one. Judging from > the code in > > http://source.winehq.org/source/objects/font.c#L1849 > > this additional flag is not yet implemented in wine. I can not tell > you when it was introduced into WIN32 itself. The MSDN says something > about WIN2000, but it isn't available on my WIN2k installation. > > Dirk > > > Shachar Shemesh wrote: > >> Hi Dirk, >> >> Judging from the email address you used to send me this email, we are >> talking about a fairly old post. Can you please direct me to the some >> context regarding what was my post exactly? A link with the post in >> question would be very helpful. >> >> Also, I hope you do realize that my post was on a Wine mailing list, >> which means that the functions there are totally disjointed from any >> particular Microsoft version or DLL. >> >> Many thanks, >> Shachar >> >> Dirk Luetjens wrote: >> >>> Hi Shachar Shemesh, >>> >>> I came across your post while researching a problem I had during >>> application development under windows. >>> >>> While digging in the MSDN I found another function that deals with >>> this problem: TranslateCharsetInfo >>> >>> --- >>> TranslateCharsetInfo function translates based on the specified >>> character set, code page, or font signature value, setting all >>> members of the destination structure to appropriate values. >>> --- >>> >>> and for the flag parameter: >>> >>> --- >>> TCI_SRCLOCALE Windows 2000: Source (that is, lpSrc) is the LCID or >>> LANGID of the keyboard layout. If it is a LANGID, the value is in >>> the low word. >>> --- >>> >>> The destination of the translation is a CHARSETINFO structure, where >>> one entry is the charset. >>> >>> I had problems using this parameter, since it isn't implemented in >>> the gdi version I use (VS6, gdi32.dll 5.00.2195.6898). I can only >>> find the definition in the newer .NET studio headers. But perhaps >>> this is a possible inclusion point for your function. >>> >>> Best regards >>> Dirk Luetjens >>> >> >> > > -- Shachar Shemesh Lingnu Open Source Consulting http://www.lingnu.com/

1 0

Re: winmm wave test
by Robert Reif April 26, 2004

April 26, 2004

Francois Gouget wrote: > Some Windows sound drivers allow wild frequencies such as the 2MHz > one. > Is this a windows driver bug? Is the test really valid and it's just finding a real bug in the windows driver for that specific sound card? I know it makes the test fail on bad windows drivers but is that bad?

2 4

email add
by kosta April 26, 2004

April 26, 2004

1 0

Re: Add some functions to msvcrt.dll
by Dmitry Timoshkov April 25, 2004

April 25, 2004

"kata 198" <kata198(a)hotmail.com> wrote: > Changelog: > Add some functions to msvcrt.dll If you could make your patch behave according to current locale that would be really nice. Otherwise it's not acceptable. Also, please make sure that your formatting style is at least consistent, and even better conforms to the formatting style of the file you are editing. -- Dmitry.

1 0

Window management protocol extensions
by Mike Hearn April 25, 2004

April 25, 2004

The latest WM problems caused by Metacity preventing windows from being moved offscreen prompted to raise the issue of having a Wine specific extension to the NETWM (EWMH) protocol into which we can easily add whatever semantics we need to map win32 windowing to X windowing more easily. The idea would be that we have a new property/set of properties we can mark windows with, _NET_WM_WINDOWS_EMULATION for instance, and standards compliant window managers will take note and not do things that break us. Short of actually doing all our own window management, this is probably the best way to fix the obscure bugs caused by different interpretations WMs have of what various hints mean. The current NETWM border hints are semantic not presentational, so we cannot rely on particular intepretations. The MWM hints are even worse - they are apparently undocumented, and most modern WM authors consider them deprecated. Thanks to their undocumentedness, WMs vary in their interpretation of these too: http://bugs.kde.org/show_bug.cgi?id=73294 So, I'd like to start gathering ideas on what sort of things we'd like to have in the NETWM spec, seeing as how many WMs conform to it these days. First up is a way to ensure the WM doesn't attempt to keep the window onscreen of course :) Other things we probably want are related to border styles, an absolutely crystal-clear no-wriggling-possible equivalent to PPosition as apparently the wording of the ICCCM allows WM authors to ignore this if they want. We might also want to look at where we use unmanaged windows and see if we can replace them with the use of managed windows+hints. We can probably hack around the lack of these things for now, but being able to use these hints where possible will probably keep a lot of us from going grey-haired early :) thanks -mike

2 2

RFH: MDI Window menu fix
by Krzysztof Foltman April 25, 2004

April 25, 2004

I've just finished fixing the behaviour of Window menu in MDI applications. However, I'm not sure if it won't cause some other MDI applications to work incorrectly. So, if you have some time and MDI apps to play with (regardless if they worked correctly before or not), check this out and tell me if I've fixed or broken the Window menu. http://www.winehq.org/hypermail/wine-patches/2004/04/0414.html Krzysztof

1 0

Re: abnt2 keyboard patch
by Mauro Carvalho Chehab April 25, 2004

April 25, 2004

Rafael, With your patch, the '\' key on the main keyboard will stop to work on the ABNT2 Keyboard, at least on the ABNT2 keyboards I've tested (some low valued keyboards and also Compaq keyboard). On ABNT2 keyboards, before the 'z' key there is the '\' key and the layout has to reflect that. That was the situation before 200303xx version. That changes was required in order to support that key. Had you tested it? If it works on your keyboard, may be you has some ABNT2 variant. I think it would be better to find another solution for the numeric keyboard that doesn't break the main keyboard. Mauro Chehab > Date: Fri, 16 Apr 2004 10:52:01 -0300 > From: Rafael Ávila de Espíndola <rafael.espindola(a)ic.unicamp.br> > To: wine-devel(a)winehq.org > Subject: Re: abnt2 keyboard patch > > > That's why I said that it's an X11 bug and we can do nothing about it. > > > > Please resubmit the patch without that part. > Attached. I will contact xfree and try to fix the rest of the problem. > > Thanks. > > > Dmitry. > Rafael >

2 1

RFH: RPC patch
by Filip Navara April 25, 2004

April 25, 2004

Hi, for about week and half I was working on a patch that makes Wine RPC runtime communicating with the Windows one. Now I'm almost done with the work and I need help from someone who can test the patch under Wine. (I was doing all my testing under regular Windows.) If someone will be willing to help me testing it, the patch is at www.volny.cz/xnavara/rpc/ along with a test programs. Regards, Filip

1 0

Wine bug with MFC/C++?
by greensh＠knology.net April 25, 2004

April 25, 2004

It appears that Wine cannot use tmpfs... so far my testing only works using swap :o( which I do have enabled at home and it crashes. At work I have a 500MB swap space and the same exact program does not crash there. It's not convenient for me to run swap at home, so does anyone know a fix for this? The code seems to require approx 500MB free memory. Since I only have 500MB RAM, I setup tmpfs on /tmp. But wine doesn't seem to use any of it. Shaun

1 0

winetest.exe output from Windows 95
by Jakob Eriksson April 25, 2004

April 25, 2004

By killing processes left and right, I managed to get a 50 - megabyte log. Maybe in the future we should link in bzip2 in winetest.exe to compress logs before submitting? regards, Jakob

2 2

Re: Remove debugmsg from guides
by Dimitrie O. Paun April 25, 2004

April 25, 2004

On Sun, 25 Apr 2004, Ivan Leo Murray-Smith wrote: > Please let me know if something isn't OK. A few points: -- I find the change confusing. There's no WINEDEBUG channel. I think most should just change from debugmsg -> debug -- Please provide only one .diff file. Multiple diffs are not necessary, and makes reviewing and applying so much more difficult. -- It would be nice if the diff would be either inlined or attached as text/plain. Inlining is prefered, this way people can quote easily when replying. -- Dimi.

1 0

Re: Remove debugmsg from guides
by Eric Pouech April 25, 2004

April 25, 2004

Ivan Leo Murray-Smith a écrit : > With this all unneeded references to debugmsg should be gone. Just one note, in > debugger.sgml I found this > [MACHINE\\Software\\Microsoft\\Windows NT\\CurrentVersion\\AeDebug] 957636538 > "Auto"=dword:00000001 > "Debugger"="winedbg --debugmsg -all %ld %ld" > As I don't know what it's meant to be now, I've left it, if it should change, > let me know. > Please let me know if something isn't OK. > > Ivan. this should be changed into "Debugger"="winedbg %ld %ld" (the issue is that the debugger will be inheriting the parent's debug channel, while this setting was turning off all debugging channels from debugger itself). A+

1 0

Re: PATCH: silence strict aliasing in dlls/opengl32
by Lionel Ulmer April 25, 2004

April 25, 2004

> Not sure how to regenerate the opemngl_ext.c file. You need the extension registry CVS (you can find it here http://oss.sgi.com/projects/ogl-sample/ - no idea where the exact CVS informations are). After, do 'make_opengl PATH_TO_EXTENSION_DIR 1.2' where PATH_TO_EXTENSION_DIR points to the '<OGL-SAMPLE-ROOT>/main/doc/registry/specs/' directory). Anyway, if it can wait, it will be updated the next time I generate the file. Lionel -- Lionel Ulmer - http://www.bbrox.org/

1 0

dlls/d3d8/shader.c
by Marcus Meissner April 25, 2004

April 25, 2004

Hi, I was just fixing this code, but I wonder why floorf(s0->w) is overwritten right next with tmp_f = powf(2.0f, s0->w); What is the right version? Ciao, Marcus void vshader_expp(D3DSHADERVECTOR* d, D3DSHADERVECTOR* s0) { float tmp_f = floorf(s0->w); DWORD tmp_d = 0; tmp_f = powf(2.0f, s0->w); tmp_d = *((DWORD*) &tmp_f) & 0xFFFFFF00;

2 1

Bad string handling in PRINTDLG_UpdatePrintDlg?
by Francois Gouget April 24, 2004

April 24, 2004

In PRINTDLG_UpdatePrintDlgA() in dlls/commdlg/printdlg.c we do: ... LPPRINTER_INFO_2A pi = PrintStructures->lpPrinterInfo; ... pi->pPortName = "FILE:"; And in PRINTDLG_UpdatePrintDlgW() we do: ... LPPRINTER_INFO_2W pi = PrintStructures->lpPrinterInfo; ... static WCHAR file[] = {'F','I','L','E',':',0}; pi->pPortName = file; ... Now in gcc all string literals are read-only, even if you declare them as char* foo="bar", *foo='\0' will cause a segfault. So in the Ansi version we point pPortName to a read-only string. But if the application gets hold of that pointer we have no garantee that it won't expect to be able to modify that string: pPortName is an LPSTR. (this is a case where -Wwrite-strings will complain) And in the Unicode case we point pPortName to a static writable buffer. But in this case if the application writes in our buffer its contents is going to be messed up forever. However I don't know this code at all. So I'm asking, shoudln't we make a copy of these strings? Which also raises the ugly issue of later freeing the corresponding memory buffer... -- Francois Gouget fgouget(a)free.fr http://fgouget.free.fr/ Any sufficiently advanced Operating System is indistinguishable from Linux

1 0

Idea to fight spam
by Ivan Leo Murray-Smith April 24, 2004

April 24, 2004

What about having a link to this page somewhere on winehq? Maybe on the mailing lists page? http://www.hostedscripts.com/scripts/antispam.html Ivan.

2 1

Re: Docu: mention piping for debugmsgs
by Ferenc Wagner April 24, 2004

April 24, 2004

Andreas Mohr <andi(a)rhlx01.fht-esslingen.de> writes: > Make people know how to reduce log size, using pipes. Hi Andi! Two notes: 1. The --debugmsg option is deprecated, people should use the WINEDEBUG environmental variable instead. 2. Won't the pipe block until you read from it, then give you all the data ever written? -- Feri.

1 0

winetest: build & publishing
by Dimitrie O. Paun April 24, 2004

April 24, 2004

Hi folks, I've checked the build this morning on SF, and it looks beautiful! So let's review the possible problems still outstanding: 1. Build Info Feri want to move that in the resources, with good reason. This will probably affect you very little, but needs to be done. We can still keep it in the .zip file (it is tiny) as it's a lot easier accessible for others. However, if we do that, the file will need to use DOS line ends, as this file will probably be viewed by folks on Windows boxes. Anyway, from your point of view it will probably mean a 'unix2dos' and a 'cp' before the build. 2. Uploading to SF This works nice now, but I'm afraid it may not when we have more than one person doing it. Namely, a release needs to be created, will that not fail if it already exists? Can we check that? If that's a problem, how are we going to deal with it? Kevin, can you please try to see if what happens if you upload two files (with differents sfutils invocations) into the same release? 3. Downloading from SF I just realized there is a time delay for the uploads to be propagated to the mirrors. What URL are we going to use for downloading? How long are we going to allow for the propagation to happen? Anyone has experience with this? 4. Release lifetime It would be nice to have the "WineTest Files" package visible on SF. I have made it so for testing purposes now. However, it will become ugly soon, because we don't hide the old releases. We can hide the entire package, and that would solve the ugliness problem, but it would be much nicer to have an automatic way of hiding old releases, and having just the latest releases visible. Can we maybe modify the sfutils to do this for us? We are very close on this front. I'm very happy with the state of things right now, I hope we can resolve these in a few days. -- Dimi.

5 6

Re: FILEDLG_MapOfnStructA bug?
by Dmitry Timoshkov April 24, 2004

April 24, 2004

"Francois Gouget" <fgouget(a)free.fr> wrote: > -static char defaultopen[]="Open File"; > -static char defaultsave[]="Save as"; > > /*********************************************************************** > * FileDlg_Init [internal] > @@ -970,9 +968,13 @@ > if (ofnA->lpstrTitle) > str = ofnA->lpstrTitle; > else > + { > /* Allocates default title (FIXME : get it from resource) */ > + static const char* defaultopen="Open File"; > + static const char* defaultsave="Save as"; Don't you really want to leave it as a static const array of characters as before and not a static const pointer? Same for your next patch. -- Dmitry.

1 0

Re: Fix for test of CreateFileMapping in dlls/kernel/tests/file.c
by Stefan Leichter April 24, 2004

April 24, 2004

Am Samstag, 24. April 2004 00:54 schrieb Jakob Eriksson: > From a run of winetest 20040423 on Windows XP: > > file.c:1075: Test failed: mapping should succeed > file.c:1077: Test failed: can't close mapping handle > file: 487696 tests executed, 0 marked as todo, 2 failures. > > > > Seems CreateFileMapping has some seriously strange behaviour. > I tested some more how XP behaves and produced a patch based on that. > > > regards, > Jakob Hello Jakob, + *handle = CreateFile(filename, GENERIC_READ|GENERIC_WRITE, 0, 0, + CREATE_ALWAYS, FILE_ATTRIBUTE_NORMAL, NULL); + if (handle != INVALID_HANDLE_VALUE) { do you have to test the pointer not the value ? Bye Stefan

2 1

Re: Fix for test of CreateFileMapping in dlls/kernel/tests/file.c
by Jakob Eriksson April 24, 2004

April 24, 2004

Jakob Eriksson wrote: > > From a run of winetest 20040423 on Windows XP: > > file.c:1075: Test failed: mapping should succeed > file.c:1077: Test failed: can't close mapping handle > file: 487696 tests executed, 0 marked as todo, 2 failures. > > > > Seems CreateFileMapping has some seriously strange behaviour. > I tested some more how XP behaves and produced a patch based on that. But I just installed Windows 95 and it fails very much under that platform: will have to think more about that... file.c:58: Test failed: couldn't create file "testfile.xxx" (err=32) file.c:105: Test failed: couldn't create file "testfile.xxx" (err=32) file.c:180: Test failed: couldn't create file "testfile.xxx" (err=32) file.c:204: Test failed: couldn't create file "testfile.xxx" (err=32) file.c:319: Test failed: couldn't create file "testfile.xxx" (err=32) file.c:354: Test failed: couldn't create file "testfile.xxx" (err=32) file.c:394: Test failed: couldn't create file "testfile.xxx" (err=32) file.c:441: Test failed: couldn't create file "testfile.xxx" (err=32) file.c:683: Test failed: DeleteFileA("nul") returned ret=0 error=5 file.c:893: Test failed: couldn't create file "testfile.xxx" (err=32) file.c:982: Test failed: couldn't create file "testfile.xxx" (err=32) file.c:1077: Test failed: Map is already created, for some reason we should be able to remap now... file.c:1081: Test failed: can't close mapping handle file.c:1086: Test failed: We should still be able to map! file.c:1087: Test failed: can't close mapping handle file.c:1088: Test failed: can't close file handle file.c:1097: Test failed: not ERROR_FILE_INVALID file.c:1108: Test failed: mapping should fail file.c:1110: Test failed: can't close file handle file.c:1111: Test failed: DeleteFile failed after map file.c:1118: Test failed: open testfile.xxx failed file.c:1120: Test failed: expected type disk got 0 file: 76 tests executed, 0 marked as todo, 22 failures.

1 0

graphical Linux
by Lonnie Cumberland April 24, 2004

April 24, 2004

Hello All, I am going to work on a project that will rely heavily on OpenGL and Wine. I wanted to get some of your opinions on which Linux distribution has the easiest setup and install for dealing with OpenGL and Wine. I will make a fresh install of the best latest version and build on it. did try Mandrake once but the OpenGL did not seem to work with my NVIDIA GeForce FX 5200 Any suggestions would be greatly appreciated. Lonnie __________________________________ Do you Yahoo!? Yahoo! Photos: High-quality 4x6 digital prints for 25� http://photos.yahoo.com/ph/print_splash

4 3

DirectPlay
by Ryan Underwood April 24, 2004

April 24, 2004

I have seen some mentions of work on DirectPlay. Who is currently working on this, and what are the implementation plans? I am mainly curious if it is planned to be network compatible with the same programs running on Windows. There could be some drawback to this, because the old versions of DirectPlay are not NAT friendly. So in the case of Wine, if the user is running a program that uses an old version of DirectPlay (i.e. the DX6 version), perhaps we could give them the choice to maintain compatibility with Windows clients (and thus be subject to NAT issues), or to use a more robust WINE-only version instead, perhaps an identical protocol to the one that would be implemented to support the later more robust versions of DirectPlay. -- Ryan Underwood, <nemesis(a)icequake.net>

3 2

winetest: dissect & gather
by Dimitrie O. Paun April 23, 2004

April 23, 2004

Hi Feri, Here is a list of things that I see as potential problems: (I haven't checked the code, so they may already be handled) 1. Multiple builds For a given BUILD_ID, we can have reports comming from different builds now, that is from different winetest-*.exe. That means a much of URLs, etc. 2. Multiple reports For a given BUILD_ID, we may get multiple reports from the same client (same Tag). This is because of (1), we can ask each client to run not one, but a bunch of the available tests (this is very useful, say we have each client run a mingw build, and a msvc build). 3. Build Info This is a minor issue, and we can safely ignore it for now, but maybe we can provide a little icon beside each URL that we list in (1), that would popup a window and display the Build Info. 4. Template Not a biggie, but maybe we can transform the generated report into a template so that it integrates nicely with the rest of the site (menu to the left, etc). Jer, what do we need to do to achive that? I think it should be very easy, but... -- Dimi.

4 8

Re: Winetest: outstanding issues (SERVING)
by Dimitrie O. Paun April 23, 2004

April 23, 2004

And the thrist and last installment: SERVING (Brian) --------------- What we need to support: -- multiple publishers -- multiple files for the same build for each publisher -- various distribution policies What we need to do: -- maintain a pool of releases as described in my previous mail. This is a map/hastable keyd by the published filename. -- persist this to disk, in case of server crash, load it from disk on startup. -- currently we use <program>/url.mask to get to the url mask. We need to add the publisher info in there, so this will need to change to: <publisher>/<program>/url.mask Similar for the other files. -- the rest remains unchanaged, as Brian correctly mentioned. We can have different distribution policies: pick a random file, send all files, etc. Given that we have few published files, and tests are short, I suggest we send them all. So what we have to do is to add a loop around most of the script generation: for file in pool: generate script commands to download it and run it It looks like there aren't that many changes after all. -- Dimi.

3 3

Re: winetest: build & publishing
by Jakob Eriksson April 23, 2004

April 23, 2004

Okay, but am I missing the point? winetest.exe is already self-extracting in itself... regards, Jakob chmorgan(a)charter.net wrote: > Lately I've been doing a lot of installer building with the NSIS > installer(http://nsis.sf.net). We could easily package winetest as a > self-extracting and executing installer that was transparent to the > user. The installer would simply extract the actual winetest.exe to > windows/temp, start it up and wait until it completed before exiting. > This should be transparent to winrash as well. > > Chris > > > > >> From: Jakob Eriksson <jakov(a)vmlinux.org> >> >> Can you also upload the exe? >> >> Ferenc touched this issue slightly - if we want to point (maybe very >> novice) users >> to a way to contribute to Wine - commandline option may be hard to >> understand. >> A zipped file can also be an obstacle. >> It's very neat just to click "open" on an exe compared to unpacking a >> zip. >> >> >

1 0

Re: Re: winetest: build & publishing
by chmorgan＠charter.net April 23, 2004

April 23, 2004

Lately I've been doing a lot of installer building with the NSIS installer(http://nsis.sf.net). We could easily package winetest as a self-extracting and executing installer that was transparent to the user. The installer would simply extract the actual winetest.exe to windows/temp, start it up and wait until it completed before exiting. This should be transparent to winrash as well. Chris > > From: Jakob Eriksson <jakov(a)vmlinux.org> > Date: 2004/04/23 Fri PM 02:24:17 EDT > To: "Dimitrie O. Paun" <dimi(a)intelliware.ca> > CC: Wine Development <wine-devel(a)winehq.org>, > Paul Millar <paulm(a)astro.gla.ac.uk>, Kevin Koltzau <kevin(a)plop.org>, > Ferenc Wagner <wferi(a)afavant.elte.hu> > Subject: Re: winetest: build & publishing > > Dimitrie O. Paun wrote: > > > 2. Uploading to SF > > This works nice now, but I'm afraid it may not when we > > have more than one person doing it. Namely, a release > > needs to be created, will that not fail if it already > > exists? Can we check that? If that's a problem, how are > > we going to deal with it? Kevin, can you please try to > > see if what happens if you upload two files (with > > differents sfutils invocations) into the same release? > > > > > > Can you also upload the exe? > > Ferenc touched this issue slightly - if we want to point (maybe very > novice) users > to a way to contribute to Wine - commandline option may be hard to > understand. > A zipped file can also be an obstacle. > It's very neat just to click "open" on an exe compared to unpacking a zip. > > regards, > Jakob > > > >

1 0

project David?
by Rein Klazes April 23, 2004

April 23, 2004

Hi, Somehow, we are doing something far to complicated: http://www.specopslabs.com/david.htm Rein. -- Rein Klazes rklazes(a)xs4all.nl

13 18

Google's cachze link update
by PETREOLLE Sylvain April 23, 2004

April 23, 2004

Hello, The following change makes the google page unavailable here. why would you change it since it still works ? Patch: http://cvs.winehq.org/patch.py?id=12121 Old revision New revision Changes Path 1.2 1.3 +1 -1 lostwages/news/2004040701.xml - (23 April 2004 - <a href="http://66.102.7.104/search?q=cache:sDglkGiMh7sJ:www-106.ibm.com/develo perworks/linux/library/l-wine/%3Fca%3Ddgr-lnxw11FineWine">Google's cache</a> still has a copy, but it's been removed from IBM's website.) + (23 April 2004 - <a href="http://www.google.com/search?q=cache:sDglkGiMh7sJ:www-106.ibm.com/deve loperworks/linux/library/l-wine/%3Fca%3Ddgr-lnxw11FineWine Cordialement, PETREOLLE Sylvain #DMIT/DATO/PSI/OPS (exploit at gefco dot fr ) +33 (0)1 49 05 29 29 "1+1=3" Bernard Werber ---------------------------------------------------------------------------- Ce message ainsi que toutes pièces jointes (le "message") sont confidentiels et sont exclusivement destinés à l'usage de la personne à laquelle ils sont adressés. Tout point de vue ou toute opinion contenus dans ce message expriment la pensée personnelle de leur auteur et ne représentent pas nécessairement la position des sociétés du Groupe GEFCO. Si vous n'êtes pas la personne à laquelle ce message est destiné, veuillez noter que vous avez reçu cet e-mail par erreur et qu'il vous est strictement interdit d'utiliser, de diffuser, de transférer, d'imprimer ou de copier ce message. Si vous avez reçu ce message par erreur, merci de contacter la personne qui vous l'a adressé et de l'effacer immédiatement. Les sociétés du Groupe GEFCO déclinent toute responsabilité en cas d'altération, de modification, d'édition, de diffusion sans autorisation de ce message ou en cas d'affection de ce message par un virus. This message and any attachments (the "message") are confidential and intended solely for the use of the individual to whom they are addressed. Any views or opinions presented are solely those of the author and do not necessarily represent those of the GEFCO Group of Companies. If you are not the intended recipient, be advised that you have received this email in error and that any use, dissemination, forwarding, printing, or copying of this message is strictly prohibited. If you have received this message in error please contact the sender and delete the message immediately. The GEFCO Group of Companies shall not be liable for the message if altered, changed, falsified, edited, diffused without authorization or affected by any virus. ----------------------------------------------------------------------------

1 0

Problem with GetVolumeInformation
by Chris McKenzie April 23, 2004

April 23, 2004

The refers to the 2004-03-09 ChangeLog, specifically: * dlls/kernel/Makefile.in, dlls/kernel/volume.c, dlls/ntdll/file.c, documentation/samples/config, documentation/wine.conf.man, files/drive.c, include/drive.h, include/winnt.h: Reimplemented GetVolumeInformation and SetVolumeLabel; volume label and serial number are now stored in the filesystem instead of in the config file (partly based on a patch by Eric Pouech). GetVolumeInformation's new implementation has 2 problems: 1. For software that was previously registered with a "made up" serial number via the "serial=" directive in the config file, the software (for me Altera MaxPlus) will no refuse to run claiming a change in serial. 2. If the function fails to claim a serial number (What actually _did_ happen with me) it will fall back on *serial=0. In my opinion this does not seem reasonable. I propose two options for wineconfig. First, one that says UseConfigSerial = [Y/N] Which will use the serial number in the config file and ignore the one on disk (This is useful for people that registered their software before the new function implementation) And the second option will be to reinstate the serial= for the drive descriptions. This should be the fallback serial number if the function fails, instead of the value *serial=0. I made a quick hack to return the value that I needed - however, I have not dealt with the kernel or the wine code extensively so I cannot put for a patch that does this at this time. However, I think this is a much more sensible design decision. Thank you. Sincerely, Christopher J. McKenzie cjm(a)ucdavis.edu

2 1

MinGW --mno-align-double
by Dimitrie O. Paun April 23, 2004

April 23, 2004

======= From: Hans Leidekker <hans(a)it.vu.nl> To: wine-devel(a)winehq.com <wine-devel(a)winehq.com> On Saturday 27 March 2004 02:20, Dmitry Timoshkov wrote: > generated.c:542: Test failed: TYPE_ALIGNMENT(*(LPWIN32_STREAM_ID)0) == 8 > (expected 4) generated: 842 tests executed, 0 marked as todo, 1 failure. > * NOTE: the same test compiled by cl.exe (MS compiler) works as expected, > * so that's most likely a gcc bug. (gcc version 3.3.1 (cygming special)) Same here (GCC 3.2.3 (mingw special 20030504-1)). It appears that this compiler defaults to --malign-double. When I specify --mno-align-double the test succeeds. ======= Alexandre, can we add this flag to our makefiles? If so, I can produce a patch... -- Dimi.

3 2

Re: Winetest: outstanding issues (BUILDING)
by Dimitrie O. Paun April 23, 2004

April 23, 2004

> From: Paul Millar <paulm(a)astro.gla.ac.uk> > One way this could work is if the general format for the BUILD_ID is > YYYYMMDDhhmm [ "-" <tag> [ "-" <tag> ] ...] > where <tag> is any metadata keyword that must not include a "-" Right, I have reached the exact same conclusion over lunch myself. So here is a fresh round of specifications: BUILDING (Paul & Kevin) ----------------------- A.The build process will produce 2 files: <name> <name>.sig where: <name>: the file being published <name>.sig: the GPG signature B. Naming <program>-<build>-<publisher>-<tools>[-<misc>].<ext> where: <program>: program name (i.e. "winetest", "winrash") <build>: build ID (for winetest, it's YYYYMMDDhhmm) <publisher>: publisher ID (i.e. "paul", "kevin") <tools>: tool chain used (i.e. "mingw", "msvc") <mic>: any other information the publisher wishes this can be used to support different builds for the same <build>, say with different flags, patches, etc. <ext>: appropriate extension (i.e. "zip", "exe") C. Publishing After the files have been successfully uploaded to the server (for winetest, this will be sf.net) a publish event will be sent to WineHQ, with the following form: http://test.winehq.org/service?publish=true&url=<url> where: <url>: is the URL to the <name> D. Winetest specific information - <program> = "winetest" - <publisher> = "paul" or "kevin" - <tools> = "mingw" (but a msvc would be good too) - <ext> = "zip" - nightly build will use sources as of 5AM EST - the <build> will use same time converted to UTC so it will mosty be YYYYMMDD1000, but it may be different when daylight-savings is in effect - you should use cvs co -D "YYYY-MM-DD 5:00 EST" to check out the code, and use date -u +%Y%m%d%H%D -d "YYYY-MM-DD 5:00 EST" to generate the <build> (I may have the format wrong, I don't have a reference handy) - the .zip file contains two files: winetest.exe: the object of this exercise winetest.info: metadata information about the build, as the tools that where used, special flags, mentions any special patches, etc. The information in here will not be automatically parsed, but rather is intended to help the developers debug problems. NOTE: should we include all the metadata present in <name> is the names of the .exe and .info? I would personally like to, in order to avoid any possible confusion on the tester's box. OK, did I miss anything? I will shotly send emails describing the other stages, I did not include them here, since this email is getting a bit long. The good thing is that there are no changes to Chris' winrash, there are a few changes (trivial) to winetest.exe, and dissect&gather, and some to service.cgi. But I'll cover those in my next emails. -- Dimi.

4 4

Microsoft: "We'd have been dead a long time ago without the Windows APIs"
by Mike Hearn April 23, 2004

April 23, 2004

http://news.zdnet.co.uk/0,39020330,39152686,00.htm At least they're honest about it - I especially love this quote: In the commission's report, Contorer's email was also blunt about Microsoft's shortcomings. "It is this switching cost that has given the customers the patience to stick with Windows through all our mistakes, our buggy drivers, our high TCO [total cost of ownership], our lack of a sexy vision at times, and many other difficulties ... Customers constantly evaluate other desktop platforms, [but] it would be so much work to move over that they hope we just improve Windows rather than force them to move. In short, without this exclusive franchise called the Windows API, we would have been dead a long time ago." Higher TCO huh? D'oh! thanks -mike

1 0

Re: project David? (Rein Klazes)
by Kevin F. Quinn April 23, 2004

April 23, 2004

A couple of statements on their web site: "One of the problems of the MS Windows OS is that it is subject to crash applications and itself. While studying the Microsoft Windows OS, we found the design flaw that causes this problem." _the_ design flaw? Marketing hype. It seems from later statements that they're talking about Microsoft's decision to relocate large sections of previously user-land code into the kernel (in particular, GDI), much commented on when NT went from 3.51 to 4.0. Later on, when comparing themselves to WINE, they state: "SpecOpS Labs believes, that [WINE] suffers from a major architectural flaw, which requires a major rewrite of the WINE code. The WINE project had been too faithful in reverse engineering the Windows Environment Subsystem, that it also inherited the architectural flaws in Windows. Among these flaws is when a problem is experienced by an application running in a Window, it can crash the whole operating system, causing it to either hang or reboot. " I think I'm right in saying an application using WINE _cannot_ crash the kernel, any more than any other user-land app. Also, from what I understand, it is _not_ WINE policy to re-implement Win32 bug-for-bug (since Microsoft do fix bugs, this is sensible). It looks like their project is on the search for large-scale funding. Assuming they can't distribute WINE code (good ol' GPL), they have a lot of work to do. Their "architecture" basically means full Win32 API implementation, with what looks like a priviledged kernel component ("WACS Driver") to interface to priviledged objects in the kernel. Incidentally, that kernel component better be rock-solid or they _will_ be able to kill the kernel from Windows apps in their subsystem... They do claim to have a "prototype" - I would guess that's either a modified WINE, or if newly written only implements a very small subset of the Win32 API. It will be interesting to see how quickly they get product to market, and when they do to see how much WINE code is used (which will need object-code comparisons, of course). One of their headline comments is: "David is not a reinvention of the wheel. It takes the best of breed pieces from previous attempts to simulate the Windows Subsystem, and integrates them into a single product." We shall see. There's a lot of "validation" garbage on the site, which to my mind isn't worth the paper it isn't printed on. It could well be they'll just raise loads of cash, then later crash and burn when they realise quite how much effort they need to spend to reimplement everything - or maybe when they realise they'll be violating the GPL if they lift WINE code. Kev.

1 0

Re: Re: i810 audio fixes 1/2
by chmorgan＠charter.net April 22, 2004

April 22, 2004

>like "play this wave file". The major remaining drawback >is latency; >click on button in app, hear sound 2 seconds later when >scheduler This is why we have sound servers like jack. Any delay is likely due to not having a fully callback based architecture in the app. >Maybe a better thing to do is rewrite arts/esd library >interfaces to >omit the daemons altogether. That way the daemons can go >away, but the >existing applications can be trivially migrated. I've heard that arts is being phased out from kde use in the near future. I'm not sure what might replace it though. Chris

6 7

Re: wine/include winternl.h
by Lionel Ulmer April 22, 2004

April 22, 2004

On Wed, Apr 21, 2004 at 05:21:36PM -0500, Alexandre Julliard wrote: > ChangeSet ID: 12085 > CVSROOT: /opt/cvs-commit > Module name: wine > Changes by: julliard(a)wine.codeweavers.com 2004/04/21 17:21:36 > > Modified files: > include : winternl.h > > Log message: > Mike McCormack <mike(a)codeweavers.com> > * add some values to the SYSTEM_INFORMATION_CLASS enumeration > > *************** > wint-sys-debug.diff (text/plain) > > Index: include/winternl.h > =================================================================== > RCS file: /home/wine/wine/include/winternl.h,v > retrieving revision 1.74 > diff -u -r1.74 winternl.h > --- include/winternl.h 16 Apr 2004 05:00:46 -0000 1.74 > +++ include/winternl.h 19 Apr 2004 06:57:43 -0000 > @@ -442,9 +442,36 @@ > SystemCacheInformation, (...) > > Patch: http://cvs.winehq.org/patch.py?id=12085 > > Old revision New revision Changes Path > 1.77 1.78 +33 -1 wine/include/winternl.h > Is this a VERY verbose Changelog or the new way to see the patch in the commit log :-) ? Lionel -- Lionel Ulmer - http://www.bbrox.org/

2 1

Winetest publishers
by Dimitrie O. Paun April 22, 2004

April 22, 2004

Hi Brian, A few more thoughts on the multiple publisher idea, and what we need to do on the server side. So, let's assume we have a number publishers out there. On server side we need to: 1. Maintain a pool of published winetest.exe's 2. Implement policy to serve from that pool (on a poll even) 1. The pool The pool will have to be maintained in a persistent manner to survive server reboots. It needs to get saved in a file. It should be managed as such: When a publish event happen we do: - check that the publisher id exists in publishers.txt file (1) - reject the event if not - check that it's really the publisher who publishes (2) - reject if not - check that the new BUILD_ID is >= than the current BUILD_ID - reject if not - if greater, reset the pool - add the published test to the pool - return 'OK' NOTES: (1) I suggest this format for publishers.txt <publisher-id>:<publisher-base-url> (2) We need to check that someone doesn't want to polute the pool. I suggest checking for the existance of: <publisher-base-url>/<published-file> 2. The Policy The sky's the limit. For now, we can just pick one at random from the pool. -- Dimi.

2 1

Invalid curdrive and curdir
by jeffl April 22, 2004

April 22, 2004

All, I am tracking a problem where an install program with an exception trying to access 0x00000000 memory. I have tracked the problem down to files/drive.c in DRIVE_GetCurrentDirectory where dos_cwd contains 0 and is used in ret = strlenW(dos_cwd) + 3; /* length of WHOLE current directory */ It seems that DRIVE_GetCurrentDrive returns a value greater than MAX_DOS_DRIVES (25) and then return DOSDrives[drive].dos_cwd; where dos_cwd=0; pTask->curdrive and pTask->curdir appear not to be set or invalid. Now it seems that a check for null ptr in DRIVE_GetCurrentDirectory is order ie. UINT ret; LPCWSTR dos_cwd = DRIVE_GetDosCwd( DRIVE_GetCurrentDrive() ); static const WCHAR driveA_rootW[] = {'A',':','\\',0}; if (dos_cwd == 0) { ERR("dos_cwd returned 0 instead of pointer to current working directory\n"); dos_cwd = "\0"; /* make a valid string for strlenW */ } ret = strlenW(dos_cwd) + 3; /* length of WHOLE current directory */ if (ret >= buflen) return ret + 1; but what would be a good value for dos_cwd? I am interested in what causes the pTask->curdrive to be greater than 25 and curdir to be set to ???????? as shown below. Wine-dbg>p *pTask {hNext=0, ss_sp=0, nEvents=0, priority=0, unused1=0, hSelf=4143, hPrevInstance=0, unused2=0, ctrlword8087=0, flags=16, error_mode=0, version=1024, hInstance=0, hModule=0, hQueue=4199, hParent=0, signal_flags=0, sighandler=0x00000000, userhandler=0x00000000, discardhandler=0x00000000, int0=0x00000000, int2=0x00000000, int4=0x00000000, int6=0x00000000, int7=0x00000000, int3e=0x00000000, int75=0x00000000, compat_flags=0, unused4="", teb=0x4002f000, unused5="", hPDB=4151, dta=272040064, curdrive=254, curdir="????????", nCmdShow=1, hYieldTo=0, dlls_to_init=0, hCSAlias=4167, thunks={0, 21584, 0, 8, 16, 0, 0, 0, 24, 0, 0, 0, 32, 0, 0, 0, 40, 0, 0, 0, 48, 0, 0, 0, 56, 0, 0, 0, 0, 0, 0, 0}, module_name="", magic=17492, hEvent=0x0000000c, pdb={int20=8397, nextParagraph=0, reserved1=0, dispatcher="?", savedint22=0x00000000, savedint23=0x00000000, savedint24=0x00000000, parentPSP=0, fileHandles="????????????????????", environment=4159, saveStack=0, nbFiles=20, fileHandlesPtr=272039960, hFileHandles=0, reserved3={0, 0, 0, ...}, ...}...} Wine-dbg> Jeff Latimer

1 0

Anyone working on WSALookupServiceBegin() and WSALookupServiceNext?
by Robbert Xerox April 22, 2004

April 22, 2004

Simple question : is anyone working on WSALookupServiceBegin() and WSALookupServiceNext() in winsock? ____________________________________________________________ Yahoo! Messenger - Communicate instantly..."Ping" your friends today! Download Messenger Now http://uk.messenger.yahoo.com/download/index.html

1 0

richedit - anybody trying to do it ?
by Krzysztof Foltman April 22, 2004

April 22, 2004

Hi, Just out of curiosity: is anyone already actively developing a (supposedly) fully featured richedit control ? The current version looks like a plain editor combined with RTF reader... -- Chris

5 5

speed up oleaut32: vartest
by Chris Morgan April 21, 2004

April 21, 2004

*dlls/oleaut32/tests/vartest.c: Chris Morgan<cmorgan(a)alum.wpi.edu> Skip over most of the unused variant types to speed up the variant tests. This look like a reasonable way to speed up the variant tests without losing testing functionality? Chris

1 1

RE: Fix for GetFileTime API - Incorrect Modified/Last access Time .
by Krishna Murthy April 21, 2004

April 21, 2004

We have Team Developer Product for MS Windows which has its own language and in the process of verifying on Wine. Few of test cases working on Windows are failing in Wine. Hence I have submitted this patch. Yes, since Pre-1970 times are not going to work on Unix anyway we could release note this in our product. Thanks for your input on this. Krishna -----Original Message----- From: Alexandre Julliard [mailto:julliard(a)winehq.org] Sent: Wednesday, April 21, 2004 6:16 PM To: Krishna Murthy Cc: wine-devel(a)winehq.org Subject: Re: Fix for GetFileTime API - Incorrect Modified/Last access Time. Krishna Murthy <Krishna.Murthy(a)guptaworldwide.com> writes: > Change of argument type DWORD to LONGLONG of function > RtlSecondsSince1970ToTime() and RtlSecondsSince1980ToTime() to accept > negative values. These are Windows APIs, you cannot simply change their arguments, we have to be compatible. Pre-1970 times are not going to work right on Unix anyway; what do you need them for? -- Alexandre Julliard julliard(a)winehq.org

1 0

winetest oddity - oleaut32: VarTest
by Chris Morgan April 21, 2004

April 21, 2004

I've been looking at vartest, specifically to speed it up since the tests I added for VarMod() are the cause of its current excessive length. Something appears to be a bit different with mingw compiled tests vs. MSVC compiled tests. Under wine these tests run just fine, three fixmes popup. I get two failures when running the oleaut32 vartest on my XP machine. This is compiled with vc6sp5+processor pack. Running the mingw compiled test from Paul Millar, http://www.astro.gla.ac.uk/users/paulm/WRT/CrossBuilt/winetest-20040421-100…, there are thousands of failures for the same test. Anyone have a clue what might be causing this? One thing that came to mind was that there are int64's used in the ole code, the test won't compile cleanly under vc6 without the processor pack. Maybe mingw is messing up the int64's? Chris

1 0

Re: Fix for GetFileTime API - Incorrect Modified/Last access Time.
by Alexandre Julliard April 21, 2004

April 21, 2004

Krishna Murthy <Krishna.Murthy(a)guptaworldwide.com> writes: > Change of argument type DWORD to LONGLONG of function > RtlSecondsSince1970ToTime() and RtlSecondsSince1980ToTime() to accept > negative values. These are Windows APIs, you cannot simply change their arguments, we have to be compatible. Pre-1970 times are not going to work right on Unix anyway; what do you need them for? -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: Wine detect patch (ignore previous whitespace patch)
by Ferenc Wagner April 21, 2004

April 21, 2004

Jakob Eriksson <jakov(a)vmlinux.org> writes: > (Sorry about the linespace noise after print_version(), > emacs seems to change stuff behind my back when I save.) Hi Jakob, You'd better educate Emacs about your ways, otherwise all your patches will be full of whitespace changes making them very hard to review. The indentation isn't consistent, but if you want to fix this, do so in a separate patch, please! > + static const TCHAR wine[128] = {"Wine"}; I think we could use static const char wine[] = "Wine"; instead. No wide characters in winetest for now. But see below! > + if (running_on_wine ()) { /* Make a an obvious signature for Wine */ > + ext = 1; > + > + ver.dwMajorVersion = 1; > + ver.dwMinorVersion = 1; > + ver.dwBuildNumber = 1; > + ver.dwPlatformId = 1; > + strncpy (ver.szCSDVersion, wine, sizeof (wine)); > + } > + I think we can do better. A single line like xprintf (" bRunningUnderWine=%d\n", running_under_wine ()); as the first line of print_version() would do a perfect job without masking anything else. -- Feri.

2 1

Re: Winetest results uploaded... where?
by Ivan Leo Murray-Smith April 21, 2004

April 21, 2004

On http://test.winehq.org/data/200404192305/ the "next build" and "previous build" buttons are broken. I get a error message, saying "You don't have permission to access /site on this server." Ivan.

4 12

Re: White-space patch for programs/winetest/main.c
by Alexandre Julliard April 21, 2004

April 21, 2004

Jakob Eriksson <jakov(a)vmlinux.org> writes: > This is also an example of how the file can store its indent > preference in a comment. > > I will run this when I'm done editing the file: > FILENAME=main.c && `grep indent $FILENAME` $FILENAME I don't think that's a good idea at all. It would require everybody to do that every time they make a change to that file, otherwise there will be plenty of spurious whitespace changes and conflicts. Please don't do that. -- Alexandre Julliard julliard(a)winehq.org

2 2