November 2002 - Wine-devel - list.winehq.org

Fix for bug #688 (documentation building)
by Vincent Béron 10 Nov '02

10 Nov '02

This one I'm not sure which one is better. So I include both patches, but only one should be applied. Functionnally, both produce the exact same html files, although the fix1 creates (by the system's db2html) a directory with images (which are unused). So I'd tend to go with fix2. Please comment, and I'll post the good one to wine-patches. Vincent

4 6

RFC: rpcss skeleton prototype #2
by Greg Turner 10 Nov '02

10 Nov '02

This one implements "pseudo-fork" behavior, a main workhorse loop, pseudo-debug log emitter, and command-line interpretation. In particular I'm looking for comments/criticism/flames/etc for: o my handling of NO_NAMELESS_STRUCT... works for me, but how correct will this be, say, under MSVC with Microsoft headers? o I differentiate between "server mode" and "lazy mode"... In lazy mode I assume that I'll have at least one incoming connection (if not I time out after 20 minutes); after that, once all endpoints are unmapped (all RPC-specific code is unimplmented, but this is what RPCSS_empty() is supposed to detect) I time out after a short delay (20 sec.) In server mode, I just wait around forever until I'm told to terminate (unimplemented, will be a command-line arg in the tradition of "wineserver -k", processed via the pipe)... reasonable? o Pseudo-debug log emitter: should I be using stderr? is this basically an OK approach? What is the hex number that most real wine debug logs begin with? I'd like to look the same. o Make/autoconf BS o the pseudo-fork crud (RPCSS_spork() and the /^ cmd-line arg.) should I be worried about accidentally invoking the real rpcss.exe? is this an unacceptably stupid approach in general? thanks for your help. diff -ur -x CVS -x 'bigdif*' ../wine.test/configure.ac ./configure.ac --- ../wine.test/configure.ac 2002-11-03 11:50:15.000000000 -0600 +++ ./configure.ac 2002-11-01 18:53:32.000000000 -0600 @@ -1501,6 +1501,7 @@ programs/regedit/Makefile programs/regsvr32/Makefile programs/regtest/Makefile +programs/rpcss/Makefile programs/uninstaller/Makefile programs/view/Makefile programs/wcmd/Makefile diff -ur -x CVS -x 'bigdif*' ../wine.test/programs/Makefile.in ./programs/Makefile.in --- ../wine.test/programs/Makefile.in 2002-11-03 11:50:15.000000000 -0600 +++ ./programs/Makefile.in 2002-11-01 18:53:18.000000000 -0600 @@ -17,6 +17,7 @@ regedit \ regsvr32 \ regtest \ + rpcss \ uninstaller \ view \ wcmd \ @@ -39,6 +40,7 @@ progman \ regedit \ regsvr32 \ + rpcss \ uninstaller \ wcmd \ wineconsole \ @@ -56,6 +58,7 @@ progman \ regedit \ regsvr32 \ + rpcss \ uninstaller \ wcmd \ wineconsole \ @@ -67,6 +70,7 @@ # Symlinks to apps that we want to run from inside the source tree SYMLINKS = \ + rpcss.exe \ wineconsole.exe \ winedbg.exe @@ -115,12 +119,16 @@ # Rules for symlinks +rpcss.exe$(DLLEXT): rpcss/rpcss.exe$(DLLEXT) + $(RM) $@ && $(LN_S) rpcss/rpcss.exe$(DLLEXT) $@ + wineconsole.exe$(DLLEXT): wineconsole/wineconsole.exe$(DLLEXT) $(RM) $@ && $(LN_S) wineconsole/wineconsole.exe$(DLLEXT) $@ winedbg.exe$(DLLEXT): winedbg/winedbg.exe$(DLLEXT) $(RM) $@ && $(LN_S) winedbg/winedbg.exe$(DLLEXT) $@ +rpcss/rpcss.exe$(DLLEXT): rpcss wineconsole/wineconsole.exe$(DLLEXT): wineconsole winedbg/winedbg.exe$(DLLEXT): winedbg --- /dev/null 1969-12-31 18:00:00.000000000 -0600 +++ ./programs/rpcss/Makefile.in 2002-11-03 18:51:44.000000000 -0600 @@ -0,0 +1,15 @@ +TOPSRCDIR = @top_srcdir@ +TOPOBJDIR = ../.. +SRCDIR = @srcdir@ +VPATH = @srcdir@ +MODULE = rpcss.exe +APPMODE = cui +IMPORTS = user32 kernel32 + +C_SRCS = \ + rpcss_main.c + +(a)MAKE_PROG_RULES@ + +### Dependencies: + --- /dev/null 1969-12-31 18:00:00.000000000 -0600 +++ ./programs/rpcss/README 2002-11-03 18:52:12.000000000 -0600 @@ -0,0 +1,23 @@ +rpcss README + +rpcss, copyright 2002, Greg Turner +rpcss is to be distributed under the LGPL/X11-MIT licenses +See the Wine License for further information. + +Wine needs a server whose role is somewhat like that +of rpcss.exe in windows. + +Enjoy. + +KNOWN BUGS / TODO: + + o Service hooks are unimplemented (if you bother to implement + these, also implement net.exe, at least for "net start" and + "net stop" (should be pretty easy I guess, assuming the rest + of the services API infrastructure works (which net.exe should + assume, even if the API is lean or bugged))). + + o Is supposed to use RPC, not random kludges, to map endpoints. + + o Who knows? Whatever (sane) things rpcss does, we ought to at + least think about doing... but what /does/ it do? --- /dev/null 1969-12-31 18:00:00.000000000 -0600 +++ ./programs/rpcss/rpcss.h 2002-11-03 18:50:26.000000000 -0600 @@ -0,0 +1,43 @@ +/* + * Copyright 2002 Greg Turner <gmturner007(a)ameritech.net> + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + */ + +#ifndef __WINE_RPCSS_H +#define __WINE_RPCSS_H + +#include <stdio.h> + +#include "windows.h" + +/* yet another crude debug logger substitute (just what wine needed!) */ +#define RPCSS_DEBUG(args...) \ + do { \ + if (in_debug_mode) {\ + printf("rpcss.exe.so:%s: ", __FUNCTION__); \ + printf(args); \ + } \ + } while (FALSE) + +extern BOOL in_debug_mode; + +extern BOOL RPCSS_is_server(); +extern void RPCSS_be_server(); +extern void RPCSS_server_terminate(); + +#endif /* __WINE_RPCSS_H */ + +/* end of header */ --- /dev/null 1969-12-31 18:00:00.000000000 -0600 +++ ./programs/rpcss/rpcss_main.c 2002-11-04 01:16:29.000000000 -0600 @@ -0,0 +1,308 @@ +/* + * rpcss (main.c) + * + * Copyright 2002 Greg Turner <gmturner007(a)ameritech.net> + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + */ + +#include <assert.h> + +#include "rpcss.h" +#include "winnt.h" + +/* in seconds */ +#define RPCSS_LONG_LAZY_TIMEOUT 1200 +#define RPCSS_SHORT_LAZY_TIMEOUT 20 + +/* are we the "fork child"? */ +static BOOL secret_fork_child; + +/* do we dump debug info to the console? */ +static BOOL in_debug_mode; + +/* are we in server mode, with no timeout? */ +static BOOL in_server_mode; + +/* was server mode terminated? */ +static BOOL server_mode_was_terminated; + +/* when do we just give up and bail? (UTC) */ +static SYSTEMTIME lazy_timeout_time; + +#if defined(NONAMELESSSTRUCT) + #define FILETIME_TO_ULARGEINT(filetime, ularge) \ + ( ularge.s.LowPart = filetime.dwLowDateTime, \ + ularge.s.HighPart = filetime.dwHighDateTime, \ + ularge ) /* r-value */ + #define ULARGEINT_TO_FILETIME(ularge, filetime) \ + ( filetime.dwLowDateTime = ularge.s.LowPart, \ + filetime.dwHighDateTime = ularge.s.HighPart, \ + filetime ) /* r-value */ +#else + #define FILETIME_TO_ULARGEINT(filetime, ularge) \ + ( ularge.LowPart = filetime.dwLowDateTime, \ + ularge.HighPart = filetime.dwHighDateTime, \ + ularge ) /* r-value */ + #define ULARGEINT_TO_FILETIME(ularge, filetime) \ + ( filetime.dwLowDateTime = ularge.LowPart, \ + filetime.dwHighDateTime = ularge.HighPart, \ + filetime ) /* r-value */ +#endif /* NONAMELESSSTRUCT */ + +#define TEN_MIL 10000000LL + +/* returns time remaining in seconds */ +long RPCSS_GetLazyTimeRemaining() +{ + SYSTEMTIME st_just_now; + FILETIME ft_jn, ft_ltt; + ULARGE_INTEGER ul_jn, ul_ltt; + + assert(! in_server_mode); /* then why are you calling? */ + + GetSystemTime(&st_just_now); + SystemTimeToFileTime(&st_just_now, &ft_jn); + FILETIME_TO_ULARGEINT(ft_jn, ul_jn); + + SystemTimeToFileTime(&lazy_timeout_time, &ft_ltt); + FILETIME_TO_ULARGEINT(ft_ltt, ul_ltt); + + return (ul_ltt.QuadPart - ul_jn.QuadPart) / TEN_MIL; +} + +void RPCSS_SetLazyTimeRemaining(long seconds) +{ + SYSTEMTIME st_just_now; + FILETIME ft_jn, ft_ltt; + ULARGE_INTEGER ul_jn, ul_ltt; + + RPCSS_DEBUG("(seconds == %ld)\n", seconds); + + assert(! in_server_mode); /* then why are you calling? */ + assert(seconds >= 0); /* negatives are not allowed */ + + GetSystemTime(&st_just_now); + SystemTimeToFileTime(&st_just_now, &ft_jn); + FILETIME_TO_ULARGEINT(ft_jn, ul_jn); + + /* we want to find the time ltt, s.t. ltt = just_now + seconds */ + ul_ltt.QuadPart = ul_jn.QuadPart + seconds * TEN_MIL; + + /* great. just remember it */ + ULARGEINT_TO_FILETIME(ul_ltt, ft_ltt); + if (! FileTimeToSystemTime(&ft_ltt, &lazy_timeout_time)) + assert(FALSE); +} + +#undef FILETIME_TO_ULARGEINT +#undef ULARGEINT_TO_FILETIME +#undef TEN_MIL + +BOOL RPCSS_is_server() +{ + return in_server_mode; +} + +void RPCSS_be_server() +{ + RPCSS_DEBUG("\n"); + in_server_mode = TRUE; +} + +void RPCSS_server_terminate() +{ + assert(in_server_mode); + server_mode_was_terminated = TRUE; +} + +BOOL RPCSS_work() +{ + RPCSS_DEBUG("\n"); + Sleep(1000); + return FALSE; /* we didn't do anything */ +} + +BOOL RPCSS_empty() +{ + BOOL rslt = TRUE; + + /* FIXME */ + + RPCSS_DEBUG("%s\n", rslt ? "TRUE" : "FALSE" ); + return rslt; +} + +BOOL RPCSS_is_ready_to_die() +{ + return + ( + RPCSS_empty() && + ( + (in_server_mode && server_mode_was_terminated) || + ( + (! in_server_mode) && + (RPCSS_GetLazyTimeRemaining() <= 0) + ) + ) + ); +} + +void RPCSS_main_loop() { + BOOL did_something_new; + + RPCSS_DEBUG("\n"); + + for (;;) { + did_something_new = FALSE; + + while ( (! did_something_new) && (! RPCSS_is_ready_to_die()) ) + did_something_new = (RPCSS_work() || did_something_new); + + if ((! did_something_new) && RPCSS_is_ready_to_die()) + break; /* that's it for us */ + + if (did_something_new) { + /* if we were in server mode, but terminated, then, + because we did something new, fall back to lazy mode */ + if (in_server_mode && server_mode_was_terminated) + in_server_mode = server_mode_was_terminated = FALSE; + + /* if we aren't in server mode, set the lazy timeout */ + if (! in_server_mode) + RPCSS_SetLazyTimeRemaining(RPCSS_SHORT_LAZY_TIMEOUT); + } + } +} + +BOOL RPCSS_ProcessArgs( int argc, char **argv ) +{ + int i; + char *c; + + for (i = 1; i < argc; i++) { + c = argv[i]; + while (*c == ' ') c++; + if ((*c != '-') && (*c != '/')) + return FALSE; + c++; + switch (*c++) { + case 's': + case 'S': + in_server_mode = TRUE; + break; + case 'd': + case 'D': + in_debug_mode = TRUE; + break; + case '^': + /* "secret" hidden command-line arg means we are the "fork child" */ + secret_fork_child = TRUE; + break; + default: + return FALSE; + break; + } + while (*c == ' ') c++; + if (*c != '\0') return FALSE; + } + + return TRUE; +} + +void RPCSS_Usage() +{ + printf("\nWine RPCSS\n"); + printf("\nsyntax: rpcss [/d] [/s]\n\n"); + printf(" /d: debug mode (print debug messages to std. output)\n"); + printf(" /s: server mode (don't time out)\n\n"); +} + +void RPCSS_Spork() +{ + /* Because we are a wine app, we can't fork. This is + a trick to invoke ourselves, so we can provide a + similar effect of creating a "background" process */ + + PROCESS_INFORMATION pi; + STARTUPINFO si; + char *oldcmdline, *newcmdline; + + /* create a string to invoke ourselves */ + oldcmdline = GetCommandLineA(); + newcmdline = LocalAlloc(LPTR, strlen(oldcmdline) + 10); /* more than enough */ + assert(newcmdline); /* must succeed */ + + sprintf(newcmdline, "%s /^", oldcmdline); + + ZeroMemory(&pi, sizeof(PROCESS_INFORMATION)); + ZeroMemory(&si, sizeof(STARTUPINFO)); + si.cb = sizeof(STARTUPINFO); + + if (!CreateProcess( + NULL, // executable + newcmdline, // command line + NULL, // process security attributes + NULL, // primary thread security attributes + FALSE, // no inherit handles + 0, // creation flags + NULL, // use parent's environment + NULL, // use parent's current directory + &si, // STARTUPINFO pointer + &pi // PROCESS_INFORMATION + )) + assert(FALSE); + + LocalFree(newcmdline); +} + +int main( int argc, char **argv ) +{ + /* + * We are invoked as a standard executable; typically, we act in a + * "lazy" manner. We open up our pipe, and hang around until we have + * nothing left to do, and then silently terminate. When we're needed + * again, rpcrt4.dll.so will invoke us automatically. + */ + + if (!RPCSS_ProcessArgs(argc, argv)) { + RPCSS_Usage(); + return 1; + } + + if (! in_debug_mode) { + if (secret_fork_child) + FreeConsole(); + else { + /* pseudo-fork */ + RPCSS_Spork(); + return 0; + } + } else if (secret_fork_child) { + /* this should never happen... perhaps they typed /d /^ on cmd line */ + RPCSS_Usage(); + return 2; + } + + if (! in_server_mode) + /* we want to wait for something to happen, and then + timeout when no activity occurs... but we will be + extra patient on the first wait */ + RPCSS_SetLazyTimeRemaining(RPCSS_LONG_LAZY_TIMEOUT); + + RPCSS_main_loop(); + + return 0; +} -- gmt "The purpose of government is to rein in the rights of the people" --President Bill Clinton, MTV interview, 1993

7 15

Re: help support
by Sylvain Petreolle 10 Nov '02

10 Nov '02

CHM is a special extension of IE. The patch to make any windows compatible with CHM (except Win 3.1) is downloadable at http://www.microsoft.com/downloads/release.asp?ReleaseID=30328 To handle and generate CHM I found a french page (tools are in english). http://www.jurixt.com/informatique/hlp_compile_aide.htm Utility to decompile CHM and many many more : http://www.keyworks.net/keytools.htm Google power ! ___________________________________________________________ Do You Yahoo!? -- Une adresse @yahoo.fr gratuite et en français ! Yahoo! Mail : http://fr.mail.yahoo.com

1 0

Re: int21 move a8
by Jukka Heinonen 09 Nov '02

09 Nov '02

Again, no full review, just quickly browsed through the patch. I have a feeling that I have just scratched the surface regarding this patch series and it would need a full review before it can be accepted. Function 0x40 change in winedos/int21.c breaks DOS VGA text console. Function 0x3f handling drops WIN16_hread which used to check for segment size. That check seemed to be for running buggy Win16 programs. At least I wouldn't have courage to delete this kind of checks. -- Jukka Heinonen <http://www.iki.fi/jhei/>

1 0

Re: winamp patch A0
by Vincent Béron 09 Nov '02

09 Nov '02

Le sam 09/11/2002 à 11:22, Andreas Mohr a écrit : > Hi all, > > IMHO we should now switch to Win98 SE as default Windows version. > Too many programs expect at least Win98 to be running already. > > Winamp also checks the Version registry keys, so I added them to > winedefault.reg. > > Index: winedefault.reg > =================================================================== > RCS file: /home/wine/wine/winedefault.reg,v > retrieving revision 1.58 > diff -u -r1.58 winedefault.reg > --- winedefault.reg 25 Oct 2002 19:05:19 -0000 1.58 > +++ winedefault.reg 9 Nov 2002 16:20:36 -0000 > @@ -72,6 +72,9 @@ > "ProgramFilesDir"="C:\\Program Files" > "RegisteredOrganization"="Change preferred organization in ~/.wine/system.reg" > "RegisteredOwner"="Change preferred owner in ~/.wine/system.reg" > +"SubVersionNumber"=" A " > +"Version"="Windows 98" > +"VersionNumber"="4.10.2222" This doesn't seem the place to do it, since if we specify something different (in config) then the registry keys do not match, and some other app could possibly have some problems. Would it be possible to generate automatically those keys with a call to GetVersionEx or something similar? Vincent

3 2

Re: GetCurrentObject
by steve.lustbader＠philips.com 09 Nov '02

09 Nov '02

GetCurrentObject returns 0, but as far as it knows, it is returning successfully. This is because in the line case OBJ_BITMAP: ret = dc->hBitmap; break; dc->hBitmap is 0, although dc is a valid DC. The HDC that is passed to GetCurrentObject to get dc is obtained by calling CWindow::BegingPaint(PAINTSTRUCT *) from a class derived from ATL's CWindow. I'm guessing that this is where the problem is probably coming from - the app is written in C++ with ATL. -Steve I've seen this occur when wine frees a gdi handle it shouldn't. Check that the OBJ_BITMAP handle is really valid. Also wierd stuff can occur if the app expects a HDIB but we provide it a HBITMAP... or vice versa. Mike steve.lustbader(a)philips.com wrote: > My app calls GetCurrentObject with a valid HDC and OBJ_BITMAP. > It returns 0, however, which causes my app to crash. What further > info can I provide to help debug this? > > -Steve >

2 1

Fwd: Re: So lets say we do it
by Dimitrie O. Paun 09 Nov '02

09 Nov '02

---------- Forwarded Message ---------- Subject: Re: So lets say we do it Date: Sat, 9 Nov 2002 02:51:47 -0800 From: Kenny Smith <kenny(a)blindinstinct.com> To: dpaun(a)rogers.com I don't know how to use mail lists, so i though I might mail you In responce to your ideas about reorganizing the winehq site, i think there should be a site for development, like winehq, and a site for users. most users don't want to know about listview overhauls, etc. They wnat to know if their precious Quicken will run on linux or not. I'm sure a web design would be EASILY feasible. I'm a web designer (never done anything that big), and a 6-month linux user and there are tons of people like me who i'm sure would love to help. I use wine because I am somewhat computer-savvy (A+ certified savvy). I hope this message can be passed along, if you can't maybe you can tell me how to quickly... would google-groups let me?? kenny smith (AIM:standsolid) ------------------------------------------------------- -- Dimi.

1 0

re: Wine 0.8: VB compatibility !!
by Dan Kegel 09 Nov '02

09 Nov '02

Andreas Mohr wrote: > I keep encountering people mentioning that VB stuff has > lotsa problems on Wine. > While I really can't say that I like VB, a lot of people > seem to use it for critical tasks in companies on a > daily basis, so it sounds like perfect > VB compatibility would be utterly important. > IMHO improving that should be a goal of 0.8. Make sense. How about this: let's add a visual basic regression test suite to Wine. I'm not kidding! Ought to be pretty easy to hit at least the non-gui parts... - Dan

6 6

Re: wine button
by Mike Hearn 09 Nov '02

09 Nov '02

[ps i assumed by what was written you meant for this to go to wine-devel but didn't hit reply-to-all, sorry if it was meant for me only] > There's the application, and button with a wine bottle, that says > 'works with wine'. You could come up with quite a few jokes with that. heh, true. It doesn't look good with the first letter capitalized, at least imho, and probably wouldn't alter the impression non-initiates would get. I tried putting "windows emulator" underneath, but it looks a bit out of place and kind of reduces the impact. It also makes it harder to scale down, at 50% of the size you can still read the "works with wine" text, but not the "windows emulator" underneath. It might also piss off people who say "Wine is not an emulator" :) Actually, letting people make jokes about it might be a good thing - I doubt many people would actually take it literally and assuming they do find out, if they're talking and joking about it, they're more likely to remember it too. On Fri, 2002-11-08 at 15:35, Rick Romero wrote: > > On Fri, 2002-11-08 at 08:53, Mike Hearn wrote: > > Here is my first attempt. I'm hardly an ace designer, so feel free to > > criticize, tell me to junk it etc. > I think it looks great, but if I were a Windows app owner, how would my > potential users see it? > > There's the application, and button with a wine bottle, that says > 'works with wine'. You could come up with quite a few jokes with that. > > Of course we all know what it means, should the 'W' be capitalized to > differentiate it from alcohol? > > I don't intend to start a logo flame war, but I think perceptions are > everything. > > Good thing I'm only around for another 90 minutes today :P > > Rick > > > > > > > -- > -- > Rick Romero > IT Manager > Valeo, Inc. ph: 262.695.4841 > Sussex, WI. fax: 262.695.4850 > Rick(a)valeoinc.com > >

3 2

wine button
by Mike Hearn 09 Nov '02

09 Nov '02

Here is my first attempt. I'm hardly an ace designer, so feel free to criticize, tell me to junk it etc. I did try at first to make a browser style button, but it was very hard to make the wine logo fit in a way that didn't look odd, as it's largely a vertical type of logo. Also, is the Windows logo usable? I'd assume it's trademarked or something, but i've seen the 4 colour blocks used in for instance icon overlays in KDE. If it's thought that this is too large, I can either scale it, or try and make another from scratch that is more button sized. thanks -mike

7 10