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

Focus problems
by Pavel Roskin Oct. 6, 2003

Oct. 6, 2003

Hello! My patch for fixing the focus of some programs (notepad, winecfg) on startup hasn't been applied, so I'm guessing what the reason might be. One reason may be that the testing could have found another bug that might seem to be related. When Wine uses non-managed windows without desktop, they don't get the keyboard focus initially. It affects even programs what show active captions, such as Winemine. When Winemine is started, the keyboard events go to the original window. The same happens after the "About" dialog is run. But when the "About" dialog is dismissed, the main window of Winemine gets keyboard focus. It seems to be another example of not doing something because the initial window is "already focused". This is a separate problem, it affects a wider class of programs and doesn't affect Wine in the desktop mode. If fixing this is considered precondition to fixing all other focus problems, I'd like to know that. Another possible reason I can think of is because SetFocus is supposed to affect widgets before the application window is shown. That's something that needs testing. I've made a patch that only disables activation of the parent window if it's invisible, but still focuses the window for which SetFocus() was called. The patch is attached. Before I start time-consuming verification, I'd like to know if I going in the right direction. There is another, third focus problem. Windows should not cover modal dialogs. Shareware Windows Commander 5.0 is useless in the desktop mode because the main window covers the "nag screen" on startup. The workaround is to enter "foo" in the command line and press Enter - there will be an error message, and after it's dismissed, the "nag screen" will pop up. Again, it's a separate problem that I'm not trying to solve at this point. I'll appreciate an advise in which order to proceed and what kind of verification is expected for the changes in the focus code. -- Regards, Pavel Roskin

2 1

Protected mode interrupt problem
by Mike McCormack Oct. 6, 2003

Oct. 6, 2003

Hi All, I've run into an interrupt related problem with Quicken 2002's installer. Two 16bit programs launch, one after the other. Each program hooks INT 0, however since our interrupt tables are global to all tasks, the second task retrieves the interrupt vector that the first task set. The first task exits, and restores the original INT 0 vector. When the second task exits, tries to restore the INT 0 vector set by the first task, which now has an invalid selector, so it crashes. After a bit of discussion and writing a test program, we discovered that it appears that some interrupt vectors are task local, and some are global. The local ones appear to be stored in the TDB (task descriptor block), in offsets 0x26 - 0x4a. The attached patch fixes this, though we're not sure that it is correct for DOS programs. I have enclosed a test program that shows the problem. Mike Index: include/task.h =================================================================== RCS file: /cvstrees/crossover/office/wine/include/task.h,v retrieving revision 1.1.1.5 diff -u -r1.1.1.5 task.h --- include/task.h 12 Sep 2003 18:37:36 -0000 1.1.1.5 +++ include/task.h 7 Oct 2003 06:29:43 -0000 @@ -98,13 +98,13 @@ FARPROC16 sighandler WINE_PACKED; /* 26 Signal handler */ FARPROC16 userhandler WINE_PACKED; /* 2a USER signal handler */ FARPROC16 discardhandler WINE_PACKED; /* 2e Handler for GlobalNotify() */ - DWORD int0 WINE_PACKED; /* 32 int 0 (divide by 0) handler */ - DWORD int2 WINE_PACKED; /* 36 int 2 (NMI) handler */ - DWORD int4 WINE_PACKED; /* 3a int 4 (INTO) handler */ - DWORD int6 WINE_PACKED; /* 3e int 6 (invalid opc) handler */ - DWORD int7 WINE_PACKED; /* 42 int 7 (coprocessor) handler */ - DWORD int3e WINE_PACKED; /* 46 int 3e (80x87 emu) handler */ - DWORD int75 WINE_PACKED; /* 4a int 75 (80x87 error) handler */ + FARPROC16 int0 WINE_PACKED; /* 32 int 0 (divide by 0) handler */ + FARPROC16 int2 WINE_PACKED; /* 36 int 2 (NMI) handler */ + FARPROC16 int4 WINE_PACKED; /* 3a int 4 (INTO) handler */ + FARPROC16 int6 WINE_PACKED; /* 3e int 6 (invalid opc) handler */ + FARPROC16 int7 WINE_PACKED; /* 42 int 7 (coprocessor) handler */ + FARPROC16 int3e WINE_PACKED; /* 46 int 3e (80x87 emu) handler */ + FARPROC16 int75 WINE_PACKED; /* 4a int 75 (80x87 error) handler */ DWORD compat_flags WINE_PACKED; /* 4e Compatibility flags */ BYTE unused4[2]; /* 52 */ struct _TEB *teb; /* 54 Pointer to thread database */ Index: dlls/winedos/interrupts.c =================================================================== RCS file: /cvstrees/crossover/office/wine/dlls/winedos/interrupts.c,v retrieving revision 1.1.1.8 diff -u -r1.1.1.8 interrupts.c --- dlls/winedos/interrupts.c 12 Sep 2003 18:37:31 -0000 1.1.1.8 +++ dlls/winedos/interrupts.c 7 Oct 2003 06:29:47 -0000 @@ -25,6 +25,7 @@ #include "wine/winbase16.h" #include "thread.h" +#include "task.h" WINE_DEFAULT_DEBUG_CHANNEL(int); WINE_DECLARE_DEBUG_CHANNEL(relay); @@ -541,9 +542,42 @@ */ FARPROC16 DOSVM_GetPMHandler16( BYTE intnum ) { + TDB *pTask; + FARPROC16 proc = 0; + + pTask = TASK_GetCurrent(); + if (pTask) + { + switch( intnum ) + { + case 0x00: + proc = pTask->int0; + break; + case 0x02: + proc = pTask->int2; + break; + case 0x04: + proc = pTask->int4; + break; + case 0x06: + proc = pTask->int6; + break; + case 0x07: + proc = pTask->int7; + break; + case 0x3e: + proc = pTask->int3e; + break; + case 0x75: + proc = pTask->int75; + break; + } + if( proc ) + return proc; + } if (!DOSVM_Vectors16[intnum]) { - FARPROC16 proc = (FARPROC16)MAKESEGPTR( DOSVM_dpmi_segments->int16_sel, + proc = (FARPROC16)MAKESEGPTR( DOSVM_dpmi_segments->int16_sel, DOSVM_STUB_PM16 * intnum ); DOSVM_Vectors16[intnum] = proc; } @@ -558,9 +592,41 @@ */ void DOSVM_SetPMHandler16( BYTE intnum, FARPROC16 handler ) { + TDB *pTask; + TRACE("Set protected mode interrupt vector %02x <- %04x:%04x\n", intnum, HIWORD(handler), LOWORD(handler) ); - DOSVM_Vectors16[intnum] = handler; + + pTask = TASK_GetCurrent(); + if (!pTask) + return; + switch( intnum ) + { + case 0x00: + pTask->int0 = handler; + break; + case 0x02: + pTask->int2 = handler; + break; + case 0x04: + pTask->int4 = handler; + break; + case 0x06: + pTask->int6 = handler; + break; + case 0x07: + pTask->int7 = handler; + break; + case 0x3e: + pTask->int3e = handler; + break; + case 0x75: + pTask->int75 = handler; + break; + default: + DOSVM_Vectors16[intnum] = handler; + break; + } } #include <stdio.h> #include <windows.h> void myintfn(void) { } int PASCAL WinMain(HANDLE hInstance, HANDLE hPrevInstance, LPSTR lpCmdLine, int nCmdShow) { int old_seg, old_ofs, new_seg, new_ofs; char str[16]; /* get the old interrupt handler for int 0 */ _asm { push bx push es mov ax,0x3502 int 21h mov old_seg, es mov old_ofs, bx pop es pop bx } /* print it in a message box */ sprintf(str, "old %04x:%04x", old_seg, old_ofs ); MessageBox( 0, str, "Int 02", MB_OK); /* set the new interrupt handler */ _asm { push ds push dx mov ax,0x2502 push cs pop ds mov dx, offset myintfn int 21h pop dx pop ds } if( !hPrevInstance) WinExec("taskint.exe",SW_SHOW); /* fetch it back again */ _asm { push bx push es mov ax,0x3502 int 21h mov new_seg, es mov new_ofs, bx pop es pop bx } sprintf(str, "new %04x:%04x", new_seg, new_ofs ); MessageBox( 0, str, "Int 02", MB_OK); /* now set the original back again */ _asm { push ds push dx mov ax,0x2502 mov ds, old_seg mov dx, old_ofs int 21h pop dx pop ds } return 0; }

1 0

Wineconf - survey
by Jeremy White Oct. 6, 2003

Oct. 6, 2003

Hi everyone, I've put up a survey page here: http://www.winehq.org/wineconf/ I'd like to ask everyone who would come to Wineconf to go and enter their preferences. If you would not come to Wineconf regardless of where it was held, please do not fill out the survey. Let's hold this open through Friday, and then let's consider it final. Sound reasonable? Cheers, Jer

1 0

Re: Problem into the mail server ?
by Hannu Valtonen Oct. 6, 2003

Oct. 6, 2003

Duane Clark wrote: > If you don't want the emails sent to you, subscribe and then turn on the > option to "Disable mail delivery". No need, I subscribed to a bunch of lists and finally configured my Mozilla Mail's filters. Took me a grand total of five minutes. The traffic is hardly a problem since I'm sitting behind a cablemodem and would've read the posts on the mailinglists anyway. But thanks for the suggestion. - Hannu Valtonen Hannu.Valtonen(a)hut.fi

1 0

Unicode strings
by Kevin Koltzau Oct. 6, 2003

Oct. 6, 2003

I've come upon an instance where I need to create a large lookup table of unicode strings, and was curious what the recommended way to handle this was. Using MSVC I would simply declare the string as L"Text" I only see two options currently, either declare the string as (const WCHAR[]){'T','e','x','t'} or store as non-unicode and use WideCharToMultiByte/MultiByteToWideChar as appropriate As this is being implemented for a unicode-only function, it seems silly to convert the string, but declaring it as (const WCHAR[]){'T','e','x','t'} is a little unwieldy for the quantity of data that I need to declare

3 2

Re: Oleaut32 7
by Alexandre Julliard Oct. 6, 2003

Oct. 6, 2003

Jon Griffiths <jon_p_griffiths(a)yahoo.com> writes: > +/* The list of localised text data */ > +const VARIANT_TEXT VARIANT_LocalisedTextList[NUM_LOCALISED_LANGS] = > +{ > + { szCzechStrings, LANG_CZECH, 7, 16, 24, 32, 36 }, > + { szDanishStrings, LANG_DANISH, 5, 11, 14, 18, 22 }, > + { szGermanStrings, LANG_GERMAN, 5, 12, 15, 20, 24 }, > + { szEnglishStrings, LANG_ENGLISH, 5, 11, 14, 18, 22 }, > + { szSpanishStrings, LANG_SPANISH, 10, 16, 19, 22, 31 }, > + { szFinnishStrings, LANG_FINNISH, 5, 13, 19, 22, 31 }, > + { szFrenchStrings, LANG_FRENCH, 5, 10, 16, 24, 28 }, > + { szHungarianStrings,LANG_HUNGARIAN, 5, 11, 16, 20, 23 }, > + { szItalianStrings, LANG_ITALIAN, 5, 11, 14, 17, 20 }, > + { szDutchStrings, LANG_DUTCH, 5, 12, 15, 19, 24 }, > + { szNorseStrings, LANG_NORWEGIAN, 5, 11, 14, 18, 21 }, > + { szPolishStrings, LANG_POLISH, 7, 13, 17, 21, 24 }, > + { szSwedishStrings, LANG_SWEDISH, 5, 12, 15, 19, 22 } > +}; It would be better IMO to store that as string resources; you then get all the localisation mechanism for free, and people can translate them without digging into the code. -- Alexandre Julliard julliard(a)winehq.com

1 0

Re: Oleaut32 4
by Alexandre Julliard Oct. 6, 2003

Oct. 6, 2003

Jon Griffiths <jon_p_griffiths(a)yahoo.com> writes: > Changelog: > > Jon Griffiths <jon_p_griffiths(a)yahoo.com> > +dlls/oleaut32/i1.c dlls/oleaut32/i2.c dlls/oleaut32/i4.c > dlls/oleaut32/r4.c dlls/oleaut32/Makefile.in > dlls/oleaut32/oleaut32.spec > dlls/oleaut32/variant.c > Implement/document i1/ui1/i2/ui2/i4/ui4/r4 low-level calls I don't think we want to create a separate file for every variant type, especially since most of the functions are just one line. A single file with all the "simple" conversion functions would be enough. -- Alexandre Julliard julliard(a)winehq.com

1 0

Re: handle quoting on the command line in uninstaller
by Alexandre Julliard Oct. 6, 2003

Oct. 6, 2003

Mike McCormack <mike(a)codeweavers.com> writes: > ChangeLog: > * handle quoting on the command line in uninstaller Is there any reason we can't use argv[] here instead? -- Alexandre Julliard julliard(a)winehq.com

1 0

Re: Oleaut32 7
by Gregory M. Turner Oct. 6, 2003

Oct. 6, 2003

On Monday 06 October 2003 10:32 am, Jon Griffiths wrote: > Hi, > > Part 7 of oleaut32 updates. > > Cheers, > Jon > > License: X11 > > Changelog: > > Jon Griffiths <jon_p_griffiths(a)yahoo.com> > +dlls/oleaut32/bool.c dlls/oleaut32/oleaut32.spec > dlls/oleaut32/variant.c > Implement the boolean low-level calls Nice patch-set. I see there were some surprises, but good job nevertheless... extra kudos for doing laborious "boring" stuff and for providing in-source doc! Wish I was so prolific, as I am a no-patch bum lately.... -- gmt "The vice I am talking about is Pride or Self-Conceit: and the virtue opposite to it, in Christian morals, is called Humility. ... According to Christian teachers, the essential vice, the utmost evil, is Pride. Unchastity, greed, drunkenness, and all that, are mere flea-bites in comparison: it was through Pride that the devil became the devil: Pride leads to every other vice: it is the complete anti-God state of mind." --C. S. Lewis

1 0

Re: What does it take to run a third-party DLL under Wine?
by Dimitrie O. Paun Oct. 6, 2003

Oct. 6, 2003

On Mon, 6 Oct 2003, John Conneely wrote: > Hey, any ideas of where I might be able to hire a WINE guru to help with > this stuff? I haven't even used Wine before, and I'm looking for someone > that can help us set up our WINE environment, test to see if the DLL we are > trying to use will run under WINE, then create a wrapper for the DLL that is > created with WineLib. Using an RPC mechanism is fine, as long as it doesn't > involve a Windows OS box. Might you be available for a little moonlighting? > I don't think we're talking about a huge amount of work here... mostly the > feasibility analysis work and some pointers on how to setup the environment. > If not you, know any other WINE hackers that might be? Sure, I can take a look. I would also recommend the CodeWeavers guys (http://www.codeweavers.com) they've done this before a number of times, it should be an easy task for them. I'm not affiliated with them, but they are great folks! If you want me to help you get started, email me off the list. -- Dimi.

1 0