- Wine-Devel - list.winehq.org

For your consideration, here's a preview of some new seh_try_macros-type stuff...
by Gregory M. Turner Feb. 14, 2003

Feb. 14, 2003

Let me know what you think of these. So far, I like them better than the original versions, although in solving some limitations of the old macros, I've created some new ones. This version avoids ({ }) entirely, in favor of Ove's original thought, which was to use __LINE__; therefore, putting multiple __(finally|except) clauses in the same line probably breaks. Otherwise, these are better macros. They are slightly more easily understood (although still pretty darn confusing, especially in the debugger), probably more resiliant against compiler optimizations, and have much less awful performance. Best of all, they ought to properly re-raise exceptions after executing finally clauses. There is some cruft in there, btw, which will be fixed before I submit anything to patches. They still break break/continue, of course, since they still rely on __TRY / __EXCEPT blocks to do the dirty-work. You can run the original seh_macros.c test against these successfully, except for the "absurdly nested" ones which use macros and presumably would fail quite dramatically due to __LINE__ duplications. FWIW, compared to the degree of "pressing need" for these macros, I have already spent way too much time on them... But the old ones kind of sucked, and, frankly, it's a fun semantic puzzle that requires pusing the limits of good old cpp just a _bit_ further than may have been intended ... so, here it is. Note that I'm declaring a variable in the header file here. I guess, I should fix this by creating an internal API in ntdll? I never quite figured out whether global non-static variables in a dll are thread-local, or what... (same for lexically scoped static ones -- real clueless here, and building a new system so I don't have much doco's at my fingertips ATM) Sooo.... Anyone want to help me decide what's the best way to get a thread-local global variable instantiated? Probably, my previous attempt with TlsAlloc was semantically correct, but that was /way/ too fat a solution, I'd like something much quicker if its at all possible. Otherwise this won't be threadsafe, which seems like a bad call, esp. since such code will probably end up being used to guard against unacceptable resource leakage.. warning, cut-and-pasted this patch in chunks so it might be bugged... should be clear enough how to duplicate, however, if it fails... --- ../wine.vanilla/include/wine/exception.h 2003-01-29 21:50:14.000000000 -0600 +++ ./include/wine/exception.h 2003-02-09 01:33:45.000000000 -0600 @@ -21,8 +21,12 @@ #ifndef __WINE_WINE_EXCEPTION_H #define __WINE_WINE_EXCEPTION_H +/* WRC can't find setjmp.h. Does it not check /usr/include? */ +#if (!(defined(__WRC__))) + #include <setjmp.h> #include "windef.h" +#include "excpt.h" /* The following definitions allow using exceptions in Wine and Winelib code * @@ -59,6 +63,16 @@ * You can use them to leave a __EXCEPT block though. * * -- AJ + * + * Now we can implement the __try/__except/__finally magic of VC++ as standard (?) macros. + * this is a "good enough" implementation, it's incomplete, but should suffice for most uses. + * + * The macros require some symbols to be found in ntdll.dll.so, so link with that DLL if you + * use them. + * + * If you are using VC++ and this header, #define USE_COMPILER_EXCEPTIONS before you include it. + * + * -gmt */ /* Define this if you want to use your compiler built-in __try/__except support. @@ -146,8 +160,90 @@ extern DWORD __wine_finally_handler( PEXCEPTION_RECORD record, EXCEPTION_FRAME *frame, CONTEXT *context, LPVOID pdispatcher ); + +#if (!defined(NO_VC_SEH_MACROS)) + +#ifndef _WINE_ALREADY_DEFINED_SEH_MACROS__ +#define _WINE_ALREADY_DEFINED_SEH_MACROS__ + +/* fixme: should be renamed since these are used for except too */ +typedef void (*__p_seh_finally_executer) (); +typedef void (*__p_seh_dotry_executer) (__p_seh_finally_executer); +__p_seh_finally_executer __thread_outermost_handler_ptr; + +#define __SEH_LINEUNIQ_CAT2(x,y) __SEH_LINEUNIQ_ ## y ## x +#define __SEH_LINEUNIQ_CAT1(x,y) __SEH_LINEUNIQ_CAT2(x,y) + +#define __SEH_LINEUNIQ_NULLMACRO(x) x +#define __SEH_LINEUNIQ_MAGIC_LINE __SEH_LINEUNIQ_NULLMACRO( __LINE__ ) + +#define __SEH_LINEUNIQ_ID(x) __SEH_LINEUNIQ_CAT1( __SEH_LINEUNIQ_MAGIC_LINE , x ) + +#define __try \ + do { \ + void dotry(__p_seh_finally_executer fin) { \ + __label__ __seh_macro_exit_spot; \ + auto WINE_EXCEPTION_FILTER(__wine_seh_macro_exception_handler); \ + __TRY { \ + { \ + +#define __finally \ + } \ + __seh_macro_exit_spot: \ + if (0) goto __seh_macro_exit_spot; \ + } __EXCEPT( __wine_seh_macro_exception_handler ) { \ + } __ENDTRY \ + (*fin)(); \ + WINE_EXCEPTION_FILTER(__wine_seh_macro_exception_handler) { \ + (*fin)(); \ + return EXCEPTION_CONTINUE_SEARCH; \ + } \ + } \ + goto __SEH_LINEUNIQ_ID(DETOUR); \ + __SEH_LINEUNIQ_ID(RETOUR): \ + dotry( __thread_outermost_handler_ptr ); \ + } while (0); \ + goto __SEH_LINEUNIQ_ID(AFTERFINALLY); \ + auto void __SEH_LINEUNIQ_ID(Finally_Executer) (); \ + __SEH_LINEUNIQ_ID(DETOUR): \ + __thread_outermost_handler_ptr = & __SEH_LINEUNIQ_ID(Finally_Executer); \ + goto __SEH_LINEUNIQ_ID(RETOUR); \ + __SEH_LINEUNIQ_ID(AFTERFINALLY): ; \ + void __SEH_LINEUNIQ_ID(Finally_Executer) () + +#define __except(...) \ + } \ + __seh_macro_exit_spot: \ + if (0) goto __seh_macro_exit_spot; \ + } __EXCEPT( __wine_seh_macro_exception_handler ) { \ + (*fin)(); \ + } __ENDTRY \ + WINE_EXCEPTION_FILTER(__wine_seh_macro_exception_handler) { \ + return ( __VA_ARGS__ ); \ + } \ + } \ + goto __SEH_LINEUNIQ_ID(DETOUR); \ + __SEH_LINEUNIQ_ID(RETOUR): \ + dotry( __thread_outermost_handler_ptr ); \ + } while (0); \ + goto __SEH_LINEUNIQ_ID(AFTEREXCEPT); \ + auto void __SEH_LINEUNIQ_ID(Except_Executer) (); \ + __SEH_LINEUNIQ_ID(DETOUR): \ + __thread_outermost_handler_ptr = & __SEH_LINEUNIQ_ID(Except_Executer); \ + goto __SEH_LINEUNIQ_ID(RETOUR); \ + __SEH_LINEUNIQ_ID(AFTEREXCEPT): ; \ + void __SEH_LINEUNIQ_ID(Except_Executer) () + +#define __leave goto __seh_macro_exit_spot + +#endif /* _WINE_ALREADY_DEFINED_SEH_MACROS__ */ + +#endif /* NO_VC_SEH_MACROS */ + #endif /* USE_COMPILER_EXCEPTIONS */ +#endif /* __WRC__ */ + static inline EXCEPTION_FRAME * WINE_UNUSED __wine_push_frame( EXCEPTION_FRAME *frame ) { #if defined(__GNUC__) && defined(__i386__) -- gmt

4 6

wcmd autoexec.bat?
by J. Grant Feb. 14, 2003

Feb. 14, 2003

Hello I have been testing wcmd to automate some process I used to use on win98. However I can not find the autoexec.bat for wcmd? Currently the wcmd shell starts up with no prompt, this has to be corrected in my batch files. prompt $p$g etc. Wine seems to have 2 autoexec.bat's but does not seem to ever use them. /var/lib/wine/autoexec.bat never seems to be executed. Neither is the one I created in my fake c drive (/home/now3d/.wine/c/autoexec.bat) Is this issue known? are their any workarounds? Regards JG

3 4

Keyboard shortcut to disable DXGrab?
by Matthew Toseland Feb. 14, 2003

Feb. 14, 2003

What would be involved in implementing a keyboard shortcut (ideally configurable) to temporarily disable DXGrab, when playing a game in a window? If such a thing was done, would it be accepted into the source? -- Matthew Toseland toad(a)amphibian.dyndns.org/amphibian(a)users.sourceforge.net Full time freenet hacker. http://freenetproject.org/ Freenet Distribution Node (temporary) at http://amphibian.dyndns.org:8889/ahNdJC6opVw/ ICTHUS.

1 0

Re: popupmenu itemsize
by Uwe Bonnes Feb. 14, 2003

Feb. 14, 2003

>>>>> "Christian" == Christian Kristukat <semmel(a)hoc.net> writes: Christian> Hi, I suspect that this again is a dirty patch but it helps Christian> in working with Microcal Origin. Maybe a more skilled person Christian> than me finds the real origin of this problem. At one point Christian> in the program a popup menu opens but only the items' width Christian> is calculated correctly, the height is not, in fact it is Christian> equal zero. For that case I added an arbitrarily chosen Christian> height of 12 points and it works fine. Here's the patch: ... Christian> lpitem->rect.bottom += 12; + Nagic constants mostly are evil? It the "12" perhaps related to some SYSPARAMS? ... Christian> PS: By the way: every cvs operation (update, diff) hangs Christian> after processing the last directory. However it seems to be Christian> complete and pressing CTRL-C seems not to corrupt Christian> anything. Any ideas? Are you using rhlx01.fht-esslingen.de? Andi works on that problem then. Bye -- Uwe Bonnes bon(a)elektron.ikp.physik.tu-darmstadt.de Institut fuer Kernphysik Schlossgartenstrasse 9 64289 Darmstadt --------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

3 2

suspended threads acquiring synchronization objects
by Peter Hunnisett Feb. 14, 2003

Feb. 14, 2003

Howdy, we've discovered that windows seems to not allow suspended threads to acquire synchronization objects, which make really really good sense, while wine does. I've attached a simple test case, the code for the test case and a ReWind licensed patch for the problem. Would people have a look at the patch and let me know if it's the best way to fix it, or if I've missed anything? Thanks, Peter peter(a)transgaming.com

3 2

Error in the server thread id fix
by Bill Medland Feb. 13, 2003

Feb. 13, 2003

Is anyone else looking into a problem that appears to have been introduced by the server change on Feb 1st changing thread and process ids to small integers? As a result ACCPAC for Windows' main desktop process shuts down after about a minute, but leaves various other processes running. I'll have a look at it but I don't know how successful I'll be. -- Bill Medland ACCPAC International, Inc. medbi01(a)accpac.com Corporate: www.accpac.com Hosted Services: www.accpaconline.com

4 7

2GB visible partition size
by J. Grant Feb. 13, 2003

Feb. 13, 2003

Hello, I noticed there is a 2GB visible partition size. This seems to be like FAT16, however I am running on ext3, is there a config option I can set to remove this limit? Regards JG

4 5

Looking for information on implemented API's in wine
by David Miller Feb. 13, 2003

Feb. 13, 2003

I am ready to begin work on the actual database aspect of cross referencing the windows dlls. As part of the database I would like to include the status of a given API in wine so it will be possible to do things like query for all unimplemented APIs in a given dll. The first questions I have are whether or not I can depend on the spec files to give a complete list, and can someone either point me to documentation on the spec files or explain everything I need to know in order to determine the status of APIs defined there?

1 0

CreateNamedPipe questions
by Dan Kegel Feb. 13, 2003

Feb. 13, 2003

Oddly enough, it looks like it's ok to pass NULL as the filename to CreateNamedPipe. Anyone know what programs do that, and why? Here's a patch that makes our CreateNamedPipe check for legal filenames (or NULL). It should probably allow forward slashes... The patch also adds a quickie regression test, but I still have to use late binding for CreateNamedPipe (or is it ok for the test to fail on win9x?). Would this test have caught the recent named pipe breakage? If not, what would I need to add? Comments welcome; I'll send it to wine-patches after addressing any feedback. - Dan -- Dan Kegel http://www.kegel.com http://counter.li.org/cgi-bin/runscript/display-person.cgi?user=78045 Index: dlls/kernel/sync.c =================================================================== RCS file: /home/wine/wine/dlls/kernel/sync.c,v retrieving revision 1.26 diff -u -p -u -r1.26 sync.c --- dlls/kernel/sync.c 25 Oct 2002 21:02:30 -0000 1.26 +++ dlls/kernel/sync.c 13 Feb 2003 07:47:27 -0000 @@ -508,6 +508,8 @@ HANDLE WINAPI CreateNamedPipeW( LPCWSTR { HANDLE ret; DWORD len = name ? strlenW(name) : 0; + static const WCHAR leadin[] = {'\\','\\','.','\\','P','I','P','E'}; + int i; TRACE("(%s, %#08lx, %#08lx, %ld, %ld, %ld, %ld, %p)\n", debugstr_w(name), dwOpenMode, dwPipeMode, nMaxInstances, @@ -516,7 +518,13 @@ HANDLE WINAPI CreateNamedPipeW( LPCWSTR if (len >= MAX_PATH) { SetLastError( ERROR_FILENAME_EXCED_RANGE ); - return 0; + return INVALID_HANDLE_VALUE; + } + /* FIXME: can name really be NULL?! */ + if (name && strncmpiW(name, leadin, sizeof(leadin)/sizeof(leadin[0]))) + { + SetLastError( ERROR_INVALID_NAME ); + return INVALID_HANDLE_VALUE; } SERVER_START_REQ( create_named_pipe ) { Index: dlls/kernel/tests/Makefile.in =================================================================== RCS file: /home/wine/wine/dlls/kernel/tests/Makefile.in,v retrieving revision 1.6 diff -u -p -u -r1.6 Makefile.in --- dlls/kernel/tests/Makefile.in 4 Oct 2002 17:42:27 -0000 1.6 +++ dlls/kernel/tests/Makefile.in 13 Feb 2003 07:47:27 -0000 @@ -17,6 +17,7 @@ CTESTS = \ generated.c \ locale.c \ path.c \ + pipe.c \ process.c \ thread.c --- /dev/null 2002-08-30 16:31:37.000000000 -0700 +++ dlls/kernel/tests/pipe.c 2003-02-12 23:44:20.000000000 -0800 @@ -0,0 +1,109 @@ +/* + * Unit tests for named pipe functions in Wine + * + * Copyright (c) 2002 Dan Kegel + * + * 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 <stdlib.h> +#include <stdio.h> +#include <time.h> + +#ifndef STANDALONE +#include "wine/test.h" +#else +#include <assert.h> +#define START_TEST(name) main(int argc, char **argv) +#define ok(condition, msg) assert(condition) +#endif + +#include <windef.h> +#include <winbase.h> +#include <winerror.h> +#include <wtypes.h> + +#define PIPENAME "\\\\.\\PiPe\\tests_" __FILE__ + +void test_CreateNamedPipeA(void) +{ + HANDLE hnp; + HANDLE hFile; + const char obuf[] = "Bit Bucket"; + char ibuf[32]; + DWORD written; + DWORD gelesen; + + /* Bad parameter checks */ + hnp = CreateNamedPipeA("not a named pipe", + PIPE_ACCESS_DUPLEX, PIPE_TYPE_BYTE|PIPE_WAIT, + /* nMaxInstances */ 1, + /* nOutBufSize */ 1024, + /* nInBufSize */ 1024, + /* nDefaultWait */ NMPWAIT_USE_DEFAULT_WAIT, + /* lpSecurityAttrib */ NULL); + + ok(hnp == INVALID_HANDLE_VALUE && GetLastError() == ERROR_INVALID_NAME, + "CreateNamedPipe should fail if name doesn't start with \\\\.\\pipe"); + +#if 0 /* Evidently it's ok in windows to pass NULL as the pipe name! */ + hnp = CreateNamedPipeA(NULL, + PIPE_ACCESS_DUPLEX, PIPE_TYPE_BYTE|PIPE_WAIT, + 1, 1024, 1024, NMPWAIT_USE_DEFAULT_WAIT, NULL); + ok(hnp == INVALID_HANDLE_VALUE && GetLastError() == ERROR_INVALID_NAME, + "CreateNamedPipe should fail if name is NULL"); +#endif + + /* Functional checks */ + + hnp = CreateNamedPipeA(PIPENAME, + PIPE_ACCESS_DUPLEX, PIPE_TYPE_BYTE|PIPE_WAIT, + /* nMaxInstances */ 1, + /* nOutBufSize */ 1024, + /* nInBufSize */ 1024, + /* nDefaultWait */ NMPWAIT_USE_DEFAULT_WAIT, + /* lpSecurityAttrib */ NULL); + ok(hnp != INVALID_HANDLE_VALUE, "CreateNamedPipe failed"); + + hFile = CreateFileA(PIPENAME, GENERIC_READ|GENERIC_WRITE, 0, + NULL, OPEN_EXISTING, 0, 0); + ok(hFile != INVALID_HANDLE_VALUE, "CreateFile failed"); + if (hFile == INVALID_HANDLE_VALUE) + return; + + /* Make sure we can read and write a few bytes in both directions*/ + memset(ibuf, 0, sizeof(ibuf)); + ok(WriteFile(hnp, obuf, sizeof(obuf), &written, NULL), "WriteFile"); + ok(written == sizeof(obuf), "write file len"); + ok(ReadFile(hFile, ibuf, sizeof(obuf), &gelesen, NULL), "ReadFile"); + ok(gelesen == sizeof(obuf), "read file len"); + ok(memcmp(obuf, ibuf, written) == 0, "content check"); + + memset(ibuf, 0, sizeof(ibuf)); + ok(WriteFile(hFile, obuf, sizeof(obuf), &written, NULL), "WriteFile"); + ok(written == sizeof(obuf), "write file len"); + ok(ReadFile(hnp, ibuf, sizeof(obuf), &gelesen, NULL), "ReadFile"); + ok(gelesen == sizeof(obuf), "read file len"); + ok(memcmp(obuf, ibuf, written) == 0, "content check"); + + CloseHandle(hFile); + CloseHandle(hnp); +} + +START_TEST(pipe) +{ + test_CreateNamedPipeA(); +} +

3 4

Fix for Treeview Regression in Putty, winamp, et al.
by Drew "DanteAliegri" Ogle Feb. 13, 2003

Feb. 13, 2003

Putty and winamp use dialogs which contain treeviews. When the treeview code sets values to be passed to the callback function, it decides there if it is returning to A or W. When it is returning to a dialog, it returns a W, however, when the user code which the dialog calls is an A function, it excepts TVN_*A, when the treeview code is sending TVN_*W. The most logical place to test for this, was in DefDlgProcW, which then modifies the code when it finds a) the msg to be WM_NOTIFY b) the lParam to be non-null c) the code to be of W form d) the code to be of a TVN_ type. As this is my second real patch, any comments are welcome :) -Dante Index: windows/defdlg.c =================================================================== RCS file: /home/wine/wine/windows/defdlg.c,v retrieving revision 1.29 diff -u -r1.29 defdlg.c --- windows/defdlg.c 14 Jan 2003 19:29:15 -0000 1.29 +++ windows/defdlg.c 13 Feb 2003 05:39:00 -0000 @@ -395,12 +395,22 @@ LRESULT WINAPI DefDlgProcW( HWND hwnd, UINT msg, WPARAM wParam, LPARAM lParam ) { BOOL result = FALSE; + BOOL forWide = FALSE; WNDPROC dlgproc; + /* make sure we only change TVN_ codes. */ + if( msg == WM_NOTIFY && lParam && ((LPNMHDR)lParam)->code & 0x30 ) + forWide = TRUE; SetWindowLongW( hwnd, DWL_MSGRESULT, 0 ); if ((dlgproc = DEFDLG_GetDlgProc( hwnd ))) { + /* if call was meant for a W funcition, but the callback + * is really an A function, change the code from a W code + * to a A code. + **/ + if( forWide && (WINPROC_GetProcType( dlgproc) == WIN_PROC_32A )) + ((LPNMHDR)lParam)->code += 0x31; /* Call dialog procedure */ result = CallWindowProcW( dlgproc, hwnd, msg, wParam, lParam ); /* 16 bit dlg procs only return BOOL16 */

2 2