- Wine-Devel - list.winehq.org

wine-user.pdf feedback 2
by Dan W. March 4, 2005

March 4, 2005

Okay, now I found what I think is a more serious missing piece of info in wine-user.pdf. In page 20, last paragraph, it says one can start off by modifying the sample config file, which "can be found in the directory 'documentation/samples'... Problem is, I did a file search for 'documentation', and I get like 10 pages of hits I did a file search for samples, and same thing, or worse. In both cases I went through all those instances visually scanning for 'wine' or 'config' and found nothing at all. I know I ought to have the sources somewhere, because when I installed fc2, I said "install everything", and that included sources; but I have no idea where to look for the wine documentation folder. I have to give up on it for now. Far past my bedtime by now, anyways.. ;-) Cheers!

3 2

Re: MSI: Radio Buttons
by Dmitry Timoshkov March 4, 2005

March 4, 2005

"Ulrich Czekalla" <ulrich.czekalla(a)utoronto.ca> wrote: > + if (control->hwnd) > + { > + WNDPROC oldproc = (WNDPROC) SetWindowLongPtrW(control->hwnd, GWLP_WNDPROC, > + (LONG)MSIRadioGroup_WndProc); Then it's better to cast to LONG_PTR instead of LONG. -- Dmitry.

1 0

Re: [ADVAPI] SetSecurityInfo
by Robert Shearman March 3, 2005

March 3, 2005

Raphael wrote: >Hi, > > Changelog: > - simple implementation of SetSecurityInfo > >(seems to be needed by War3 frozen throne, cf >http://forums.gentoo.org/viewtopic-t-303306.html?sid=3f962c702ae58ad8902ca8…) > > I have made a few comments on the implementation below: > SRCDIR = @srcdir@ > VPATH = @srcdir@ > MODULE = advapi32.dll >-IMPORTS = kernel32 ntdll >+IMPORTS = winspool kernel32 ntdll > > Should be in DELAYIMPORTS. We don't want to slow down the loading of all applications because of this one uncommon case. >@@ -1418,6 +1422,87 @@ > { > FIXME("(%s) : stub\n", debugstr_w(lpFileName) ); > return TRUE; >+} >+ >+/****************************************************************************** >+ * SetSecurityInfo [ADVAPI32.@] >+ */ >+BOOL WINAPI SetSecurityInfo(HANDLE handle, >+ SE_OBJECT_TYPE ObjectType, >+ SECURITY_INFORMATION SecurityInfo, >+ PSID psidOwner, >+ PSID psidGroup, >+ PACL pDacl, >+ PACL pSacl) >+{ >+ SECURITY_DESCRIPTOR sd; >+ BOOL test; >+ >+ TRACE("(%p, %x, %x, %p, %p, %p, %p) : alomst stub\n", handle, ObjectType, SecurityInfo, psidOwner, psidGroup, pDacl, pSacl); > > Remove the "almost stub" part of it and just print FIXME's for the parts that aren't implemented, otherwise the next person that comes along to this function will spend hours trying to work out what isn't implemented, if they even realise this function is to blame. >+ >+ test = InitializeSecurityDescriptor(&sd, SECURITY_DESCRIPTOR_REVISION); >+ if (!test) return FALSE; >+ >+ if (SecurityInfo & DACL_SECURITY_INFORMATION) { >+ test = SetSecurityDescriptorDacl(&sd, TRUE, pDacl, TRUE); >+ } >+ if (SecurityInfo & SACL_SECURITY_INFORMATION) { >+ test = SetSecurityDescriptorSacl(&sd, TRUE, pSacl, TRUE); >+ } >+ if (SecurityInfo & GROUP_SECURITY_INFORMATION) { >+ if (NULL == psidGroup) { >+ SetLastError( ERROR_INVALID_PARAMETER ); >+ return FALSE; >+ } >+ test = SetSecurityDescriptorGroup(&sd, psidGroup, TRUE); >+ } >+ if (SecurityInfo & OWNER_SECURITY_INFORMATION) { >+ if (NULL == psidOwner) { >+ SetLastError( ERROR_INVALID_PARAMETER ); >+ return FALSE; >+ } >+ test = SetSecurityDescriptorOwner(&sd, psidOwner, TRUE); >+ } >+ >+ switch (ObjectType) { >+ case SE_KERNEL_OBJECT: >+ test = SetKernelObjectSecurity(handle, SecurityInfo, &sd); >+ break; >+ case SE_SERVICE: >+ test = SetServiceObjectSecurity(handle, SecurityInfo, &sd); >+ break; >+ case SE_REGISTRY_WOW64_32KEY: >+ case SE_REGISTRY_KEY: >+ test = RegSetKeySecurity(handle, SecurityInfo, &sd); >+ break; >+ case SE_LMSHARE: >+ /*test = NetShareSetInfo(handle, SecurityInfo, &sd); TODO: don't exit on netapi */ >+ break; >+ case SE_PRINTER: >+ { >+ PRINTER_INFO_2A pr_info; >+ memset(&pr_info, 0, sizeof(pr_info)); >+ /** maybe we should do GetPrinter before SetSecurityDesc* ? */ >+ test = GetPrinterA(handle, 2, (LPBYTE) &pr_info, 0, NULL); >+ pr_info.pSecurityDescriptor = &sd; >+ test = SetPrinterA(handle, 2, (LPBYTE) &pr_info, PRINTER_CONTROL_SET_STATUS); >+ break; >+ } >+ >+ case SE_FILE_OBJECT: >+ /*SetFileSecurity ? */ >+ case SE_WINDOW_OBJECT: > > Call SetUserObjectSecurity. >+ case SE_DS_OBJECT: >+ case SE_DS_OBJECT_ALL: >+ case SE_PROVIDER_DEFINED_OBJECT: >+ case SE_WMIGUID_OBJECT: >+ case SE_UNKNOWN_OBJECT_TYPE: >+ default: >+ WARN("unsupported ObjectType %d\n", ObjectType); > > Change this to a FIXME. >+ test = TRUE; >+ } >+ >+ return test; > > This function returns a DWORD error code, not a BOOL. > } > > /****************************************************************************** > > Rob

1 0

Re: ntdll: Add NtCreateKey tests
by Alexandre Julliard March 3, 2005

March 3, 2005

James Hawkins <truiken(a)gmail.com> writes: > What are the steps I need to take to break registry initialisation? I > did wonder how anything in the real win32 registry could be created > though if you can't create a key directly under HKLM or HKU. If this > really does break initialisation, can we implement a different > create_key that has little restrictions and would only be used > internally, or would this just be scrapped entirely? I don't think there's any point in replicating that behavior. I doubt there's any such check on Windows either, it's most likely just a side effect of the way the registry files are "mounted" at different places in the tree. Once we have a proper implementation of RegLoadKey/RegUnloadKey we should get the same behavior automatically. -- Alexandre Julliard julliard(a)winehq.org

1 0

any way to unlink the symbolic link, not its target?
by Kuba Ober March 3, 2005

March 3, 2005

Hi, I'm using wine to cross compile projects using vc7.1 from a software configuration tool called aegis (http://aegis.sf.net) I'm maintaining a symlinked baseline with source and derived (.obj, .exe, .pdb) files in it. A change only has copies of changed files. So, when the compiler tries to output something, it first tries to unlink the file. But that won't work as wine tries to unlink target of the symlink, instead of the symlink itself. Is there any switch somewhere to revert to deleting the symbolic links, or if not -- what source file should I look into to patch that? Thanks a lot, Kuba Ober

1 0

Re: ntdll: Add NtCreateKey tests
by Robert Shearman March 3, 2005

March 3, 2005

James Hawkins wrote: >On 03 Mar 2005 11:02:24 +0100, Alexandre Julliard <julliard(a)winehq.org> wrote: > > >>James Hawkins <truiken(a)gmail.com> writes: >> >> >> >>>I've been discerning the behavior of RegCreateKey and NtCreateKey when >>>creating a key directly under HKLM or HKU, and this test reveals that >>>NtCreateKey does check for this case and returns a >>>STATUS_INVALID_PARAMETER. >>> >>> >>This is going to break registry initialisation. Do you really have an >>app that depends on that? >> >> >> > >What are the steps I need to take to break registry initialisation? I >did wonder how anything in the real win32 registry could be created >though if you can't create a key directly under HKLM or HKU. If this >really does break initialisation, can we implement a different >create_key that has little restrictions and would only be used >internally, or would this just be scrapped entirely? > > It looks like Windows only uses RegLoadKey to create keys under HKLM and HKU. Maybe we should do the same. Rob

2 4

Re: Wine cabinet.dll FDI Conformance Test Patch
by Dimitrie O. Paun March 3, 2005

March 3, 2005

On Wed, Mar 02, 2005 at 03:41:21PM -0800, Rizwan Kassim wrote: Good stuff guys! Here are a few comments: +#ifndef STANDALONE +#include <wine/test.h> +#define ok2 ok +#else +/* To build outside Wine tree, compile with cl -DSTANDALONE -D_X86_ tvfs.c cabinet_fdi.c FDI.lib + These are only called if standalone are used, defining ok and START_TEST, which would normally be declared in winetree */ +#include <assert.h> +#include <stdio.h> +#define START_TEST(name) main(int argc, char **argv) +#define ok(condition, msg) \ + do { if(!(condition)) { \ + fprintf(stderr,"failed at %d, msg:" msg "\n",__LINE__); \ + exit(1); \ + } } while(0) +#define ok2(condition, msg, arg) \ + do { if(!(condition)) { \ + fprintf(stderr,"failed at %d, msg:" msg "\n",__LINE__, arg); \ + exit(1); \ + } } while(0) +#define todo_wine Please get rid of this. It is useless in the Wine tree. If you need it, just put it in a private header file, and include it there. +#ifndef DEBUG_ALLOC +FNALLOC(final_alloc) { + return malloc(cb); +} +FNFREE(final_free) { + free(pv); + return; +} +#else +FNALLOC(final_alloc) { + trace(" FNALLOC just called with %d\n",cb); + return malloc(cb); +} +FNFREE(final_free) { + trace(" FNFREE just called with %d\n",pv); + free(pv); + return; +} +#endif No need for this either. It's a test after all, there's no big deal if you have a memory leak... :) +#ifdef VERBOSE + trace(" FNFDINOTIFY real just called with %d, %d \n",fdint,pfdin); +#endif Don't do it this way. Define trace such that you can use it unconditionally. In general, try to avoid using the preprocessor in the code. + trace(" Cabinet Data : cbC %d cF %d cFi %d si %d iC %d fr %d hp %d hn %d\n", + cabinfo.cbCabinet, + cabinfo.cFolders , + cabinfo.cFiles , + cabinfo.setID, + cabinfo.iCabinet, + cabinfo.fReserve , + cabinfo.hasprev , + cabinfo.hasnext ); These should fit on 2-3 lines, no? + ok2 ( cabinfo.cbCabinet == TcbCabinet, "FDIIsCabinet,cabinfo %s data did not match! Failed!\n", cabname); + ok2 ( cabinfo.cFolders == TcFolders, "FDIIsCabinet,cabinfo %s data did not match! Failed!\n", cabname); + ok2 ( cabinfo.cFiles == TcFiles, "FDIIsCabinet,cabinfo %s data did not match! Failed!\n", cabname); + ok2 ( cabinfo.setID == TsetID, "FDIIsCabinet,cabinfo %s data did not match! Failed!\n", cabname); + ok2 ( cabinfo.iCabinet == TiCabinet, "FDIIsCabinet,cabinfo %s data did not match! Failed!\n", cabname); + ok2 ( cabinfo.fReserve == TfReserve, "FDIIsCabinet,cabinfo %s data did not match! Failed!\n", cabname); + ok2 ( cabinfo.hasprev == Thasprev, "FDIIsCabinet,cabinfo %s data did not match! Failed!\n", cabname); + ok2 ( cabinfo.hasnext == Thasnext, "FDIIsCabinet,cabinfo %s data did not match! Failed!\n", cabname); Use ok() instead of ok2(). Also, it's probably better if the message contains the values of the variables that don't match, so we know why it failed. +#ifdef TVFS_DEBUG + trace("tvfs_create called with %s, %d, %d\n", fname, buf, len); +#endif Similar to the one above, do the test inside the body of trace(), not on every use. + switch(whither) { + case SEEK_SET: { + handles[h]->pos = whence; + break; + } + case SEEK_CUR: { + handles[h]->pos += whence; + break; + } This one needs a bit of indentation, no? +int tvfs_close(int h){ + int inode = handles[h]->inode; +#ifdef TVFS_DEBUG + trace("tvfs_close called with %d\n", h); +#endif + if (!files[inode]){ + return -1; + } + free(handles[h]); + /* Currently does NOT enabled open to reuse this handle */ + return 0; + } The formatting on this is funny. It needs indenting, spaces before '{', and the opening braket for the function on a new line. +main(){ Ditto. -- Dimi.

2 1

Re: MSI trouble with ALLUSERPROFILE
by Juan Lang March 3, 2005

March 3, 2005

> I get err with %ALLUSERPROFILE during install program: <snip> > Where I will set this env variable? (I know I can do it in shell, but > what is the right way?) Hm.. I never did get this quite right. For consistency, it should be in wine.inf. That's where USERPROFILE is: HKLM,System\CurrentControlSet\Control\Session Manager\Environment,"USERPROFILE",,"%53%" But I don't think it can go here, because setupapi (which is what interprets those funny "%53" type strings) doesn't have a magic constant for this, and there isn't a CSIDL value for it. So maybe shell32 should set it in SHELL_RegisterShellFolders? --Juan __________________________________ Celebrate Yahoo!'s 10th Birthday! Yahoo! Netrospective: 100 Moments of the Web http://birthday.yahoo.com/netrospective/

1 0

Use of STLPort with winegcc
by Boaz Harrosh March 3, 2005

March 3, 2005

Winegcc will prepend any of the command line switches with "-I wine/include/msvcrt" if the "-no-cygwin" flag is specified. This breaks STLPort because STLPort must be first on the search path, before the regular c-runtime of the compiler. This way it can do it's tricks of what to use and what not to use from the standard c-runtime. What is the current way to solve it. I could remove the "-no-cygwin" flag and Manually add the "-I wine/include/msvcrt" and "-D__MSVCRT__". But how do I add the "-dmsvcrt" on the Winebuild command line, as it is all internal to winegcc. What would be the right way to do it? I would submit a patch, but what should I do? Just append the "-I .../msvcrt" instead of prepend, or maybe use "-isystem .../msvcrt". The docs for gcc say that "-isystem" folders will be searched before the default gcc folders but after any -I folders. Free Life Boaz

2 1

Re: MSI: Radio Buttons
by Dmitry Timoshkov March 3, 2005

March 3, 2005

"Ulrich Czekalla" <ulrich.czekalla(a)utoronto.ca> wrote: > + if (control->hwnd) > + { > + WNDPROC oldproc = (WNDPROC) SetWindowLongW(control->hwnd, GWLP_WNDPROC, > + (LONG)MSIRadioGroup_WndProc); I'd suggest to use SetWindowLongPtrW even it's just a macro in Win32 we still need to write 64-bit clean code. -- Dmitry.

1 0