- Wine-Devel - list.winehq.org

Re: Re-implement MSVCRT *printf
by Mike McCormack Oct. 28, 2004

Oct. 28, 2004

Aneurin Price wrote: > Comments, anyone? Looks like you've put quite a bit of effort into that. I think it's a worthwhile effort, however, please consider: * using a style more consistent with the rest of the Wine codebase * not abusing the preprocessor in printf.h, * not writing C code in the header in printf.h * preventing code duplication by implementing the A functions using the W functions * writing some regression tests to show your code is correct thanks, Mike

4 3

Windows XP Setup
by Jacobus Erasmus Oct. 28, 2004

Oct. 28, 2004

Just a quick question, How much effort would it take to get setup.exe on windows XP CD to run from begining to end. (And in the process install the DLL's fonts etc.) The second question is if somebody is willing to take on this task and how much would it cost me. Regards jacobus(a)kewlconcepts.co.za

1 0

Re: shell32: SHFileOperation improvements
by Mike McCormack Oct. 28, 2004

Oct. 28, 2004

Rolf, Thanks for cleaning up SHFileOperation. Can you please try restructuring your code further to get rid of the excessive number of branching points in that function? Last time I counted, SHFileOperation had more than 60 if/while/for/case statements in the one +500 line function... that makes it hard to read. If you were to make use of helper functions more often, you'd improve your code's readability, eliminate the need for 2-space indent, the pointless /* end-if */ comments, reduce the excessive number of local variables and reduce side effects, thus making it easier for other people to understand and fix the code. Try make small functions that fit in ~50 lines. If a function gets over 50 lines, cut it up into two new functions. Mike Rolf Kalbermatter wrote: > + if (b_Multi || IsAttribFile(FromAttr)) > + continue; > + if (IsAttribDir(FromAttr)) > + { > + /* in w98/w2k/wXp endless recursive move until it aborts due to filename getting to > + long, we don't do this stupidity but return the according error anyhow here */ > + retCode = ERROR_FILENAME_EXCED_RANGE; > + break; > + } > } > - } > + if (( b_SameRoot_MOVE && (b_Multi) && !(b_MultiFrom) && IsAttribFile(FromAttr) && IsAttribDir(ToAttr)) > + || (!(b_MultiFrom) && IsAttribFile(FromAttr) && IsAttribFile(ToAttr)) > + || ( (b_Multi) && IsAttribDir(FromAttr) && IsAttribDir(ToAttr)) > + || (!b_SameRoot_MOVE && IsAttribDir(FromAttr) && IsAttribFile(ToAttr)) > + ) > + { > + ToAttr = SHRenameOnCollision(&nFileOp, pToFile + 1, pFromFile + 1); > + } /* endif */ > + } /* endif */ > + if (!(b_MultiFrom) && (!FlagRenameOnCollision(&nFileOp) || b_SameRoot_MOVE) && > + ((SFSC_IDENTIICAL | SFSC_SOURCE_CONTAINS_TARGET) & f_SameDrive)) > + { > + if (!FlagRenameOnCollision(&nFileOp) && (IsAttribFile(FromAttr))) > + { > + /* target is the same as source */ > + retCode = 0x71; > + } > + break; > + } /* endif */ > + > + /* Analyzing for moving SourceName to TargetName */ > + if ((b_MultiFrom || !b_Multi) && ((IsAttribFile(FromAttr) && IsAttribDir(ToAttr)) || > + (!b_MultiTo && IsAttribDir(ToAttr)) || > + (!b_MultiTo && IsAttribDir(FromAttr) && b_MultiFrom && !b_SameRoot_MOVE))) > + { > + SHFileStrCpyCatW(pTempTo, NULL, pFromFile); > + /* without FOF_MULTIDESTFILES shlfileop would create dir's recursively */ > + nFileOp.fFlags |= FOF_MULTIDESTFILES; > + retCode = SHFileOperationW(&nFileOp); > + continue; > + } > + /* What can we do with one pair and FO_MOVE/FO_COPY ? */ > + /* w98se, nt40 does create recursive dirs, W2K not! wMe ist not tested */ > + if (IsAttribDir(ToPathAttr) || !b_SameRoot_MOVE) > + { > + if (!b_MultiFrom) > + { > + if (IsAttribFile(FromAttr)) > + { > + if (b_SameRoot_MOVE && > + /* windows-bug, MOVE for File also with pToTailSlash, COPY not for this */ > + (!(IsAttribFile(ToAttr)) || (!(FlagAskOverwrite(&nFileOp)) && FlagNotOverwrite(&nFileOp))) && > + (IsAttribDir(ToPathAttr) || b_ToMask)) > + { > + /* With the same drive, we can move for FO_MOVE, dir to dir is solved later */ > + retCode = SHNotifyMoveFileW(nFileOp.pFrom, nFileOp.pTo); > + continue; > + } /* endif */ > + if (IsAttribDir(ToPathAttr) && !pToTailSlash && !IsAttribDir(ToAttr)) > + { > + if (!(FOI_NeverOverwrite & nFileOp.fAnyOperationsAborted)) > + { > + if (!FlagNotOverwrite(&nFileOp)) > + { > + ToAttr = INVALID_FILE_ATTRIBUTES; > + } > + if (IsAttribFile(ToAttr) && FlagNotOverwrite(&nFileOp)) > + { > + if (FlagAskOverwrite(&nFileOp)) > + { > + /* FIXME: we must change the dialog in the future to allow 3-4 cases, > + * 'Yes','No','Yes for all','Never ?' */ > + if ((INVALID_FILE_ATTRIBUTES != ToAttr) && !SHELL_ConfirmDialogW(ASK_OVERWRITE_FILE, pTempTo)) > + { > + retCode = 0x10008; > + nFileOp.fAnyOperationsAborted |= TRUE; > + break; > + } > + ToAttr = INVALID_FILE_ATTRIBUTES; > + } /* endif */ > + } /* endif */ > + } /* endif */ > + if (INVALID_FILE_ATTRIBUTES == ToAttr) > + { > + if (SHNotifyCopyFileW(pTempFrom, pTempTo, FlagRenameOnCollision(&nFileOp))) > + { > + /* the retcode for this case is unknown */ > + retCode = 0x10009; > + } > + if ((FO_COPY != FuncSwitch) && !retCode) > + { > + nFileOp.wFunc = ((level + 1) << FO_LevelShift) + FO_DELETE; > + retCode = SHFileOperationW(&nFileOp); > + } > + continue; > + } > + } > + } > + if (IsAttribDir(FromAttr)) > + { > + if (INVALID_FILE_ATTRIBUTES == ToAttr) > + { > + SetIfPointer(pToTailSlash, '\0'); > + ToAttr = SHRenameOnCollision(&nFileOp, pToFile + 1, pFromFile + 1); > + if (INVALID_FILE_ATTRIBUTES != ToAttr) > + { > + /* w2k returns ERROR_ALREADY_EXISTS, if it found the target with mask */ > +#if 1 > + retCode = ERROR_ALREADY_EXISTS; > +#else > + retCode = ERROR_INVALID_NAME; > +#endif > + nFileOp.fAnyOperationsAborted |= TRUE; > + break; > + } > + lpFileName = &pToFile[lstrlenW(pToFile)]; > + if (pToTailSlash) > + { > + pToTailSlash = lpFileName; > + lpFileName++; > + } > + ((DWORD*)lpFileName)[0] = '\0'; > + retCode = SHNotifyCreateDirectoryW(pTempTo, NULL); > + > + if (retCode) > + { > + retCode = (ERROR_PATH_NOT_FOUND | 0x10000); /* nt40,w98se,w2k,wxp */ > + break; > + } > + ToAttr = GetFileAttributesW(pTempTo); > + SetIfPointer(pToTailSlash,'\\'); > + nFileOp.fFlags &= ~FOF_RENAMEONCOLLISION; > + } > + if (IsAttribDir(ToAttr)) > + { > + /* both are existing dirs, we (FO_)MOVE/COPY the content */ > + pToFile = SHFileStrCpyCatW(pTempFrom, NULL, wWildcardFile); > + nFileOp.fFlags &= ~FOF_MULTIDESTFILES; > + retCode = SHFileOperationW(&nFileOp); > + if ((FO_COPY != FuncSwitch) && !retCode) > + { > + ((DWORD*)pToFile)[0] = '\0'; > + nFileOp.wFunc = ((level + 1) << FO_LevelShift) + FO_DELETE; > + retCode = SHFileOperationW(&nFileOp); > + } > + continue; > + } > + } /* endif IsAttribDir(from) */ > + } /* endif !b_MultiFrom */ > + } > + if (!b_SameRoot_MOVE) > + { > + nFileOp.fAnyOperationsAborted |= TRUE; > + } /* endif */ > + retCode = ERROR_CANCELLED; > + } /* end-while */

1 0

Re: rsaenh
by Juan Lang Oct. 27, 2004

Oct. 27, 2004

Hi Michael, that was quick! Okay, more feedback: 1. English note: destruct_handle_table really should be destroy_handle_table. I know, "destructor" came into common use, but the verb is to destroy. Same with destruct_key, destruct_hash, etc. 2. A CSP isn't allocating a "real" handle, it's sort of a pseudohandle, so the multiple-of-four business isn't necessary. When I was testing SSPs, I found some of them return 0 as a valid handle value. So I think using the index directly would be fine. 3. Do you really need to separate all the functions into an impl version and a non-impl version? It looks as though the non-impl version does the locking, while the impl assumes the table is locked. It seems you could at least document this, and it also doesn't seem necessary in every case, if at all--EnterCriticalSection won't deadlock if a thread that's already entered it calls it again. 4. Some of the OpenSSL checks don't seem valid. In init_hash_impl, for example, you fail if libcrypto isn't available, even if it isn't necessary for the hash. Please fail only if you really need an OpenSSL func. E.g., if (!pMD2_Init) { SetLastError()... } seems better to me. 5. This may be open to debate, but as far as a "more elegant way": perhaps you should protect each typedef with the header file that defines it, e.g. in HASH_CONTEXT, rather than: #ifdef HAVE_OPENSSL MD2_CTX md2; #endif you could use: #ifdef HAVE_OPENSSL_MD2_H MD2_CTX md2; #endif This way you could support the algorithms you were compiled with, even if they aren't the entire set. Sorry I didn't mention these last time.. --Juan __________________________________ Do you Yahoo!? Yahoo! Mail Address AutoComplete - You start. We finish. http://promotions.yahoo.com/new_mail

2 2

Re: Re-implement MSVCRT *printf
by Mike Hearn Oct. 27, 2004

Oct. 27, 2004

I'm not sure if this is just Mozilla acting up but the patch doesn't seem to be fully indented. Is this just UA wierdness? Aneurin Price wrote: > Comments, anyone? >

1 0

Re: Gradient Color for Caption Bar
by Dmitry Timoshkov Oct. 27, 2004

Oct. 27, 2004

"William Poetra Yoga H" <williampoetra(a)yahoo.com> wrote: > This patch is for filling the caption bar (a.k.a. title bar) with gradient > colors, like in Win98/WinME/Win2K/WinXP. > > I've modified nonclient.c too heavily (work on other parts of it), so this > patch may generate offsets or things like that. Please comment on wine-devel. > Thanks. > -IMPORTS = gdi32 advapi32 kernel32 ntdll > +IMPORTS = msimg32 gdi32 advapi32 kernel32 ntdll Use gdi32.GdiGradientFill instead of msimg32.GradientFill, the latter one is forwarded to gdi32 anyway. -- Dmitry.

3 3

Re: GetTimeZoneInformation return value fix
by Alexandre Julliard Oct. 27, 2004

Oct. 27, 2004

Rein Klazes <rklazes(a)xs4all.nl> writes: > - reordered some code from the existing helper _GetTimeZoneBias > to make it more general useful (and changed the helper names to > TIME_xyz). > - Make GetTimeZoneInformation return the correct zone id > constants. > - changed some "const LPX_Y_Z" declarations to "const X_Y_Z *" This breaks the msvcrt tests: ../../../tools/runtest -q -P wine -M msvcrt.dll -T ../../.. -p msvcrt_test.exe.so time.c && touch time.ok time.c:101: Test failed: Wrong date:Year 70 mon 0 yday 1 mday 2 wday 5 hour 0 min 0 sec 0 dst -1 time.c:112: Test failed: Wrong date:Year 70 mon 0 yday 1 mday 2 wday 5 hour 0 min 0 sec 0 dst -1 -- Alexandre Julliard julliard(a)winehq.org

2 1

RE: Wine User Guide Improvements
by Scott Ritchie Oct. 27, 2004

Oct. 27, 2004

Brian, I agree with your plan to condense the guide down. I would still keep some parts of the Getting Wine chapter and put it at the end of the introduction, or possibly at the beginning of configuring wine. I agree about changing it to be mostly a reference to the relevant parts of WineHQ. Would it be duplicitous of me to just go ahead and redo the Getting Wine chapter and leave it as it is now, and perhaps merge it into configuring/introduction once you're done? That way we'll have 4 chapters: Introduction Configuring Wine Running Wine Troubleshooting The running wine chapter eventually needs to be rewritten too (outdated switches, etc.) It could probably be reorganized a bit to read nicer as well (for example, directions for uninstalling programs are given before running them.) I'd like to have the documentation revamped by the next release. Combined with the coming winecfg patches, the major red marks at http://www.winehq.org/site/todo_lists can be replaced with happy green bars. I think we can easily have the yellow bit for the User Guide finished by then too. After that, the only major things in the way of 0.9 seem to be the developer and winelib guide, which are a bit beyond my abilities. (note: forwarded to wine-devel for further feedback) Thanks, Scott Ritchie

2 1

rsaenh
by Michael Jung Oct. 26, 2004

Oct. 26, 2004

Hello, this is an updated version of the rsaenh implementation from yesterday. Updates: 1.) Broke long lines to not exceed a length of 100 characters. 2.) Substituted tabs with spaces. 3.) Removed magic constants. 4.) Rsaenh now uses MD4, MD5 and SHA from advapi32. 5.) Regression tests should work now without failures even if OpenSSL is not available. 6.) Fixed filenames in comments. If you want to try the patch note the following: - After you have applied the patch you have to run autoheader and autoconf. Then cd dlls ; make_dlls ; cd .. ; configure. - After you have compiled and installed wine, you have to unregister rsabase and register rsaenh (regsvr32 /U rsabase ; regsvr32 rsaenh). rsaenh.dll is meant to replace rsabase.dll sometime (Windows XP also has rsaenh.dll only). Greetings, Michael

1 0

Re: WSCEnumProtocols
by Alexandre Julliard Oct. 26, 2004

Oct. 26, 2004

Hans Leidekker <hans(a)it.vu.nl> writes: > +INT WINAPI WSAEnumProtocolsA( LPINT lpiProtocols, LPWSAPROTOCOL_INFOA lpBuffer, LPDWORD lpdwLength ) > +{ > + return WINSOCK_EnumProtocols( lpiProtocols, (WSAPROTOCOL_INFOW *)lpBuffer, lpdwLength, FALSE ); > +} You cannot cast a WSAPROTOCOL_INFOA* to a WSAPROTOCOL_INFOW*, the structures don't have the same size. -- Alexandre Julliard julliard(a)winehq.org

2 2