October 2008 - Wine-devel - list.winehq.org

wine users forum registration issue
by Nicolas Le Cam 23 Oct '08

23 Oct '08

2008/10/23 Zachary Goldberg <zgold(a)bluesata.com>: > Captchas in general are not so good, and there are much better > anti-spam techniques out there (Such as asking an actual question. > There are phpbb mods which allow you to input a list of questions > (with perhaps links to wiki pages with answers) from which it'll chose > a random one to present a user at registration. This is very simple > for the user to get passed, just check the wiki for the answer to > hopefully an obvious and perhaps informative question, and nearly > impossible for the bots). > > So perhaps ditching the captcha in favor of something less error prone > is in order? 2008/10/23 Rosanne DiMesio <dimesio(a)earthlink.net>: > It took me about five tries to register for the forum several months > ago. The reason is simple: aging eyesight. I have problems with captchas > in general, but I have to say that some of the ones used for the forum > are especially difficult to see because the background of the image also > contains letters and number, and it is not always obvious which letters > and numbers are supposed to be "right" ones. Some captchas are really difficult to read (especially those with colors for people like me that are color blindless) but some other aren't. For exemple recaptchas [1] are easily readable. And it also helps archive.org digitalizing books. Perhaps switching to such a system will avoid crazy captchas and help people register onto the forum. [1] http://recaptcha.net/learnmore.html (Resending because I have again forgotten wine-devel ...)

1 0

Re: [Try 5] msvcrt: Set permissions of file in sopen
by Dmitry Timoshkov 23 Oct '08

23 Oct '08

"Ivan Peevski" <cyberhorse(a)gmail.com> wrote: > the pmode flag test should be: > if it has the IREAD but not the IWRITE flag (read only) > does that make sense? Is it doing what I think it is? It should be: if ((pmode & MSVCRT__S_IREAD) && !(pmode & MSVCRT__S_IWRITE)) ...; IMO it would be more convenient to use switch, that would cover all the possible cases: switch (pmode & (MSVCRT__S_IREAD | MSVCRT__S_IWRITE)) { case 0: ... case MSVCRT__S_IREAD | MSVCRT__S_IWRITE: ... case MSVCRT__S_IREAD: ... case MSVCRT__S_IWRITE: ... } Of course a test is needed to prove that. -- Dmitry.

1 0

Re: [1/2] richedit: Added test for correct behaviour for callingEditWordBreakProc.
by Dmitry Timoshkov 23 Oct '08

23 Oct '08

"Dylan Smith" <dylan.ah.smith(a)gmail.com> wrote: > EditWordBreakProc documented the third parameter as being the number of > unicode characters in the string for richedit 2.0 and up. It turns out > that it should actually be the number of bytes in the string. Have you tried to set a unicode word break proc? The behaviour you see could be explained by the use of SendMessageA in your test, or even most likely an ANSI richedit window (RICHEDIT_CLASS vs. RICHEDIT_CLASSW). -- Dmitry.

1 0

Re: [Try 5] msvcrt: Set permissions of file in sopen
by Dmitry Timoshkov 23 Oct '08

23 Oct '08

"Ivan Peevski" <cyberhorse(a)gmail.com> wrote: > + if (pmode & (MSVCRT__S_IREAD & !MSVCRT__S_IWRITE)) > + attrib |= FILE_ATTRIBUTE_READONLY; The pmode flags test doesn't look right. Also the 'attrib |= ' line has a trailing space. -- Dmitry.

1 0

Re: [Try 5] msvcrt: Set permissions of file in sopen
by Austin English 23 Oct '08

23 Oct '08

On Wed, Oct 22, 2008 at 10:13 PM, Ivan Peevski <cyberhorse(a)gmail.com> wrote: > Followed the documentation at http://msdn.microsoft.com/en-us/library/w7sa2b22(VS.80).aspx. Ignore pmode if the file already exist and emulate permissions as best we can during file creation. > > --- > dlls/msvcrt/file.c | 13 ++++++++----- > 1 files changed, 8 insertions(+), 5 deletions(-) > > diff --git a/dlls/msvcrt/file.c b/dlls/msvcrt/file.c > index 7c0f74d..7b2a33e 100644 > --- a/dlls/msvcrt/file.c > +++ b/dlls/msvcrt/file.c > @@ -1422,6 +1422,7 @@ int CDECL MSVCRT__sopen( const char *path, int oflags, int shflags, ... ) > int pmode; > DWORD access = 0, creation = 0, attrib; > DWORD sharing; > + struct MSVCRT__stat buf; > int wxflag = 0, fd; > HANDLE hand; > SECURITY_ATTRIBUTES sa; > @@ -1430,6 +1431,8 @@ int CDECL MSVCRT__sopen( const char *path, int oflags, int shflags, ... ) > TRACE(":file (%s) oflags: 0x%04x shflags: 0x%04x\n", > path, oflags, shflags); > > + attrib = FILE_ATTRIBUTE_NORMAL; > + > wxflag = split_oflags(oflags); > switch (oflags & (MSVCRT__O_RDONLY | MSVCRT__O_WRONLY | MSVCRT__O_RDWR)) > { > @@ -1444,10 +1447,11 @@ int CDECL MSVCRT__sopen( const char *path, int oflags, int shflags, ... ) > pmode = va_arg(ap, int); > va_end(ap); > > - if(pmode & ~(MSVCRT__S_IREAD | MSVCRT__S_IWRITE)) > - FIXME(": pmode 0x%04x ignored\n", pmode); > - else > - WARN(": pmode 0x%04x ignored\n", pmode); > + /* Ignore _S_IREAD and _S_IWRITE if file exists. */ > + if (MSVCRT__stat(path, &buf) != 0) > + if (pmode & (MSVCRT__S_IREAD & !MSVCRT__S_IWRITE)) > + attrib |= FILE_ATTRIBUTE_READONLY; > + /* Ignore any irrelevant flags. */ > > if (oflags & MSVCRT__O_EXCL) > creation = CREATE_NEW; > @@ -1482,7 +1486,6 @@ int CDECL MSVCRT__sopen( const char *path, int oflags, int shflags, ... ) > ERR( "Unhandled shflags 0x%x\n", shflags ); > return -1; > } > - attrib = FILE_ATTRIBUTE_NORMAL; > > if (oflags & MSVCRT__O_TEMPORARY) > { > -- > 1.5.4.3 > > > > Can you add a testcase for this to dlls/msvcrt/tests/file.c? -- -Austin

1 0

Re: [include] - add dde.rh
by Dmitry Timoshkov 23 Oct '08

23 Oct '08

"Vijay Kiran Kamuju" <infyquest(a)gmail.com> wrote: > Changelog > -------------- > Add dde.rh Why are you plain copying it from PSDK? -- Dmitry.

1 0

make test solution
by chris ahrendt 22 Oct '08

22 Oct '08

NM figured it out... some odd reason the git directory has to be mapped to a windows drive.. once I did that.. it all cleaned up... chris

1 0

make test issue..
by chris ahrendt 22 Oct '08

22 Oct '08

I have a general question on the make files... specifically the make test portion of it.. Since talking to dan about things this would be where I could be of most help. I have an issue with the make test currently.. I am getting : [cahrendt(a)stinky wine-git]$ make test make[1]: Entering directory `/home/cahrendt/wine-git/tools' make[1]: `makedep' is up to date. make[1]: Leaving directory `/home/cahrendt/wine-git/tools' make[1]: Entering directory `/home/cahrendt/wine-git/dlls' make[2]: Entering directory `/home/cahrendt/wine-git/dlls/advapi32/tests' ../../../tools/runtest -q -P wine -M advapi32.dll -T ../../.. -p advapi32_test.exe.so cred.c && touch cred.ok Warning: could not find DOS drive for current working directory '/home/cahrendt/wine-git/dlls/advapi32/tests', starting in the Windows directory. wine: could not load L"C:\\windows\\system32\\advapi32_test.exe.so": Module not found make[2]: *** [cred.ok] Error 126 make[2]: Leaving directory `/home/cahrendt/wine-git/dlls/advapi32/tests' make[1]: *** [advapi32/tests/__test__] Error 2 make[1]: Leaving directory `/home/cahrendt/wine-git/dlls' make: *** [dlls/__test__] Error 2 which is not a code issue as far as I am able to see but rather a make file issue. As you see the advapi32 test can't be found. However if you look in the directory the .so is there... so I need some assistance to help me understand where and how the make test works so I can figure out why its looking in the system32 directory and not the current directory for the shared object. Any help? Chris

1 0

Re: [PATCH 1/2] jscript: Added Math.floor implementation.
by Vincent Povirk 22 Oct '08

22 Oct '08

+tmp = Math.floor(1.1); +ok(tmp === 1, "Math.round(1.1) = " + tmp); Copy/paste error? Vincent Povirk 2008/10/21 Jacek Caban <jacek(a)codeweavers.com>: > --- > dlls/jscript/math.c | 20 ++++++++++++++++++-- > dlls/jscript/tests/api.js | 18 ++++++++++++++++++ > 2 files changed, 36 insertions(+), 2 deletions(-) > > > >

2 1

re: Dylan Smith : richedit:Added tests for ctrl-key shortcut handling in WM_KEYDOWN.
by Dan Kegel 22 Oct '08

22 Oct '08

Dmitry wrote: >> + BYTE key_state[256]; >> +... >> + result = GetKeyboardState((LPBYTE)&key_state); > the casts in the above code clearly > show that the pointer syntax is wrong there. Nah. It's ugly, but not wrong. key_state == &key_state here. - Dan

1 0