January 2010 - Wine-Devel - list.winehq.org

Re: Implement OleCreatePropertyFrame
by Piotr Caban Jan. 11, 2010

Jan. 11, 2010

Hi, In prop_sheet_proc you wrote: case PSN_APPLY: FIXME("(%p, %s, %s)\n", hwnd, "WM_NOTIFY", "PSN_APPLY"); IPropertyPage_Apply(opf->propPage); return TRUE; If I understand it correctly applying changes is fully implemented. Why there's a FIXME? Probably you wanted to use TRACE here. default: FIXME("(%p, %d, %ld, %p)\n", hwnd, msg, wparam, (void *)lparam); There's lots of messages that your're not going to handle (e.g. WM_SETCURSOR) so this FIXME will be very loud. In pixels_to_dialog_units function: You're using some hardcoded values here. Probably something like this would be better, but I'm not sure if it's correct: static void pixels_to_dialog_units(short *x_pixels, short *y_pixels) { int basey; float basex; HFONT font = GetStockObject(DEFAULT_GUI_FONT); HDC hdc = GetDC(NULL); TEXTMETRICW tm; SIZE size; font = SelectObject(hdc, font); GetTextMetricsW(hdc, &tm); GetTextExtentPoint32A(hdc, "ABCDEFGHIJKLMNOPQRSTUVWXYZ" "abcdefghijklmnopqrstuvwxyz", 52, &size); SelectObject(hdc, font); basex = ((float)size.cx/26+1)/2; basey = tm.tmHeight; *x_pixels = (float)(*x_pixels)*4/basex; *y_pixels = MulDiv(*y_pixels, 8, basey); } Of course the basex and basey values should be computed once. In OleCreatePropertyFrameIndirect it's probably good to add a FIXME, that dispidInitialProperty value is ignored. Please try to use the same formatting in whole file (e.g. you sometimes use white chars before brackets). Cheers, Piotr

1 0

Re: advpack: Use GetModuleHandle instead of LoadLibrary.
by Paul Vriens Jan. 10, 2010

Jan. 10, 2010

On 01/05/2010 02:46 AM, James Hawkins wrote: > --- > dlls/advpack/tests/advpack.c | 7 +------ > 1 files changed, 1 insertions(+), 6 deletions(-) > Hi James, Don't think this works: advpack: 0 tests executed (0 marked as todo, 0 failures), 0 skipped. -- Cheers, Paul.

2 2

Re: [2/4] avifil32/tests: Initial-test-data-and-simple-test-for-avifil32-depends
by James McKenzie Jan. 10, 2010

Jan. 10, 2010

Julius Schwartzenberg wrote: > These are resents of last week which are cleaned up more and with > Detlef's patch on which they depend in front. > > While looking at avifile, I noticed some strange behaviour. Because > some fixes may seem strange, I felt a test suite was really necessary > to show that this is really what's happening on Windows. All the tests > pass on both Windows 98 and Windows XP. > > C++ type comments are not allowed. Please mark second and subsequent submissions with [Try #] in the message header. It keeps us from guessing which submission is the latest. There is no requirement to submit subsequent parts of a patchset if there are no changes to them if changes in prior sections do not affect the code in them. James McKenzie

2 1

Re: avifil32: Overwrite dwRate in the audiostream with the nBlockAlign from the wave header
by Paul Vriens Jan. 10, 2010

Jan. 10, 2010

On 01/10/2010 03:18 AM, Julius Schwartzenberg wrote: > I was stupid last week and sent patches in a serie that didn't actually > depend on each other. I'm resending them properly now. All fixes are > actually independent from each other. > > I started looking into some problems avifil32 had because Civilization > II would crash if it tried to play it's intro movie. > I created a small program to print information about the different > headers in AVI files and to play back the audio stream. Also with this > program, many AVI files failed to play (while they worked on Windows > with it). The fixes I introduce also allow my testing program to play > back everything correctly. > > These patches solve this problem and several other problems. > Hi Julius, Do any of these patches fix any of the tests? If so you should also change the tests by removing one or more todo_wine()'s. This could become tricky though if your tests patches are not committed before these ones. -- Cheers, Paul.

3 2

Re: ntdll/tests: Fix a test failure on NT4
by Paul Vriens Jan. 10, 2010

Jan. 10, 2010

On 01/10/2010 05:44 PM, André Hentschel wrote: > --- > dlls/ntdll/tests/om.c | 3 ++- > 1 files changed, 2 insertions(+), 1 deletions(-) > > diff --git a/dlls/ntdll/tests/om.c b/dlls/ntdll/tests/om.c > index 1bab21d..e351813 100644 > --- a/dlls/ntdll/tests/om.c > +++ b/dlls/ntdll/tests/om.c > @@ -677,7 +677,8 @@ static void test_query_object(void) > ok( status == STATUS_SUCCESS, "NtQueryObject failed %x\n", status ); > ok( len> sizeof(UNICODE_STRING), "unexpected len %u\n", len ); > str = (UNICODE_STRING *)buffer; > - ok( sizeof(UNICODE_STRING) + str->Length + sizeof(WCHAR) == len, "unexpected len %u\n", len ); > + ok( sizeof(UNICODE_STRING) + str->Length == len /* WinNT */ || > + sizeof(UNICODE_STRING) + str->Length + sizeof(WCHAR) == len, "unexpected len %u\n", len ); > trace( "got %s len %u\n", wine_dbgstr_w(str->Buffer), len ); > pNtClose( handle ); > } Looks a bit like: http://source.winehq.org/patches/data/57319 Except that I used broken() for NT4 which IMHO makes more sense as this is the only platform with this issue. -- Cheers, Paul.

2 2

Re: Resent of [1/4] and [2/4] from avifil32/test serie
by Paul Vriens Jan. 10, 2010

Jan. 10, 2010

On 01/10/2010 05:35 PM, Julius Schwartzenberg wrote: > First of all, my apologies for all the junk. I didn't fully understand > how the mail moderation exactly worked and made some wrong guesses. > > I'm going to use git send-email now as Henri suggested me, to resent two > patches from the serie as replies to this e-mail. Both with included > suggestions from Paul & Henri wine-devel. > > [1/4]: This is Detlef's patch from June 2008 with just the import > changed from "vfw32" to "avifil32" to work around bug 14085. > > http://www.winehq.org/pipermail/wine-patches/2008-June/056310.html > > > [2/4]: These are resents of last week which are cleaned up more and with > Detlef's patch on which they depend in front. > > While looking at avifile, I noticed some strange behaviour. Because some > fixes may seem strange, I felt a test suite was really necessary to show > that this is really what's happening on Windows. All the tests pass on > both Windows 98 and Windows XP. Remember that we have that great winetestbot (https://winetestbot.geldorp.nl/index.pl) which I encourage people to use as it covers NT4 up to W2K8R2 if needed. -- Cheers, Paul.

1 0

Re: Sorry!! Please drop the previous mails ^^^
by Henri Verbeet Jan. 10, 2010

Jan. 10, 2010

2010/1/10 Julius Schwartzenberg <j.schwartzenberg(a)student.utwente.nl>: > My apologies to who'll read this. I found out a bit late my mailclient > sucked and used the wrong 'from' address. I've resent all my patches with > the proper address now, so all other mails from me may be dropped. > I'm not sure what was wrong with it? Afaik the address in the "From: " line in the patch is leading, unless it's obviously invalid. Note that just using "git send-email" avoids a lot of the common issues with patch submission.

2 1

Re: d3drm: The input quaternions are not changed when calling D3DRMQuaternionSlerp
by Henri Verbeet Jan. 10, 2010

Jan. 10, 2010

2010/1/10 David Adam <david.adam.cnrs(a)gmail.com>: > + > +/* Test to show that the input quaternions are not changed */ > + expect_quat(q1,q1final); > + expect_quat(q2,q2final); > + You have some trailing spaces here.

1 0

Re: [1/4] avifil32/tests: Add initial tests
by Henri Verbeet Jan. 10, 2010

Jan. 10, 2010

2010/1/10 Julius Schwartzenberg <julius.schwartzenberg(a)gmail.com>: > This is Detlef's patch from June 2008 with just the import changed from > "vfw32" to "avifil32" to work around bug 14085. > > http://www.winehq.org/pipermail/wine-patches/2008-June/056310.html > You should add a proper "From: " line to the patch when submitting patches for other people. If you use "git format-patch" it will do this automatically, provided you gave correct author information to "git commit".

1 0

Re: [PATCH] shell32: handle bad menu handle (Coverity)
by Nikolay Sivov Jan. 10, 2010

Jan. 10, 2010

On 1/9/2010 16:59, Marcus Meissner wrote: > Hi, > > handle "invalid menu" return -1. Now with test. > > Ciao, Marcus > --- > dlls/shell32/shlmenu.c | 5 +++-- > dlls/shell32/tests/shlfileop.c | 13 +++++++++++++ > 2 files changed, 16 insertions(+), 2 deletions(-) > > > { > diff --git a/dlls/shell32/tests/shlfileop.c b/dlls/shell32/tests/shlfileop.c > index 0a5cdca..4d74b93 100644 > --- a/dlls/shell32/tests/shlfileop.c > +++ b/dlls/shell32/tests/shlfileop.c > Wrong file to place test in I think. > @@ -2224,6 +2224,17 @@ static void test_unicode(void) > ok(!file_existsW(UNICODE_PATH), "The directory should have been removed\n"); > } > > +extern HRESULT WINAPI Shell_MergeMenus (HMENU hmDst, HMENU hmSrc, UINT uInsert, UINT uIDAdjust, UINT uIDAdjustMax, ULONG uFlags); > + > This should got to shlobj.h. > +static void > +test_shlmenu(void) { > + HRESULT hres; > + hres = Shell_MergeMenus (0, 0, 0x42, 0x4242, 0x424242, 0); > + ok (hres == 0x4242, "expected 0x4242 but got %x\n", hres); > + hres = Shell_MergeMenus ((HMENU)42, 0, 0x42, 0x4242, 0x424242, 0); > + ok (hres == 0x4242, "expected 0x4242 but got %x\n", hres); > +} > + > Use a valid empty menu handle here as first parameter. Also I'd like to use more sane const values. > START_TEST(shlfileop) > { > InitFunctionPointers(); > @@ -2263,4 +2274,6 @@ START_TEST(shlfileop) > clean_after_shfo_tests(); > > test_unicode(); > + > + test_shlmenu(); > } > Function is available starting XP SP2 per msdn, so dynamic load should be used.

3 4