November 2017 - Wine-devel - list.winehq.org

When to mark a bug as STAGED
by Josh DuBois 06 Nov '17

06 Nov '17

I was recently looking at bug 18070 in relation to a problem I had installing Office 2010: https://bugs.winehq.org/show_bug.cgi?id=18070 The last comment asks whether the bug can be marked STAGED. In talking with Sebastian he questioned whether it would make sense in this case, because the patch in the staging tree which improves behavior is clearly a hack as opposed to something which is progress toward a more permanent solution. To me it made sense to go ahead and mark the bug as STAGED anyhow, to record the fact that one of the patches in the stating tree addresses the behavior. Perhaps this is one way to help associate patchsets with bugs so that we can tell the motivation for a particular patchset in staging. (In this case, a URL on the bug also has information about the motivation, here: https://dev.wine-staging.com/patches/132/ ). So the question: If a staging patchset is more along the lines of a dead-end hack than it is progress toward a long-term solution, should bugs which are fixed by the patchset still be marked STAGED? My gut differed from Sebastian's, so I thought it best to ask here to reach a consensus. (I can see the value either way - just looking for a standard approach.)

3 2

Re: [PATCH 3/5] gdiplus/metafile: Support linear gradient brushes in playback
by Vincent Povirk 06 Nov '17

06 Nov '17

+ GpLineGradient *gradient; ... + if (status == Ok) + *brush = (GpBrush *)gradient; + else + GdipDeleteBrush((GpBrush *)gradient); This can free gradient without initializing it.

1 0

Re: [PATCH 3/3] quartz/tests: Add test AVI and MPG files.
by Alexandre Julliard 06 Nov '17

06 Nov '17

Zebediah Figura <z.figura12(a)gmail.com> writes: > Generated with ffmpeg using the following commands: > > ffmpeg -f lavfi -i smptebars -t 1 -r 1 -f avi -vcodec rawvideo -pix_fmt yuv420p -vf scale=32x24 test.avi > > ffmpeg -f lavfi -i smptebars -f lavfi -i "sine=frequency=1000" -t 0.04 -r 25 -f mpeg -vcodec mpeg1video -vf scale=32x24 -acodec mp2 test.mpg > > Signed-off-by: Zebediah Figura <z.figura12(a)gmail.com> > --- > dlls/quartz/tests/Makefile.in | 3 +++ > dlls/quartz/tests/filtergraph.c | 42 +++++++++++++++++++++++++++++++++++----- > dlls/quartz/tests/rsrc.rc | 27 ++++++++++++++++++++++++++ > dlls/quartz/tests/test.avi | Bin 0 -> 6938 bytes > dlls/quartz/tests/test.mpg | Bin 0 -> 6144 bytes > 5 files changed, 67 insertions(+), 5 deletions(-) > create mode 100644 dlls/quartz/tests/rsrc.rc > create mode 100644 dlls/quartz/tests/test.avi > create mode 100644 dlls/quartz/tests/test.mpg It doesn't work here: ../../../tools/runtest -q -P wine -T ../../.. -M quartz.dll -p quartz_test.exe.so avisplitter && touch avisplitter.ok avisplitter.c:498: Test failed: The thread count should be 3 not 2 avisplitter.c:574: Test succeeded inside todo block: The thread count should be 1 not 1 Makefile:194: recipe for target 'avisplitter.ok' failed make: *** [avisplitter.ok] Error 2 -- Alexandre Julliard julliard(a)winehq.org

2 1

FOSDEM participation - talk?
by Marcus Meissner 06 Nov '17

06 Nov '17

Hi, We can still consider a FOSDEM participation in scope of a Talk/Lecture. We could fit to the Retrocomputing DevRoom with a winky smiley: - Emulation of old systems to run videogames, legacy software, etc - Open source software emulation/simulation - Open hardware - Operating systems/executives for retrocomputers/retrosystems - Uses of retrocomputing today - Other retrosystems topics - Opportunities in retrocomputing https://fosdem.org/2018/schedule/track/retrocomputing/ https://lists.fosdem.org/pipermail/fosdem/2017-October/002623.html Cfp runs until Nov 25. Ciao, Marcus

3 3

Re: [PATCH 2/2] riched20: Protect read-only Richedit against pasting and cutting data.
by Huw Davies 06 Nov '17

06 Nov '17

This looks good apart from some style issues - see inline below. On Sun, Nov 05, 2017 at 07:28:39PM +0100, Rafał Harabień wrote: > Signed-off-by: Rafał Harabień <rafalh1992(a)o2.pl> > --- > dlls/riched20/editor.c | 60 +++++++++++++++++++++++--------------------- > dlls/riched20/tests/editor.c | 15 +++++++++++ > 2 files changed, 47 insertions(+), 28 deletions(-) > > diff --git a/dlls/riched20/editor.c b/dlls/riched20/editor.c > index a7b1bc7..3da19fa 100644 > --- a/dlls/riched20/editor.c > +++ b/dlls/riched20/editor.c > @@ -2294,6 +2294,13 @@ static BOOL paste_special(ME_TextEditor *editor, UINT cf, REPASTESPECIAL *ps, BO > struct paste_format *format; > IDataObject *data; > > + /* Protect read-only edit control from modification */ > + if (editor->styleFlags & ES_READONLY) { Please put the open brace on a new line matching the rest of the function. > + if (!check_only) > + MessageBeep(MB_ICONERROR); > + return FALSE; > + } > + > init_paste_formats(); > > if (ps && ps->dwAspect != DVASPECT_CONTENT) > @@ -2351,6 +2358,29 @@ static BOOL ME_Copy(ME_TextEditor *editor, const ME_Cursor *start, int nChars) > return SUCCEEDED(hr); > } > > +static BOOL copy_or_cut(ME_TextEditor *editor, BOOL cut) > +{ > + BOOL result; > + int nOfs, nChars; > + int nStartCur = ME_GetSelectionOfs(editor, &nOfs, &nChars); > + ME_Cursor *selStart = &editor->pCursors[nStartCur]; Let's take the opportunity to use lower case variable names. How about offs, num_chars, start_cursor and sel_start. > + > + if (cut && (editor->styleFlags & ES_READONLY)) { Again, opening brace to new line. > + MessageBeep(MB_ICONERROR); > + return FALSE; > + } > + > + nChars -= nOfs; > + result = ME_Copy(editor, selStart, nChars); > + if (result && cut) > + { > + ME_InternalDeleteText(editor, selStart, nChars, FALSE); Indent this block 4 spaces. > + ME_CommitUndo(editor); > + ME_UpdateRepaint(editor, TRUE); > + } > + return result; > +} > + Huw.

1 0

Re: [v2] comctl32/taskdialog: Implement nDefaultButton and add tests
by Fabian Maurer 06 Nov '17

06 Nov '17

I reworked my patch according to your suggestions. >>> We already iterate over custom buttons and check every common button >>> flag before this point, so it shouldn't be necessary to have another loop. >> >> I'm open to suggestion, but how would you do that? >> >> We could add three checks each time before taskdialog_init_button is called, >> or alternatively one check in taskdialog_init_button. Like giving it a pointer >> to default_button, and checking if the id equal the taskdialog default button. >> But I find my current solution to be cleaner to be honest, what's your >> suggestion? >Yes, something like that. taskdialog_init_button() has all the info to > do that. In the beginning I wanted to have default_button be -1 if no button was found yet, but as the tests show this is impossible since windows accepts a negative ID. Now we have to pass 2 pointers around.. Your thoughts on that? I find that the second loop makes things a lot cleaner. Regards, Fabian Maurer

2 1

Re: [PATCH] msvcp100: add time_get::get function with format
by Piotr Caban 06 Nov '17

06 Nov '17

Hi, On 11/05/17 13:26, Vijay Kiran Kamuju wrote: > diff --git a/dlls/msvcp110/msvcp110.spec b/dlls/msvcp110/msvcp110.spec > index bf54ac5f53..efb13361ff 100644 > --- a/dlls/msvcp110/msvcp110.spec > +++ b/dlls/msvcp110/msvcp110.spec > @@ -2599,8 +2599,8 @@ > @ thiscall -arch=i386 ?get@?$time_get(a)DV?$istreambuf_iterator(a)DU?$char_traits(a)D@std@@@std@@@std@@QBE?AV?$istreambuf_iterator(a)DU?$char_traits(a)D@std@@@2(a)V32@0AAVios_base(a)2@AAHPAUtm@@DD(a)Z(ptr ptr ptr long ptr long ptr ptr ptr long long) time_get_char_get > @ cdecl -arch=win64 ?get@?$time_get(a)DV?$istreambuf_iterator(a)DU?$char_traits(a)D@std@@@std@@@std@@QEBA?AV?$istreambuf_iterator(a)DU?$char_traits(a)D@std@@@2(a)V32@0AEAVios_base(a)2@AEAHPEAUtm@@DD(a)Z(ptr ptr ptr long ptr long ptr ptr ptr long long) time_get_char_get > @ stub -arch=arm ?get@?$time_get(a)DV?$istreambuf_iterator(a)DU?$char_traits(a)D@std@@@std@@@std@@QBA?AV?$istreambuf_iterator(a)DU?$char_traits(a)D@std@@@2(a)V32@0AAVios_base(a)2@AAHPAUtm@@PBD4(a)Z > -@ stub -arch=i386 ?get@?$time_get(a)DV?$istreambuf_iterator(a)DU?$char_traits(a)D@std@@@std@@@std@@QBE?AV?$istreambuf_iterator(a)DU?$char_traits(a)D@std@@@2(a)V32@0AAVios_base(a)2@AAHPAUtm@@PBD4(a)Z > -@ stub -arch=win64 ?get@?$time_get(a)DV?$istreambuf_iterator(a)DU?$char_traits(a)D@std@@@std@@@std@@QEBA?AV?$istreambuf_iterator(a)DU?$char_traits(a)D@std@@@2(a)V32@0AEAVios_base(a)2@AEAHPEAUtm@@PEBD4(a)Z > +@ thiscall -arch=i386 ?get@?$time_get(a)DV?$istreambuf_iterator(a)DU?$char_traits(a)D@std@@@std@@@std@@QBE?AV?$istreambuf_iterator(a)DU?$char_traits(a)D@std@@@2(a)V32@0AAVios_base(a)2@AAHPAUtm@@PBD4(a)Z(ptr ptr ptr long ptr long ptr ptr ptr ptr ptr) time_get_char_get_fmt > +@ cdecl -arch=win64 ?get@?$time_get(a)DV?$istreambuf_iterator(a)DU?$char_traits(a)D@std@@@std@@@std@@QEBA?AV?$istreambuf_iterator(a)DU?$char_traits(a)D@std@@@2(a)V32@0AEAVios_base(a)2@AEAHPEAUtm@@PEBD4(a)Z(ptr ptr ptr long ptr long ptr ptr ptr ptr ptr) time_get_char_get_fmt Please also add arm export. Also win64 arguments needs fixing (structures are passed as pointers in this case). > +istreambuf_iterator_char* __thiscall time_get_char_get_fmt(const time_get_char *this, > + istreambuf_iterator_char *ret, istreambuf_iterator_char s, istreambuf_iterator_char e, > + ios_base *base, int *err, struct tm *t, char *fmtstart, char *fmtend) The fmtstart and fmtend arguments should be const. Thanks, Piotr

1 0

Re: [v2 PATCH] cmd: Implement builtin mklink command
by Alexandre Julliard 06 Nov '17

06 Nov '17

Alistair Leslie-Hughes <leslie_alistair(a)hotmail.com> writes: > @@ -310,6 +310,15 @@ CHOICE is mainly used to build a menu selection in a batch file.\n" > "EXIT terminates the current command session and returns to the operating\n\ > system or shell from which you invoked cmd.\n" > > + WCMD_MKLINK, > +"Creates a symbolic link.\n\ > +MKLINK [[/D] | [/H] | [/J]] Link Target\n\ > +/D Creates a directory symbolic link. Default is a file symbolic link.\n\ > +/H Creates a hard link instead of a symbolic link.\n\ > +/J Creates a Directory Junction.\n\ > +Link specifies the new symbolic link name.\n\ > +Target specifies the path (relative or absolute) that the new link refers to.\n" This is a straight copy from the Microsoft version. Not only is this not allowed, it also doesn't fit in with the description style of the other commands. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [PATCH] user32: check GetClipboardFormatName length
by Marvin 06 Nov '17

06 Nov '17

Hi, While running your changed tests on Windows, I think I found new failures. Being a bot and all I'm not very good at pattern recognition, so I might be wrong, but could you please double-check? Full results can be found at https://testbot.winehq.org/JobDetails.pl?Key=33944 Your paranoid android. === w8 (32 bit clipboard) === clipboard.c:402: Test failed: GlobalGetAtomNameA should fail clipboard.c:403: Test failed: err -559038737

2 1

Re: [v3] cmd: Added support for "&&" and "||" condition checking.
by Marvin 06 Nov '17

06 Nov '17

Hi, While running your changed tests on Windows, I think I found new failures. Being a bot and all I'm not very good at pattern recognition, so I might be wrong, but could you please double-check? Full results can be found at https://testbot.winehq.org/JobDetails.pl?Key=33950 Your paranoid android. === build (build) === Patch failed to apply

1 0