April 2014 - Wine-devel - list.winehq.org

Re: typo fixes - rebased
by Marvin 16 Apr '14

16 Apr '14

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=6325 Your paranoid android. === w2000pro (32 bit shlfileop) === Failure running script in VM: network read timed out (wait2:ListSize:0/4)

1 0

Re: user32: Fix DrawMenuBar test failures
by Marvin 16 Apr '14

16 Apr '14

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=6320 Your paranoid android. === w2000pro (32 bit msg) === Failure running script in VM: network read timed out (wait2:ListSize:0/4)

1 0

Re: [2/2] kernel32: Add a workaround for broken apps that pass negative values to ReadConsole.
by Alexandre Julliard 15 Apr '14

15 Apr '14

Dmitry Timoshkov <dmitry(a)baikal.ru> writes: > --- > dlls/kernel32/console.c | 7 +++++++ > dlls/kernel32/tests/console.c | 6 ------ > 2 files changed, 7 insertions(+), 6 deletions(-) Still no good I'm afraid: ../../../../wine/tools/runtest -q -P wine -T ../../.. -M kernel32.dll -p kernel32_test.exe.so console && touch console.ok console.c:2581: Test failed: expected 0xdeadbeef, 0 make: *** [console.ok] Error 1 -- Alexandre Julliard julliard(a)winehq.org

1 0

Application for GSoC 2014 - riched20: implement ITextDocument::Range and ITextRange
by Jactry Zeng 15 Apr '14

15 Apr '14

Hello folks, I am Jactry Zeng, a junior who majoring in Electronic Science from China. And also a very fresh developer of Wine. My nickname in #winehackers is jactry. :) I would like to apply for this year's GSoC to work for Wine. I have entered Wine's GSoC last year. What I have work for is ITextDocument, I implemented some functions of it. I got some experience of Wine development and fixed some other bugs[0] of richedit during that time. I continued the work of ITextDocument after GSoC, I know there is still a lot of work worth to do. Such as ITextDocument::Range[1], which is used to get an ITextRange object of the content. And functions of ITextRange[2] are used to operate this range of text. Now in Wine, ITextRange was implemented mix with ITextSelection[3][4]. And ITextDocument::GetSelection[5] which is used to get an ITextSelection object is similar to ITextRange. In an easy way, we can implement ITextDocument::Range using some cast (I have a dirty hack in attachment). But I don't think it is a permanent solution. There are four parts in my plan: --- Part 0: Do some COM cleanup. --- Because ITextRange is implemented mix with ITextSelection now, so we should do some COM cleanup to create an ITextRange interface independently. And create a ITextSelection interface inherits from ITextRange. This is just the preparatory work for implementing ITextRange and I will begin this part before GSoC. ( So it is signed as part zore :b ) --- Part 1: Implement ITextDocument::Range. --- I will write more tests for ITextDocument::Range and ITextDocument::GetSelection first. Than try to implement ITextDocumen::Range. When this part is finished, bug 12458[6] and bug 18303[7] will benefit from it. --- Part 2: Implement functions of ITextRange --- In this part, I will try my best to implement some functions of ITextRange. I will write some tests for them first and than implement them. I think this is the hard part. And these functions will be implemented preferentially: GetText, GetDuplicate, GetStar, GetEnd, Copy, SetText. Because GetDuplicate and GetEnd were needed by bug 18303, and these functions seem was depend on the others. After these functions were implemented, I think bug 18303 will be fixed! ( Oh, this one of my pursues! ;) ) --- Part 3: Imporve all the code for being merged by upstream. --- This is the last part, I will improve all the patches for being merged by upstream, including what I did last year. After this part is finished, Wine will really has an initial implementation of ITextDocument and ITextRange. Maybe it will spend a lot of time, and I will continue this work after GSoC if I can't finish all of this part during GSoC. I think after all these are finished not only bug 18303 and bug 12458 will benefit from it, but also many other softwares. Because TOM interfaces are so important and be used frequently in real world applications. I appreciate for any comment. Thank you. :) attachments: dirty hack of ITextDocument::Range: dirty_hack.txt patch 0001: tests for ITextDocument_Range [0] http://source.winehq.org/git/wine.git/?a=search&h=HEAD&st=author&am…<http://source.winehq.org/git/wine.git/?a=search&h=HEAD&st=author&s=jactry> [1] http://msdn.microsoft.com/en-us/library/windows/desktop/bb774097(v=vs.85).a… [2] http://msdn.microsoft.com/en-us/library/windows/desktop/bb774058(v=vs.85).a… [3] http://msdn.microsoft.com/en-us/library/windows/desktop/bb774060(v=vs.85).a… [4] http://source.winehq.org/git/wine.git/blob/HEAD:/dlls/riched20/richole.c#l7… [5] http://msdn.microsoft.com/en-us/library/windows/desktop/bb774013(v=vs.85).a… [6] http://bugs.winehq.org/show_bug.cgi?id=12458 [7] http://bugs.winehq.org/show_bug.cgi?id=18303 -- Regards, Jactry Zeng

1 1

Re: user32: Fix DrawMenuBar test failures
by Alexandre Julliard 15 Apr '14

15 Apr '14

Piotr Caban <piotr(a)codeweavers.com> writes: > @@ -4819,6 +4819,8 @@ static void test_messages(void) > after_end_dialog = FALSE; > test_def_id = FALSE; > > + ok(SetCursorPos(109, 109), "SetCursorPos failed\n"); > + Please try to restore the cursor position after the test. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [PATCH 1/5] d3d9/tests: Add a test for timestamp queries. (try 2)
by Henri Verbeet 15 Apr '14

15 Apr '14

On 14 April 2014 22:51, Matteo Bruni <mbruni(a)codeweavers.com> wrote: > + hr = IDirect3DDevice9_CreateQuery(device, D3DQUERYTYPE_TIMESTAMPDISJOINT, &disjoint_query); > + ok(hr == D3D_OK || hr == D3DERR_NOTAVAILABLE, "Got unexpected hr %#x.\n", hr); > + data_size = IDirect3DQuery9_GetDataSize(disjoint_query); ... > + hr = IDirect3DDevice9_CreateQuery(device, D3DQUERYTYPE_TIMESTAMP, &query); > + ok(hr == D3D_OK || hr == D3DERR_NOTAVAILABLE, "Got unexpected hr %#x.\n", hr); > + data_size = IDirect3DQuery9_GetDataSize(query); I don't think this really works. Either creating these queries can legitimately fail and you need to handle those failures, or TIMESTAMPDISJOINT and TIMESTAMP are always supported if TIMESTAMPFREQ is, and D3DERR_NOTAVAILABLE is not a legitimate return value.

1 0

Re: [PATCH 3/5] wined3d: Add support for timestamp queries. (try 2)
by Henri Verbeet 15 Apr '14

15 Apr '14

On 14 April 2014 22:52, Matteo Bruni <mbruni(a)codeweavers.com> wrote: > + D3DQUERYTYPE type; ... > + type = wined3d_query_get_type(query->wined3d_query); > + if (type == D3DQUERYTYPE_TIMESTAMPDISJOINT) > + ret = sizeof(BOOL); > + else > + ret = wined3d_query_get_data_size(query->wined3d_query); Strictly speaking wined3d_query_get_type() returns an enum wined3d_query_type. > + if (type == D3DQUERYTYPE_TIMESTAMPDISJOINT && data && size == sizeof(BOOL)) > + { > + struct wined3d_query_data_timestamp_disjoint data_disjoint; > + > + hr = wined3d_query_get_data(query->wined3d_query, &data_disjoint, sizeof(data_disjoint), flags); > + *(BOOL *)data = data_disjoint.disjoint; > + } > + else > + { > + hr = wined3d_query_get_data(query->wined3d_query, data, size, flags); > + } Does this do the right thing if an application were to pass sizeof(data_disjoint) as size? > + if (gl_info->supported[ARB_TIMER_QUERY]) > + { > + GL_EXTCALL(glGenQueriesARB(1, &query->id)); > + checkGLcall("glGenQueriesARB"); > + > + TRACE("Allocated timestamp query %u in context %p.\n", query->id, context); > + } > + else > + { > + WARN("Timestamp queries not supported, not allocating query id.\n"); > + query->id = 0; > + } This should never happen, so this should either be an ERR(), or just be removed. Though I'll grant that e.g. the occlusion query code has some room for improvement as well in that regard. > @@ -969,6 +1024,9 @@ static void context_destroy_gl_resources(struct wined3d_context *context) > GL_EXTCALL(glDeleteProgramsARB(1, &context->dummy_arbfp_prog)); > } > > + if (gl_info->supported[ARB_TIMER_QUERY]) > + GL_EXTCALL(glDeleteQueriesARB(context->free_timestamp_query_count, context->free_timestamp_queries)); > + What about queries that are still on the context->timestamp_queries list? > +static HRESULT wined3d_timestamp_query_ops_get_data(struct wined3d_query *query, ... > + if (!gl_info->supported[ARB_TIMER_QUERY]) > + { > + WARN("%p timestamp queries not supported. Returning 0.\n", query); > + if (u64data) > + *u64data = 0; > + return S_OK; > + } ... > +static HRESULT wined3d_timestamp_query_ops_issue(struct wined3d_query *query, DWORD flags) ... > + if (gl_info->supported[ARB_TIMER_QUERY]) > + { ... > + } > + else > + { > + FIXME("%p timestamp queries not supported.\n", query); > + } Similar to above, the query should never have been created in these cases.

1 0

Re: typo fixes - rebased
by Nikolay Sivov 14 Apr '14

14 Apr '14

> diff --git a/ANNOUNCE b/ANNOUNCE It's generated, I don't think it makes sense to fix it. > ref = IHTMLLocation_Release(location); > - ok(!ref, "location chould be destroyed here\n"); > + ok(!ref, "location could be destroyed here\n"); Not exactly what it's supposed to mean. > -# @ stub RtlGetLengthWithoutTrailingPathSeperators > +# @ stub RtlGetLengthWithoutTrailingPathSeparators This is wrong. > -static void test_Mod( int line, VARIANT *left, VARIANT *right, VARIANT *expected, HRESULT expres ) > +static void test_Mod( int line, VARIANT *left, VARIANT *right, VARIANT *expected, HRESULT express ) Wrong too. > -#define NETWORKPLACESSHELLVIEWCOLUMNS sizeof(networkplaces_header)/sizeof(shvheader) > +#define NETWORKPLACCESSHELLVIEWCOLUMNS sizeof(networkplaces_header)/sizeof(shvheader) Network placcess hell view columns.

1 0

Re: wined3d: Add NVIDIA Geforce GTX 750 to device list. (try 2)
by Bruno Jesus 14 Apr '14

14 Apr '14

While you are at it, please if possible take a look at bugs: http://bugs.winehq.org/show_bug.cgi?id=35847 http://bugs.winehq.org/show_bug.cgi?id=32694 Best regards, Bruno

1 0

Re: [2/2] kernel32: Handle memory allocation failure in ReadConsoleA.
by Alexandre Julliard 14 Apr '14

14 Apr '14

Dmitry Timoshkov <dmitry(a)baikal.ru> writes: > @@ -2566,8 +2566,6 @@ static void test_ReadConsole(void) > ok(ret == INVALID_FILE_SIZE, "expected INVALID_FILE_SIZE, got %#x\n", ret); > ok(GetLastError() == ERROR_INVALID_HANDLE, "expected ERROR_INVALID_HANDLE, got %d\n", GetLastError()); > > -if (0) /* FIXME: uncomment once Wine doesn't hang forever */ > -{ It still hangs here on 64-bit. -- Alexandre Julliard julliard(a)winehq.org

1 0