On Nov 3, 2007 1:10 PM, Gerald Pfeifer <gerald(a)pfeifer.com> wrote:
> While removing the dead code I noticed that we can actually also
> strengthen the const-ness of this function.
>
> Gerald
>
> ChangeLog:
> Remove a check which never could trigger (due to the domain of the
> variable in question) and increase const-ness of STREAMS_CreateView().
>
> Index: dlls/msi/streams.c
> ===================================================================
> RCS file: /home/wine/wine/dlls/msi/streams.c,v
> retrieving revision 1.7
> diff -u -3 -p -r1.7 streams.c
> --- dlls/msi/streams.c 18 Oct 2007 13:00:57 -0000 1.7
> +++ dlls/msi/streams.c 3 Nov 2007 18:06:27 -0000
> @@ -430,7 +430,7 @@ static UINT add_streams_to_table(MSISTRE
> return count;
> }
>
> -UINT STREAMS_CreateView(MSIDATABASE *db, MSIVIEW **view)
> +UINT STREAMS_CreateView(const MSIDATABASE *db, MSIVIEW **view)
> {
> MSISTREAMSVIEW *sv;
>
> @@ -444,9 +444,6 @@ UINT STREAMS_CreateView(MSIDATABASE *db,
> sv->db = db;
> sv->num_rows = add_streams_to_table(sv);
>
> - if (sv->num_rows < 0)
> - return ERROR_FUNCTION_FAILED;
> -
> *view = (MSIVIEW *)sv;
>
> return ERROR_SUCCESS;
>
This change is wrong. If you'd actually read what the code intended
to do instead of just fixing warnings, you'd see that
add_streams_to_table returns -1 on error. While the check for < 0 is
not correct, removing the check entirely is wrong. The check should
be if (sv->num_rows == -1) return ERROR_FUNCTION_FAILED.
--
James Hawkins
Howdy,
I just finished moving over the last of the obsolete components.
_obsolete_binary, _obsolete_directx, and _obsolete_gui can all be
removed now.
-Austin
James McKenzie wrote:
> Please submit a git diff not a UNIX diff.
Now, hang on there. Jens submitted a perfectly
good unified diff patch taken from the right directory.
There is no reason to turn up our noses at it.
It will apply using patch in exactly the same way as
a diff generated by git.
Jens, you're fine, I wouldn't bother rediffing. You might
however resend your patch next week when Alexandre
the maintainer is back from vacation!
- Dan
These tests are all failing consistently with identical failures on
these platforms:
combobox - 21 failures; all identical across the platforms.
datetime - 2 failures; all identical across the platforms.
The rebar tests have 146 failures on some (Win95 and NT4) and 545
failures on others (Win98 and Win2K). This looks like it is due to a
different comctl32 version as a result of a different IE version.
Several of the others have failures on Win95 and Win98 only.
Is anyone going to look at these?
- Reece
I have created a /dlls/d3dx8/texture.c file where to put all functions
related to textures and d3dx8.
I have included D3DXIMAGE_FILEFORMAT in the patch because it is referred
by D3DXIMAGE_FILEFORMAT which is referred in the function. If not
included, it would not compile. They both are now in /include/d3dx8.h
whci I think is the appropiate file. Do you propose to put them elsewhere?
Do you think I am missing anything or is it ok?
Hi folks,
I've been trying some debugging with kdbg (Graphical kde debugger), but
it fails on wine because it can't attach to a process.
While it is of course possible to use WINEDEBUG flags, I want to be able
to use a full debugger, and not just deduce from logs what happened, but
rather be able to see the corresponding lines of source as well. so
would it be possible to make gdb work with winedbg again, and perhaps
make it understand threads this time?
Cheers,
Maarten.
Hello,
i just noticed that the test results for advapi32:service tests are not
available for the last build (http://test.winehq.org/data/200801301937/). It
looks like the parser did not create the files.
Can someone familar with this stuff have a look?
Thanks Stefan