You should write some tests to examine the behavior of VerQueryValue
with NULL puLen.
I just found out that wine has a test for it, but it has been commented out:
if (0)
{
/* this test crashes on WinNT4
*/
boolret = VerQueryValueA( pVersionInfo, backslash, (LPVOID
*)&pFi
Quentin Paris informs me that the root directory is
user-writable on MacOSX (shudder), which means
our beloved z: mapping leads to lots of garbage
directories named e.g. /fd91216cc2ce81eb2ce16a058619
after running msi installers.
Is this something Wine's msi could avoid doing?
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
http://testbot.winehq.org/JobDetails.pl?Key=8130
Your paranoid android.
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
http://testbot.winehq.org/JobDetails.pl?Key=8127
Your paranoid android.
On Mon, 3 Jan 2011, Alexandre Julliard wrote:
>> I verified this does not cause any warning on FreeBSD 8.1 test
>> builds, and all the tools like bison and flex in somewhat current
>> versions.
> It's broken here with bison 2.4.1:
>
> make[1]: Entering directory `/home/julliard/wine/wine/libs/wpp'
Alexandre wrote:
>> +#define WINE_ONCE(do_once) ...
>
> I don't think that buys anything over spelling it out explicitly.
It makes it easier to grep for, which might be useful if you
want to find all the places it's used to throttle messages
that shouldn't be happening in the first place.
Sven Baars writes:
> @@ -250,6 +250,12 @@ static inline const char *wine_dbgstr_longlong(
> ULONGLONG ll )
> #define WINE_DPRINTF wine_dbg_printf
> #define WINE_MESSAGE wine_dbg_printf
>
> +#define WINE_ONCE(do_once) \
> +do{ static int wine_once_var; \
> +
André Hentschel writes:
> @@ -2480,7 +2498,9 @@ sub generate_project_files($)
>my $canon=canonize("@$target[$T_NAME]");
>$canon =~ s/_so$//;
>
> - if (@$target[$T_TYPE] == $TT_DLL) {
> + if (@$target[$T_TYPE] == $TT_DLL && (@$project_settings[$T_FLAGS] &
> $TF_HASDEF)
André Hentschel writes:
> @@ -232,6 +233,10 @@ BOOL WINAPI InitiateSystemShutdownExA( LPSTR
> lpMachineName, LPSTR lpMessage,
> DWORD dwTimeout, BOOL bForceAppsClosed, BOOL bRebootAfterShutdown,
> DWORD dwReason)
> {
> + if (!lpMachineName && !lpMessage && !dwTimeout)
>
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Yep, the patch(f41fdb22...) should be reverted.
As I explained on IRC, the correct approach is to implement a replacement
shader for the vertex pipeline and implement vertex blending in it. Henri has a
rebased version of my ARBvp experiments from 20
Gerald Pfeifer writes:
> Checking this via autoconf sounds a bit involved. Do you have a
> recommendation? Perhaps just wait until "Alexandre runs bison 2.4.2
> or later on all his machines"? ;-)
Yes, but maybe filing a bug with Debian could help speeding up the
process...
--
Alexandre Julli
11 matches
Mail list logo