Here's an odd one. On my Levnovo x60 running Hardy,
with or without metacity-2.23.8 built from source,
the user32 test input.c fails when run from make
nearly all of the time, but doesn't fail if run standalone.
The problem persists even with a trivial makefile, e.g.
foo.ok:
env WINETEST_DEBUG=1 WINETEST_PLATFORM=wine ~/wine-git/wine
user32_test.exe.so input.c && touch foo.ok
This usually fails with
input.c:648: Test failed: 49: a1 from 00 -> 00 instead of 00 -> 80
or
On Wed, Sep 24, 2008 at 1:12 PM, Louis. Lenders
<xerox_xerox2000(a)yahoo.co.uk> wrote:
> This is the 2nd stub needed to let .net 2.0 SP1 installer finish fine. I'll
> resend the first one (corrected)
>
The vast majority of this file does not put spaces around parenthesis,
so can you resend with that fixed?
+UINT WINAPI MsiDetermineApplicablePatchesW( LPCWSTR szProductPackagePath,
+ DWORD cPatchInfo,
PMSIPATCHSEQUENCEINFOW pPatchInfo)
Also, why don't you add MsiDetermineApplicablePatchesA at the same time?
--
James Hawkins
On Wed, Sep 24, 2008 at 10:06 AM, James Hawkins
<jhawkins(a)codeweavers.com> wrote:
> Hi,
>
> Changelog:
> * Fix a failing test in win95.
>
> dlls/secur32/tests/main.c | 3 ++-
> 1 files changed, 2 insertions(+), 1 deletions(-)
>
> --
> James Hawkins
>
But if the result is SEC_E_UNSUPPORTED_FUNCTION it's not broken is it?
Shouldn't we use some kind of skip here? (The end result will be the
same, as we should use win_skip, it just doesn't look correct).
Cheers,
Paul.
Andrew Fenn wrote:
> I just submitted a patch to wine patches (Started implementing the
> xinput library for xbox 360 joystick support ) and I believe patch
> watcher failed for a few reasons.
>
>1) Although it applied my patch it didn't build it as
>./tools/make_makefiles isn't called so it isn't built?
I think make_makefiles is called:
http://code.google.com/p/winezeug/source/browse/trunk/patchwatcher/patchwat…
Maybe I should make the log more verbose...
> 2) Patch watcher seems to fail because of another patch, perhaps the
> latest known good build should be tested if the patch fails and then
> compared with the first failed attempt?
We don't accumulate patches. There was no other patch active
when yours was tested. This was probably just a rare failure
of a slightly flaky test. When I verify that tests are flaky, I
add them to a blacklist. You just got unlucky, I think.
Once you figure out why your patch didn't trigger your stuff
to be built, resubmit...
- Dan
2008/9/24 Francois Gouget <fgouget(a)free.fr>:
> ---
>
> On my system I have libgnutls11-dev 1.0.16-14+b1 which has
> gnutls_certificate_credentials but not the corresponding ..._t type. So
> this patch lets it compile on my system. Hopefully it will still work
> with the newer headers (otherwise we'll have to introduce a configure
> check :-( )
>
I don't think that'll work, this is what newer versions of gnutls have:
struct gnutls_certificate_credentials_st;
typedef struct gnutls_certificate_credentials_st
*gnutls_certificate_credentials_t;
typedef gnutls_certificate_credentials_t
gnutls_certificate_server_credentials;
typedef gnutls_certificate_credentials_t
gnutls_certificate_client_credentials;
gnutls_certificate_client_credentials and
gnutls_certificate_server_credentials should be available in both
versions though. Considering we're only setting client credentials
anyway for the moment, just using
gnutls_certificate_client_credentials would probably be fine.