October 2013 - Wine-devel - list.winehq.org

Re: [4/5] gdi32: Fix the B spacing value of empty glyph. (try 2)
by Alexandre Julliard 01 Oct '13

01 Oct '13

Akihiro Sagawa <sagawa.aki(a)gmail.com> writes: > Try 2: > - Get rid of redundant trace message in the last patch. > > --- > dlls/gdi32/freetype.c | 2 +- > dlls/gdi32/tests/font.c | 8 +++++++- > 2 files changed, 8 insertions(+), 2 deletions(-) This is causing failures here: ../../../tools/runtest -q -P wine -M gdi32.dll -T ../../.. -p gdi32_test.exe.so font.c && touch font.ok font.c:1133: Test succeeded inside todo block: LTR -1 advanced: abcA's sign should be unchanged font.c:1141: Test succeeded inside todo block: LTR -1 advanced: abcfA's sign should be unchanged font.c:1133: Test succeeded inside todo block: RTL 1 advanced: abcA's sign should be unchanged font.c:1141: Test succeeded inside todo block: RTL 1 advanced: abcfA's sign should be unchanged -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: kernel32/tests: Fix the test_waittxempty() timeout.
by Francois Gouget 01 Oct '13

01 Oct '13

On Tue, 1 Oct 2013, Francois Gouget wrote: [...] > The timeout was too short and would only work for UARTs that have a > FIFO and where the trigger level was set just right. With this patch > we calculate how much time sending the test data will actually take > and tack on a small margin to take into account scheduling delays. I should note that this does not fix the test failuers on the QEmu VMs, specifically the WineTestBot ones. If QEmu 1.1.2 is anything like QEmu 1.6.0, then the serial port seems to operate at a fixed speed of about 1800000bps. This is so out of line with the 150bps that we request that it causes the test to fail. [...] > * Windows reports that all bytes have been sent not when the UART's > buffer is empty but when they all made it to the UART's buffer. Note > that this can be tested (see next mail). And here is this test. IT IS NOT MEANT TO BE COMMITTED. (but seeing the WineTestBot results would still be interesting) On my one Windows PC with a real UART (Intel(R) 82801DBM LPC Interface Controller - 24CC, i.e. ICH4 which emulates a 16550A UART) I get the following results which clearly illustrate the impact of the FIFO/trigger level. comm.c:1100: WaitCommEvent(1 bytes) for EV_TXEMPTY took 0 ms comm.c:1100: WaitCommEvent(2 bytes) for EV_TXEMPTY took 0 ms comm.c:1100: WaitCommEvent(3 bytes) for EV_TXEMPTY took 0 ms comm.c:1100: WaitCommEvent(4 bytes) for EV_TXEMPTY took 0 ms comm.c:1100: WaitCommEvent(5 bytes) for EV_TXEMPTY took 0 ms comm.c:1100: WaitCommEvent(6 bytes) for EV_TXEMPTY took 0 ms comm.c:1100: WaitCommEvent(7 bytes) for EV_TXEMPTY took 0 ms comm.c:1100: WaitCommEvent(8 bytes) for EV_TXEMPTY took 0 ms comm.c:1100: WaitCommEvent(9 bytes) for EV_TXEMPTY took 0 ms comm.c:1100: WaitCommEvent(10 bytes) for EV_TXEMPTY took 0 ms comm.c:1100: WaitCommEvent(11 bytes) for EV_TXEMPTY took 0 ms comm.c:1100: WaitCommEvent(12 bytes) for EV_TXEMPTY took 0 ms comm.c:1100: WaitCommEvent(13 bytes) for EV_TXEMPTY took 0 ms comm.c:1100: WaitCommEvent(14 bytes) for EV_TXEMPTY took 0 ms comm.c:1100: WaitCommEvent(15 bytes) for EV_TXEMPTY took 871 ms comm.c:1100: WaitCommEvent(16 bytes) for EV_TXEMPTY took 872 ms comm.c:1100: WaitCommEvent(17 bytes) for EV_TXEMPTY took 872 ms comm.c:1100: WaitCommEvent(18 bytes) for EV_TXEMPTY took 872 ms comm.c:1100: WaitCommEvent(19 bytes) for EV_TXEMPTY took 871 ms comm.c:1100: WaitCommEvent(20 bytes) for EV_TXEMPTY took 871 ms comm.c:1100: WaitCommEvent(21 bytes) for EV_TXEMPTY took 871 ms comm.c:1100: WaitCommEvent(22 bytes) for EV_TXEMPTY took 871 ms comm.c:1100: WaitCommEvent(23 bytes) for EV_TXEMPTY took 871 ms comm.c:1100: WaitCommEvent(24 bytes) for EV_TXEMPTY took 871 ms comm.c:1100: WaitCommEvent(25 bytes) for EV_TXEMPTY took 871 ms comm.c:1100: WaitCommEvent(26 bytes) for EV_TXEMPTY took 871 ms comm.c:1100: WaitCommEvent(27 bytes) for EV_TXEMPTY took 871 ms comm.c:1100: WaitCommEvent(28 bytes) for EV_TXEMPTY took 871 ms comm.c:1100: WaitCommEvent(29 bytes) for EV_TXEMPTY took 1802 ms comm.c:1100: WaitCommEvent(30 bytes) for EV_TXEMPTY took 1803 ms comm.c:1100: WaitCommEvent(31 bytes) for EV_TXEMPTY took 1803 ms comm.c:1100: WaitCommEvent(32 bytes) for EV_TXEMPTY took 1802 ms comm.c:1100: WaitCommEvent(33 bytes) for EV_TXEMPTY took 1803 ms comm.c:1100: WaitCommEvent(34 bytes) for EV_TXEMPTY took 1802 ms comm.c:1100: WaitCommEvent(35 bytes) for EV_TXEMPTY took 1803 ms comm.c:1100: WaitCommEvent(36 bytes) for EV_TXEMPTY took 1803 ms comm.c:1100: WaitCommEvent(37 bytes) for EV_TXEMPTY took 1803 ms comm.c:1100: WaitCommEvent(38 bytes) for EV_TXEMPTY took 1803 ms comm.c:1100: WaitCommEvent(39 bytes) for EV_TXEMPTY took 1803 ms comm.c:1100: WaitCommEvent(40 bytes) for EV_TXEMPTY took 1803 ms comm.c:1100: WaitCommEvent(41 bytes) for EV_TXEMPTY took 1802 ms comm.c:1100: WaitCommEvent(42 bytes) for EV_TXEMPTY took 1803 ms comm.c:1100: WaitCommEvent(43 bytes) for EV_TXEMPTY took 2734 ms comm.c:1100: WaitCommEvent(44 bytes) for EV_TXEMPTY took 2734 ms comm.c:1100: WaitCommEvent(45 bytes) for EV_TXEMPTY took 2734 ms comm.c:1100: WaitCommEvent(46 bytes) for EV_TXEMPTY took 2734 ms comm.c:1100: WaitCommEvent(47 bytes) for EV_TXEMPTY took 2734 ms comm.c:1100: WaitCommEvent(48 bytes) for EV_TXEMPTY took 2734 ms comm.c:1100: WaitCommEvent(49 bytes) for EV_TXEMPTY took 2734 ms comm.c:1100: WaitCommEvent(50 bytes) for EV_TXEMPTY took 2734 ms -- Francois Gouget <fgouget(a)free.fr> http://fgouget.free.fr/ "Utilisateur" (nom commun) : Mot utilisé par les informaticiens en lieu et place d'"idiot".

1 0

Re: [3/3] ws2_32: Always return the source address from WSAAccept.
by Marvin 01 Oct '13

01 Oct '13

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://newtestbot.winehq.org/JobDetails.pl?Key=2421 Your paranoid android. === wxppro (32 bit sock) === sock.c:519: Test failed: oob_server (b0): unexpectedly at the OOB mark: 0 sock: unhandled exception c0000005 at 71AB8D62 === w2008s64 (32 bit sock) === sock.c:519: Test failed: oob_server (5a0): unexpectedly at the OOB mark: 0 === w2008s64 (64 bit sock) === sock.c:519: Test failed: oob_server (624): unexpectedly at the OOB mark: 0

1 0

Re: [2/3] ws2_32: Return an error from accept if the address buffer is too small.
by Marvin 01 Oct '13

01 Oct '13

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://newtestbot.winehq.org/JobDetails.pl?Key=2420 Your paranoid android. === w2008s64 (32 bit sock) === sock.c:519: Test failed: oob_server (5a0): unexpectedly at the OOB mark: 0 sock: unhandled exception c0000005 at 757A9799 === w7pro64 (32 bit sock) === sock.c:528: Test failed: oob_server (778): not at the OOB mark: 1 === w2008s64 (64 bit sock) === sock.c:519: Test failed: oob_server (7f8): unexpectedly at the OOB mark: 0

1 0

Re: [PATCH] dsound: fixup IDirectSoundCaptureBuffer_QueryInterface
by Andrew Eikum 01 Oct '13

01 Oct '13

Seems reasonable, but could you put together some tests to show this? Andrew On Sat, Sep 28, 2013 at 10:41:15AM +0200, Maarten Lankhorst wrote: > diff --git a/dlls/dsound/capture.c b/dlls/dsound/capture.c > index 40f1702..0fe300c 100644 > --- a/dlls/dsound/capture.c > +++ b/dlls/dsound/capture.c > @@ -51,7 +51,7 @@ typedef struct IDirectSoundCaptureBufferImpl > IDirectSoundCaptureBuffer8 IDirectSoundCaptureBuffer8_iface; > IDirectSoundNotify IDirectSoundNotify_iface; > LONG numIfaces; /* "in use interfaces" refcount */ > - LONG ref, refn; > + LONG ref, refn, has_dsc8; > /* IDirectSoundCaptureBuffer fields */ > DirectSoundCaptureDevice *device; > DSCBUFFERDESC *pdscbd; > @@ -241,8 +241,9 @@ static HRESULT WINAPI IDirectSoundCaptureBufferImpl_QueryInterface(IDirectSoundC > > *ppobj = NULL; > > - if ( IsEqualGUID( &IID_IDirectSoundCaptureBuffer, riid ) || > - IsEqualGUID( &IID_IDirectSoundCaptureBuffer8, riid ) ) { > + if ( IsEqualIID( &IID_IUnknown, riid ) || > + IsEqualIID( &IID_IDirectSoundCaptureBuffer, riid ) || > + (This->has_dsc8 && IsEqualIID( &IID_IDirectSoundCaptureBuffer8, riid )) ) { > IDirectSoundCaptureBuffer8_AddRef(iface); > *ppobj = iface; > return S_OK; > @@ -1239,6 +1240,8 @@ static HRESULT WINAPI IDirectSoundCaptureImpl_CreateCaptureBuffer(IDirectSoundCa > > if (hr != DS_OK) > WARN("IDirectSoundCaptureBufferImpl_Create failed\n"); > + else > + This->device->capture_buffer->has_dsc8 = This->has_dsc8; > > return hr; > } > > >

1 0

Re: [3/3] ws2_32: Always return the source address from WSAAccept.
by Bruno Jesus 01 Oct '13

01 Oct '13

On Tue, Oct 1, 2013 at 6:37 AM, Hans Leidekker <hans(a)codeweavers.com> wrote: > --- > dlls/ws2_32/socket.c | 9 +-------- > dlls/ws2_32/tests/sock.c | 6 +++--- > 2 files changed, 4 insertions(+), 11 deletions(-) > > diff --git a/dlls/ws2_32/socket.c b/dlls/ws2_32/socket.c > index c34de4b..6855298 100644 > --- a/dlls/ws2_32/socket.c > +++ b/dlls/ws2_32/socket.c > @@ -6546,12 +6546,8 @@ SOCKET WINAPI WSAAccept( SOCKET s, struct WS_sockaddr *addr, LPINT addrlen, > TRACE("Socket %04lx, sockaddr %p, addrlen %p, fnCondition %p, dwCallbackData %ld\n", > s, addr, addrlen, lpfnCondition, dwCallbackData); > > - > - size = sizeof(src_addr); > - cs = WS_accept(s, &src_addr, &size); > - > + cs = WS_accept(s, addr, addrlen); The parameter addr may be NULL in WSAAccept, I can't check the source code right now but what will WS_accept do in this case? Best wishes, Bruno

2 1

Re: [1/3] ws2_32: Always clear res on error in getaddrinfo/GetAddrInfoW.
by Marvin 01 Oct '13

01 Oct '13

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://newtestbot.winehq.org/JobDetails.pl?Key=2419 Your paranoid android. === wxppro (32 bit sock) === sock.c:519: Test failed: oob_server (2a4): unexpectedly at the OOB mark: 0 === w2008s64 (32 bit sock) === sock.c:519: Test failed: oob_server (774): unexpectedly at the OOB mark: 0 === w2008s64 (64 bit sock) === sock.c:519: Test failed: oob_server (884): unexpectedly at the OOB mark: 0 === w7pro64 (64 bit sock) === sock.c:509: Test failed: oob_server (af8): unexpectedly at the OOB mark: 0 sock.c:519: Test failed: oob_server (af8): unexpectedly at the OOB mark: 0 sock.c:288: Test failed: do_synchronous_recv (af8): error 10054: sock.c:534: Test failed: oob_server (af8): not at the OOB mark: 1

1 0

Re: [PATCH 2/5] wined3d: Move location management into the resource.
by Henri Verbeet 01 Oct '13

01 Oct '13

Please split this.

1 0

Re: [PATCH 2/5] wined3d: Move location management into the resource.
by Stefan Dösinger 01 Oct '13

01 Oct '13

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, To illustrate the goal of these patches I've attached an archive of all surface cleanup patches to this mail. The goal is to merge GL buffer object handling of buffers, surfaces and volumes. The main motivation for this is to handle WINED3D_MAP_DISCARD in one central place once dynamic resource handling is implemented in the command stream. The 4 patches in this series move location and buffer handling code from volume.c to resource.c. The next patches remove broken(overlay, partial rectangle handling) and half-broken(direct DIB maps, buffers) surface features. The next step is migrating surfaces to the resource location and loading code in resource.c. After that WINED3D_LOCATION_DISCARDED handling is improved and previously removed surface features re-added. Due to the removal of some surface features, performance regressions are possible. I think it's unlikely because the existing surface PBO code doesn't really do what it is supposed to do. Functional regressions should not happen. This patchset does not deal with d3d buffers. The central aspect of merging buffer handling is removing color and rhw conversion support. (or, alternatively, adding a separate location WINED3D_LOCATION_CONVBUFFER if we really want to keep the conversion code.) There are some problems later in the patchset that need fixing: Patch 24 is too big, patches 48 and 52 need to be moved. Non-GL ddraw rendering is broken at the moment (black screen). My patches don't make this any worse, but I'd like to fix it to make sure non-GL operation works as intended. Cheers, Stefan -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.20 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iQIcBAEBAgAGBQJSSqzhAAoJEN0/YqbEcdMww1wP/jYYo0B/igDHd0QwOkwk8DIz ip4PKHxRgqaSYzswcBGZge+iBrJrvXSHh3cScG9KJ1a/VeaRlPyhDsFDzPce1VDK 0IX7iF/gKpZazPVy21lSIPQMlErSVy58fkIsd/PqKV/zyU8zNyZDkYU3n80itYrq Y3hPhaKbJIDN8QbhIOZu0egaiHObcwomBgZCfrUv+hr+xf+gMERYbqN9gj2JWFwJ dx4EzdeTFw1CZkFExbAJKt5zWA2KeFyitVU887WOSJpa3UbqemGpOID8vnYZlABZ mmqrS22XWjXlo2+0s0NufxeQJHvzpkcludLkg8a12r4FDs+ORYYnGbUeUy2o2Rud uV2quoeucA1TEAyF20JCrVNuCvDz5nvQq19rzR51k7FF/IsKGiI823ZGA54F4Dkn DZkyDcE0rnI2Re+br1aM31ZUt/maTxQMpk10BAAjwz1jbXQ4plZ9dtO9/DObdw8r nfPDJz+DsuBv19Th7MPjTxvXRjxVFTzlfuXNk/cOfAhgNdpEFSOGCAexGi6moSq1 6slQeLb8PqmO/83JQBuz4OI86dSoGSC6O9c/TtWXdW1PgCDyHgFvNb3jmQHsf7ya U1o2TkkNhLvfYz5Eo85Ep8oyVIF09CDrqQzUo10q33K5Y2E3RM8Xbxc+crGtjgxA TGwCwR8SRvAW+DfeP7bq =MEk/ -----END PGP SIGNATURE-----

1 0

Re: [4/5] ntdll: Make asynchronous NtWriteFile on a disk file always return STATUS_PENDING.
by Alexandre Julliard 01 Oct '13

01 Oct '13

Dmitry Timoshkov <dmitry(a)baikal.ru> writes: > --- > dlls/ntdll/file.c | 3 ++- > dlls/ntdll/tests/file.c | 5 +++-- > 2 files changed, 5 insertions(+), 3 deletions(-) I don't see the point. Do you actually have an app that depends on this? -- Alexandre Julliard julliard(a)winehq.org

2 6