September 2013 - Wine-Devel

Re: user32: Avoid using CONST.
by Marvin Sept. 25, 2013

Sept. 25, 2013

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=2332 Your paranoid android. === build (build) === Patch failed to apply

1 0

Re: [2/5] winex11: Splitted functions X11DRV_FocusOut and X11DRV_ReparentNotify
by Alexandre Julliard Sept. 25, 2013

Sept. 25, 2013

Sebastian Lackner <sebastian(a)fds-team.de> writes: > --- > dlls/winex11.drv/event.c | 99 > +++++++++++++++++++++++++++------------------- > 1 file changed, 58 insertions(+), 41 deletions(-) These should go in the patches that need them. Also you should start your series with reparent support, the rest doesn't make sense without it. -- Alexandre Julliard julliard(a)winehq.org

2 1

Re: [PATCH] d3dx9: Use struct d3dx_object for objects.
by Matteo Bruni Sept. 25, 2013

Sept. 25, 2013

2013/9/24 Rico Schüller <kgbricola(a)web.de>: > --- > dlls/d3dx9_36/effect.c | 248 > +++++++++++++++++++------------------------------ > 1 Datei geändert, 95 Zeilen hinzugefügt(+), 153 Zeilen entfernt(-) > I definitely like the direction this patch takes. @@ -5068,6 +5009,9 @@ static HRESULT d3dx9_parse_resource(struct d3dx9_base_effect *base, const char * param = state->parameter; + /* + * TODO: Do we need to create the shader/string here or later when we access them? + */ Here it's fine AFAICS. What are the other options you are thinking about? Something like creating the shaders at the first Begin / BeginPass call?

3 2

Re: wininet: Don't assume that end of chunk means end of stream. (try 2)
by Jacek Caban Sept. 25, 2013

Sept. 25, 2013

Hi Hans, On 09/18/13 13:40, Hans Leidekker wrote: > diff --git a/dlls/wininet/tests/http.c b/dlls/wininet/tests/http.c > index bf57223..2526681 100644 > --- a/dlls/wininet/tests/http.c > +++ b/dlls/wininet/tests/http.c > @@ -4777,8 +4777,37 @@ static const struct notification async_send_request_ex_resolve_failure_test[] = > { internet_close_handle, INTERNET_STATUS_HANDLE_CLOSING, 0, } > }; > > +static const struct notification async_send_request_ex_chunked_test[] = > +{ > + { internet_connect, INTERNET_STATUS_HANDLE_CREATED }, > + { http_open_request, INTERNET_STATUS_HANDLE_CREATED }, > + { http_send_request_ex, INTERNET_STATUS_DETECTING_PROXY, 1, 0, 1 }, > + { http_send_request_ex, INTERNET_STATUS_COOKIE_SENT, 1, 0, 1 }, > + { http_send_request_ex, INTERNET_STATUS_RESOLVING_NAME, 1, 0, 1 }, > + { http_send_request_ex, INTERNET_STATUS_NAME_RESOLVED, 1, 0, 1 }, > + { http_send_request_ex, INTERNET_STATUS_CONNECTING_TO_SERVER, 1 }, > + { http_send_request_ex, INTERNET_STATUS_CONNECTED_TO_SERVER, 1 }, > + { http_send_request_ex, INTERNET_STATUS_SENDING_REQUEST, 1 }, > + { http_send_request_ex, INTERNET_STATUS_REQUEST_SENT, 1 }, > + { http_send_request_ex, INTERNET_STATUS_REQUEST_COMPLETE, 1 }, > + { http_end_request, INTERNET_STATUS_RECEIVING_RESPONSE, 1 }, > + { http_end_request, INTERNET_STATUS_RESPONSE_RECEIVED, 1 }, > + { http_end_request, INTERNET_STATUS_REQUEST_COMPLETE, 1 }, > + { internet_close_handle, INTERNET_STATUS_CLOSING_CONNECTION }, > + { internet_close_handle, INTERNET_STATUS_CONNECTION_CLOSED }, > + { internet_close_handle, INTERNET_STATUS_HANDLE_CLOSING }, > + { internet_close_handle, INTERNET_STATUS_HANDLE_CLOSING } > +}; > + > static const struct notification_data notification_data[] = { > { > + async_send_request_ex_chunked_test, > + sizeof(async_send_request_ex_chunked_test)/sizeof(async_send_request_ex_chunked_test[0]), > + "GET", > + "test.winehq.org", > + "tests/data.php" > + }, > + { > async_send_request_ex_test, > sizeof(async_send_request_ex_test)/sizeof(async_send_request_ex_test[0]), > "POST", > @@ -5100,6 +5129,7 @@ START_TEST(http) > test_async_HttpSendRequestEx(&notification_data[0]); > test_async_HttpSendRequestEx(&notification_data[1]); > test_async_HttpSendRequestEx(&notification_data[2]); > + test_async_HttpSendRequestEx(&notification_data[3]); > InternetOpenRequest_test(); > test_http_cache(); > InternetOpenUrlA_test(); I was hoping for a test like the attached one, which shows that there is still more work to do. But it's a step in the right direction, so I'm fine with your patch. Thanks, Jacek

2 2

Re: (try 5) [2/4] imm32: do not let ImmDestroyContext destroy any default contexts
by Nikolay Sivov Sept. 25, 2013

Sept. 25, 2013

On 9/25/2013 16:03, Aric Stewart wrote: > defaultContext = ImmCreateContext(); > + if (defaultContext) > + ((InputContextData*)defaultContext)->threadDefault = TRUE; I think a nicer way is to let ImmCreateContext() set this attribute.

2 1

Re: (try 5)[1/4] imm32: Move thread data from TLSEntry to an internal list
by Nikolay Sivov Sept. 25, 2013

Sept. 25, 2013

On 9/25/2013 16:03, Aric Stewart wrote: > + if (!thread_data->defaultContext) > + { > + HIMC defaultContext; > + LeaveCriticalSection(&threaddata_cs); > + defaultContext = ImmCreateContext(); > + thread_data = IMM_GetThreadData(0); Why do you need to unlock/lock around ImmCreateContext()? Is it related to SendMessage() it calls?

2 1

Re: (try 4)[1/4] imm32: Move thread data from TLSEntry to an internal list
by Alexandre Julliard Sept. 25, 2013

Sept. 25, 2013

Aric Stewart <aric(a)codeweavers.com> writes: > --- > dlls/imm32/imm.c | 153 +++++++++++++++++++++++++++++++++++++++++-------------- > 1 file changed, 116 insertions(+), 37 deletions(-) It breaks the tests: ../../../tools/runtest -q -P wine -M urlmon.dll -T ../../.. -p urlmon_test.exe.so url.c && touch url.ok err:ntdll:RtlpWaitForCriticalSection section 0x55ec38e8 "imm.c: threaddata_cs" wait timed out in thread 005c, blocked by 0031, retrying (60 sec) err:ntdll:RtlpWaitForCriticalSection section 0x7bccc200 "loader.c: loader_section" wait timed out in thread 0060, blocked by 005c, retrying (60 sec) -- Alexandre Julliard julliard(a)winehq.org

1 0

Wine Gecko 2.24-beta1
by Jacek Caban Sept. 25, 2013

Sept. 25, 2013

Hi all, It's time to do another update of Wine Gecko, so I prepared a new beta build. It's already on Sourceforge [1]. Until server-side support for auto install of the package is committed [2], its manual installation is required [3]. I attached a patch to Wine that is required for the new Gecko. Anything that uses MSHTML is worth testing. All help with testing is appreciated! Cheers, Jacek [1] https://sourceforge.net/projects/wine/files/Wine%20Gecko/2.24-beta1/ [2] http://source.winehq.org/patches/data/97881 [3] http://wiki.winehq.org/Gecko

2 2

Re: crypt32: Support HCCE_LOCAL_MACHINE.
by Juan Lang Sept. 25, 2013

Sept. 25, 2013

Hi Ben, thanks for having a whack at this. Some tests would be nice. -static HCERTCHAINENGINE CRYPT_defaultChainEngine; +/* There are two default chain engines which correspond to HCCE_CURRENT_USER and + * HCCE_LOCAL_MACHINE. +*/ +static HCERTCHAINENGINE CRYPT_defaultChainEngine[2] = { NULL, NULL }; C automatically initializes statics to 0, so the initialization here is unnecessary. I'm also a little uncertain about the use of an array of two objects, I'm not sure that two distinct objects wouldn't be easier to follow, but I'm not religious on this point. + if (hChainEngine > HCCE_LOCAL_MACHINE + && InterlockedDecrement(&engine->ref) == 0) I think a function that returns whether an HCERTCHAINENGINE is one of the special ones would make this easier to read, e.g.: static int is_special_chain_engine(HCERTCHAINENGINE h) { return h == HCCE_CURRENT_USER || h == HCCE_LOCAL_MACHINE; } then: + if (is_special_chain_engine(hChainEngine) + && InterlockedDecrement(&engine->ref) == 0) +static HCERTCHAINENGINE CRYPT_GetDefaultChainEngine(HCERTCHAINENGINE h) { - if (!CRYPT_defaultChainEngine) + if (!CRYPT_defaultChainEngine[(int)h]) The constant casting is a little awkward. At least introduce a local pointer to the one you intend to modify, so we're not constantly having to re-read that cast. + if (hChainEngine <= HCCE_LOCAL_MACHINE) + engine = (CertificateChainEngine*)CRYPT_GetDefaultChainEngine( + hChainEngine); if (TRACE_ON(chain)) dump_chain_para(pChainPara); /* FIXME: what about HCCE_LOCAL_MACHINE? */ See my earlier suggestion on is_special_chain_engine. Also, what about that comment three lines down? Can't it be removed? -void default_chain_engine_free(void) +void default_chain_engine_free(HCERTCHAINENGINE h) { - CertFreeCertificateChainEngine(CRYPT_defaultChainEngine); + CertFreeCertificateChainEngine(CRYPT_defaultChainEngine[(int)h]); The function default_chain_engine_free is a thin wrapper around CryptFreeCertificateChainEngine; its intended use is to clear memory at shutdown. Rather than shift the responsibility of knowing which engines to free to the caller, just have it free the two engines itself. Thanks, --Juan

1 0

Re: [2/2] oleaut32: LoadRegTypeLib() should check actual typelib version
by Marvin Sept. 24, 2013

Sept. 24, 2013

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=2330 Your paranoid android. === build (build) === Patch failed to apply

1 0