February 2016 - Wine-Devel - list.winehq.org

Re: reg: Fail if the data value passed with REG_DWORD will cause an overflow
by Alexandre Julliard Feb. 20, 2016

Feb. 20, 2016

Hugh McMaster <hugh.mcmaster(a)outlook.com> writes: > @@ -243,8 +243,17 @@ static LPBYTE get_regdata(LPWSTR data, DWORD reg_type, WCHAR separator, DWORD *r > { > LPWSTR rest; > DWORD val; > + BOOL overflow = FALSE; > + if (data[1] == 'x') > + { > + WCHAR *ptr = &data[2]; > + while (*ptr == '0') > + ptr++; > + if (strlenW(ptr) > 8) > + overflow = TRUE; > + } > val = strtoulW(data, &rest, (data[1] == 'x') ? 16 : 10); You won't be able to catch base 10 overflows this way. It probably needs a more general solution (and some more test cases...) -- Alexandre Julliard julliard(a)winehq.org

2 1

Re: urlmon/tests: Use todo_wine_if() in tests
by Marvin Feb. 19, 2016

Feb. 19, 2016

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://testbot.winehq.org/JobDetails.pl?Key=20745 Your paranoid android. === w1064 (32 bit misc) === misc.c:756: Test failed: [52] wrong mime: L"image/x-icon"

1 0

Re: winmm/tests: Use todo_wine_if() in tests
by Marvin Feb. 19, 2016

Feb. 19, 2016

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://testbot.winehq.org/JobDetails.pl?Key=20744 Your paranoid android. === w1064 (32 bit midi) === midi.c:290: Test failed: midiOutOpen(dev=0) rc=MMSYSERR_ERROR midi.c:489: Test failed: midiStreamOpen(dev=-2) rc=MMSYSERR_ERROR midi.c:290: Test failed: midiOutOpen(dev=-1) rc=MMSYSERR_ERROR midi.c:489: Test failed: midiStreamOpen(dev=-2) rc=MMSYSERR_ERROR === w1064 (32 bit mci) === mci: unhandled exception c0000005 at 7698851A

1 0

Re: [PATCH] packet: create packet.dll and Implement GetAdapterNames(try 4)
by Sebastian Lackner Feb. 19, 2016

Feb. 19, 2016

Thanks for sending an updated version of your patch. I have some more comments below. Nevertheless, before you spent too much time to fix all the issues, I would suggest to discuss with André first, if packet.dll is acceptable for Wine. One reason why we might consider adding it is because the original version heavily reads information from the registry (SYSTEM\\CurrentControlSet), which is not populated by Wine, but still there might be alternative ways to solve it. If we really reimplement this DLL, we might also need tests in the long term, otherwise its difficult to say if everything is implemented correctly. On 18.02.2016 14:21, Jianqiu Zhang wrote: > > 0001-packet-create-packet.dll-and-Implement-GetAdapterNames.txt > > > From 06eb579dba41bc473429d262150677687f1f092a Mon Sep 17 00:00:00 2001 > From: Jianqiu Zhang <zhangjianqiu_133(a)yeah.net> > Date: Sun, 14 Feb 2016 17:27:37 +0800 > Subject: [PATCH] packet: create packet.dll and Implement GetAdapterNames > > Signed-off-by: Jianqiu Zhang <zhangjianqiu_133(a)yeah.net> > --- > configure.ac | 7 +++++++ > dlls/packet/Makefile.in | 5 +++++ > dlls/packet/packet.c | 56 +++++++++++++++++++++++++++++++++++++++++++++++++ > dlls/packet/packet.spec | 32 ++++++++++++++++++++++++++++ > 4 files changed, 100 insertions(+) > create mode 100644 dlls/packet/Makefile.in > create mode 100644 dlls/packet/packet.c > create mode 100644 dlls/packet/packet.spec > > diff --git a/configure.ac b/configure.ac > index c9445e7..504d46e 100644 > --- a/configure.ac > +++ b/configure.ac > @@ -69,6 +69,8 @@ AC_ARG_WITH(osmesa, AS_HELP_STRING([--without-osmesa],[do not use the OSMesa > AC_ARG_WITH(oss, AS_HELP_STRING([--without-oss],[do not use the OSS sound support])) > AC_ARG_WITH(pcap, AS_HELP_STRING([--without-pcap],[do not use the Packet Capture library]), > [if test "x$withval" = "xno"; then ac_cv_header_pcap_pcap_h=no; fi]) > +AC_ARG_WITH(packet, AS_HELP_STRING([--without-packet],[do not use the Packet DLL]), > + [if test "x$withval" = "xno"; then ac_cv_header_pcap_pcap_h=no; fi]) You can leave this part out, packet.dll does not require separate dependencies. > AC_ARG_WITH(png, AS_HELP_STRING([--without-png],[do not use PNG])) > AC_ARG_WITH(pthread, AS_HELP_STRING([--without-pthread],[do not use the pthread library]), > [if test "x$withval" = "xno"; then ac_cv_header_pthread_h=no; fi]) > @@ -1186,7 +1188,11 @@ then > fi > WINE_NOTICE_WITH(pcap,[test "x$ac_cv_lib_pcap_pcap_create" != xyes], > [pcap ${notice_platform}development files not found, wpcap won't be supported.]) > +WINE_NOTICE_WITH(packet,[test "x$ac_cv_lib_pcap_pcap_create" != xyes], > + [pcap ${notice_platform}development files not found, packet won't be supported.]) A separate WINE_NOTICE_WITH is also not necessary. Instead just modify the existing one, for example: """pcap ${notice_platform}development files not found, wpcap and packet won't be supported.""" > test "x$ac_cv_lib_pcap_pcap_create" != xyes && enable_wpcap=${enable_wpcap:-no} > +test "x$ac_cv_lib_pcap_pcap_create" != xyes && enable_packet=${enable_packet:-no} > + > > dnl **** Check for libxml2 **** > > @@ -3389,6 +3395,7 @@ WINE_CONFIG_DLL(wmvcore) > WINE_CONFIG_DLL(wnaspi32,,[implib]) > WINE_CONFIG_DLL(wow32,enable_win16,[implib]) > WINE_CONFIG_DLL(wpcap) > +WINE_CONFIG_DLL(packet) These lines are alphabetically sorted, so you would have to insert it at a different position. Please note that you can also use ./tools/make_makefiles to add this line for you, after you have added files for a new dll to git. > WINE_CONFIG_DLL(ws2_32,,[implib]) > WINE_CONFIG_TEST(dlls/ws2_32/tests) > WINE_CONFIG_DLL(wshom.ocx,,[clean]) > diff --git a/dlls/packet/Makefile.in b/dlls/packet/Makefile.in > new file mode 100644 > index 0000000..cecf07c > --- /dev/null > +++ b/dlls/packet/Makefile.in > @@ -0,0 +1,5 @@ > +MODULE = packet.dll > +IMPORTS = iphlpapi > + > +C_SRCS = \ > + packet.c > diff --git a/dlls/packet/packet.c b/dlls/packet/packet.c > new file mode 100644 > index 0000000..1074067 > --- /dev/null > +++ b/dlls/packet/packet.c > @@ -0,0 +1,56 @@ > +#include "config.h" > + > +#include "windows.h" > +#include "wine/debug.h" > + > +#include "iphlpapi.h" > + > +WINE_DEFAULT_DEBUG_CHANNEL(packet); > + > +BOOLEAN CDECL PacketGetAdapterNames(char *name_list, ULONG *size) > +{ > + IP_ADAPTER_INFO *adInfo, *adInfoList = NULL; > + DWORD errcode; > + ULONG needsize = 0, minDescLen = 0, minAdNameLen = 0, minTotLen = 0, nameOffset, descOffset; > + > + TRACE("name_list %p &size %p size %d\n", name_list, size, *size); > + > + errcode = GetAdaptersInfo(adInfoList, &needsize); > + adInfoList = HeapAlloc(GetProcessHeap(), 0, needsize); > + if(adInfoList == NULL) return FALSE; > + > + errcode = GetAdaptersInfo(adInfoList, &needsize); > + if(errcode) > + { > + HeapFree(GetProcessHeap(), 0, adInfoList); > + SetLastError(errcode); > + return FALSE; > + } > + for(adInfo = adInfoList; adInfo != NULL; adInfo = adInfo->Next) > + { > + minAdNameLen += lstrlenA(adInfo->AdapterName) + 1; > + minDescLen += lstrlenA(adInfo->Description) + 1; Usually a direct invocation of strlen() is preferred in the wine source. > + } > + minTotLen = minAdNameLen + minDescLen + 2; > + if(name_list == NULL || *size < minTotLen) > + { > + SetLastError(ERROR_INSUFFICIENT_BUFFER); > + HeapFree(GetProcessHeap(), 0, adInfoList); > + *size = minTotLen; When looking through the original implementation of packet.dll, I do not see any location which changes *size. > + return FALSE; > + } > + > + nameOffset = descOffset = 0; > + for(adInfo = adInfoList; adInfo != NULL; adInfo = adInfo->Next) > + { > + lstrcpyA(name_list + nameOffset, adInfo->AdapterName); > + lstrcpyA(name_list + minAdNameLen + descOffset + 1, adInfo->Description); > + nameOffset += lstrlenA(adInfo->AdapterName) + 1; > + descOffset += lstrlenA(adInfo->Description) + 1; strcpy / strlen would be preferred here. > + } > + name_list[minAdNameLen] = '\0'; > + name_list[minTotLen - 1] = '\0'; > + > + HeapFree(GetProcessHeap(), 0, adInfoList); > + return TRUE; > +} > diff --git a/dlls/packet/packet.spec b/dlls/packet/packet.spec > new file mode 100644 > index 0000000..32b9bc0 > --- /dev/null > +++ b/dlls/packet/packet.spec > @@ -0,0 +1,32 @@ > +@ stub PacketAllocatePacket > +@ stub PacketCloseAdapter > +@ stub PacketFreePacket > +@ cdecl PacketGetAdapterNames(ptr ptr) > +@ stub PacketGetAirPcapHandle > +@ stub PacketGetDriverVersion > +@ stub PacketGetNetInfoEx > +@ stub PacketGetNetType > +@ stub PacketGetReadEvent > +@ stub PacketGetStats > +@ stub PacketGetStatsEx > +@ stub PacketGetVersion > +@ stub PacketInitPacket > +@ stub PacketIsDumpEnded > +@ stub PacketLibraryVersion > +@ stub PacketOpenAdapter > +@ stub PacketReceivePacket > +@ stub PacketRequest > +@ stub PacketSendPacket > +@ stub PacketSendPackets > +@ stub PacketSetBpf > +@ stub PacketSetBuff > +@ stub PacketSetDumpLimits > +@ stub PacketSetDumpName > +@ stub PacketSetHwFilter > +@ stub PacketSetLoopbackBehavior > +@ stub PacketSetMinToCopy > +@ stub PacketSetMode > +@ stub PacketSetNumWrites > +@ stub PacketSetReadTimeout > +@ stub PacketSetSnapLen > +@ stub PacketStopDriver > -- 2.7.1 > > >

2 1

Re: [PATCH 1/2] msvcrt: Don't close HANDLE if stdout or stderr is sharing it
by Qian Hong Feb. 19, 2016

Feb. 19, 2016

Hi Piotr, Thanks for your work! I didn't have time to improve the patches so I didn't sent it upstream, I appreciate a lot for your work, now I can remove it from my todo list :)

1 0

Re: [PATCH 1/2] msvcp110: Add tr2_sys__Unlink implementation and test.(resend)
by Marvin Feb. 19, 2016

Feb. 19, 2016

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://testbot.winehq.org/JobDetails.pl?Key=20725 Your paranoid android. === build (build) === Patch failed to apply

2 1

Re: [PATCH 2/2] msvcp110: Add tr2_sys__Stat_wchar and tr2_sys__Lstat_wchar implementation and test.(resend)
by Marvin Feb. 19, 2016

Feb. 19, 2016

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://testbot.winehq.org/JobDetails.pl?Key=20726 Your paranoid android. === build (build) === Patch failed to apply

1 0

Re: msctf/tests: Use todo_wine_if() in tests
by Marvin Feb. 19, 2016

Feb. 19, 2016

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://testbot.winehq.org/JobDetails.pl?Key=20718 Your paranoid android. === w1064 (32 bit inputprocessor) === inputprocessor.c:787: Test failed: Deactivation came unexpectedly

1 0

[RFC PATCH 2/2] Load device drivers of the same group into the same winedevice process
by Aric Stewart Feb. 19, 2016

Feb. 19, 2016

Signed-off-by: Aric Stewart <aric(a)codeweavers.com> --- programs/services/services.c | 94 +++++++++++++++++++++++++++++-- programs/winedevice/device.c | 128 ++++++++++++++++++++++++++++++++++++++----- 2 files changed, 204 insertions(+), 18 deletions(-)

1 0

[RFC PATCH 1/2] programs/services: Break services into Loader groups
by Aric Stewart Feb. 19, 2016

Feb. 19, 2016

Sebastian was requesting that multiple drivers be able to be loaded into the same winedevice process. Discussing this with Alexandre he expressed that instead of my approach to the plug and play manager as a service it may be more proper to have it be a kernel driver that then depends on other drivers being loaded into the same process. So here is a proposed process to allow this to happen. Right now these 2 patches depend on my moving the driver loading code, but that is not a hard dependency at all, it would be relatively straight forward to be able to keep all the driver loading code into winedevice and still have these patches. This is a request for comment before submitting it to wine-patches. Also to decide if we should move the driver loading or not, as that does affect what set of patches are submitted. -aric Sort the groups by the service's tag to start in order --- programs/services/services.c | 101 ++++++++++++++++++++++++++++++++----------- 1 file changed, 75 insertions(+), 26 deletions(-)

1 0

Re: [PATCH] kernel32: Use setprogname(3) to set the program name if available.
by Alexandre Julliard Feb. 19, 2016

Feb. 19, 2016

Charles Davis <cdavis5x(a)gmail.com> writes: > @@ -1118,21 +1118,29 @@ static void set_process_name( int argc, char *argv[] ) > { > #ifdef HAVE_SETPROCTITLE > setproctitle("-%s", argv[1]); > -#endif > - > -#ifdef HAVE_PRCTL > +#elif defined(HAVE_PRCTL) || defined(HAVE_SETPROGNAME) > int i, offset; > char *p, *prctl_name = argv[1]; > char *end = argv[argc-1] + strlen(argv[argc-1]) + 1; > - > -#ifndef PR_SET_NAME > -# define PR_SET_NAME 15 > +#ifdef HAVE_SETPROGNAME > + static char progname[PATH_MAX]; > #endif > > if ((p = strrchr( prctl_name, '\\' ))) prctl_name = p + 1; > if ((p = strrchr( prctl_name, '/' ))) prctl_name = p + 1; > > +#ifdef HAVE_SETPROGNAME > + lstrcpynA( progname, prctl_name, sizeof(progname)/sizeof(progname[0]) ); > + progname[PATH_MAX-1] = 0; > + setprogname( progname ); > + if (1) > +#elif defined(HAVE_PRCTL) > +#ifndef PR_SET_NAME > +# define PR_SET_NAME 15 > +#endif > + > if (prctl( PR_SET_NAME, prctl_name ) != -1) > +#endif This is becoming hard to follow, please try to clean up these #ifdefs a little. Also you shouldn't need to make a copy, it can point directly into argv. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: advapi32: add stub for AddAuditAccessObjectAce
by Nikolay Sivov Feb. 19, 2016

Feb. 19, 2016

On 19.02.2016 11:40, Piet Jankbal wrote: > For bug https://bugs.winehq.org/show_bug.cgi?id=39918 > <https://bugs.winehq.org/show_bug.cgi?id=29862> > > > Signed-off-by: Louis Lenders <xerox.xerox2000x(a)gmail.com> > > > > This should be marked as a part of series, as they depend on each other. Also I think it's better to do a little bit more and forward them to ntdll Rtl* calls, like some existing calls already do, it's safe to assume prototype are the same between the two.

1 0

Re: advapi32: add stub for AddAuditAccessObjectAce
by Austin English Feb. 19, 2016

Feb. 19, 2016

On Fri, Feb 19, 2016 at 2:40 AM, Piet Jankbal <xerox.xerox2000x(a)gmail.com> wrote: > For bug https://bugs.winehq.org/show_bug.cgi?id=39918 > > > Signed-off-by: Louis Lenders <xerox.xerox2000x(a)gmail.com> The name on your email, the named on the signed off by, and the name on the patch are all different. Why? -- -Austin

1 0

Re: [4/4] ntdll: Set correct number of arguments for breakpoint exceptions.
by Marvin Feb. 19, 2016

Feb. 19, 2016

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://testbot.winehq.org/JobDetails.pl?Key=20712 Your paranoid android. === wvistau64 (32 bit exception) === exception.c:490: Test failed: 42: Unexpected eip 0x2e0000/0x2e0001 exception: unhandled exception 80000004 at 002E0000 exception.c:490: Test failed: 42: Unexpected eip 0x77baed97/0x2e0001 === w7u (32 bit exception) === exception.c:490: Test failed: 42: Unexpected eip 0x3f0000/0x3f0001 exception: unhandled exception 80000004 at 003F0000 exception.c:490: Test failed: 42: Unexpected eip 0x77cde34f/0x3f0001 === w1064 (32 bit exception) === Failure running script in VM: network read timed out (Connect:AgentVersion.h:0/9) === w1064 (64 bit exception) === Failure running script in VM: network read timed out (Connect:AgentVersion.h:0/9)

1 0

Re: [3/4] ntdll: Fix up instruction pointer in context for EXCEPTION_BREAKPOINT on x86_64.
by Marvin Feb. 19, 2016

Feb. 19, 2016

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://testbot.winehq.org/JobDetails.pl?Key=20711 Your paranoid android. === wvistau64 (32 bit exception) === exception.c:490: Test failed: 42: Unexpected eip 0x3c0000/0x3c0001 exception: unhandled exception 80000004 at 003C0000 exception.c:490: Test failed: 42: Unexpected eip 0x77baed97/0x3c0001 === w7u (32 bit exception) === exception.c:490: Test failed: 42: Unexpected eip 0x2f0000/0x2f0001 exception: unhandled exception 80000004 at 002F0000 exception.c:490: Test failed: 42: Unexpected eip 0x77cde34f/0x2f0001 === w1064 (32 bit exception) === Failure running script in VM: network read timed out (Connect:AgentVersion.h:0/9) === w1064 (64 bit exception) === Failure running script in VM: network read timed out (Connect:AgentVersion.h:0/9)

1 0

Re: [2/4] ntdll/tests: Add separate tests for exception handling of int $0x3.
by Marvin Feb. 19, 2016

Feb. 19, 2016

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://testbot.winehq.org/JobDetails.pl?Key=20710 Your paranoid android. === wvistau64 (32 bit exception) === exception.c:490: Test failed: 42: Unexpected eip 0x2c0000/0x2c0001 exception: unhandled exception 80000004 at 002C0000 exception.c:490: Test failed: 42: Unexpected eip 0x77baed97/0x2c0001 === w7u (32 bit exception) === exception.c:490: Test failed: 42: Unexpected eip 0x2f0000/0x2f0001 exception: unhandled exception 80000004 at 002F0000 exception.c:490: Test failed: 42: Unexpected eip 0x77cde34f/0x2f0001 === w1064 (32 bit exception) === Failure running script in VM: network read timed out (Connect:AgentVersion.h:0/9) === w1064 (64 bit exception) === Failure running script in VM: network read timed out (Connect:AgentVersion.h:0/9)

1 0

Re: [1/4] ntdll/tests: Add additional tests for instruction pointer in debug service tests.
by Marvin Feb. 19, 2016

Feb. 19, 2016

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://testbot.winehq.org/JobDetails.pl?Key=20709 Your paranoid android. === wvistau64 (32 bit exception) === exception.c:490: Test failed: 42: Unexpected eip 0x3c0000/0x3c0001 exception: unhandled exception 80000004 at 003C0000 exception.c:490: Test failed: 42: Unexpected eip 0x77baed97/0x3c0001 === w7u (32 bit exception) === exception.c:490: Test failed: 42: Unexpected eip 0x700000/0x700001 exception: unhandled exception 80000004 at 00700000 exception.c:490: Test failed: 42: Unexpected eip 0x77cde34f/0x700001 === w1064 (32 bit exception) === Failure running script in VM: network read timed out (Connect:AgentVersion.h:0/9) === w1064 (64 bit exception) === Failure running script in VM: network read timed out (Connect:AgentVersion.h:0/9)

1 0

Re: [PATCH] kernel32/tests: Additional test for GetStringTypeW() and Zl/Zp categories
by Marvin Feb. 18, 2016

Feb. 18, 2016

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://testbot.winehq.org/JobDetails.pl?Key=20683 Your paranoid android. === w1064 (32 bit locale) === locale.c:2365: Test failed: LCMapStringEx should fail with bad locale name locale.c:2366: Test failed: unexpected error code -559038737 locale.c:2457: Test failed: Expected lcid == 0, got got 00001000, error -559038737 locale.c:4175: Test failed: IsValidLocaleName should have failed locale.c:4177: Test failed: IsValidLocaleName should have failed locale.c:4542: Test failed: For id 7, expected ret == 4, got 1, error 87 locale.c:4544: Test failed: For id 7, Expected IVC, got ''

2 1

Re: [PATCH] secur32: Don't print 32-bit values as long integers. (Clang)
by Ken Thomases Feb. 18, 2016

Feb. 18, 2016

On Feb 18, 2016, at 2:20 PM, Charles Davis <cdavis5x(a)gmail.com> wrote: > > Signed-off-by: Charles Davis <cdavis5x(a)gmail.com> > --- > dlls/secur32/schannel_macosx.c | 16 ++++++++-------- > 1 file changed, 8 insertions(+), 8 deletions(-) In several other places throughout this file, the same problem was solved by casting to long in the logging call. Of course, that means that Clang doesn't warn about those. I don't mind your approach, but I think this should be handled consistently throughout the file, one way or the other. -Ken

2 1

Re: [PATCH 1/2 v2] winemac.drv: Actually make the GL context current even if no drawables were given.
by Ken Thomases Feb. 18, 2016

Feb. 18, 2016

On Feb 18, 2016, at 3:17 AM, Charles Davis <cdavis5x(a)gmail.com> wrote: > @@ -3108,7 +3108,7 @@ static BOOL macdrv_wglMakeContextCurrentARB(HDC draw_hdc, HDC read_hdc, struct w > (InterlockedCompareExchange(&context->update_swap_interval, FALSE, TRUE) || pbuffer != context->draw_pbuffer)) > set_swap_interval(context, 0); > } > - else > + else if (draw_hdc) > { > WARN("no window or pbuffer for DC\n"); > LeaveCriticalSection(&dc_pbuffers_section); I would prefer an explicit "if (!draw_hdc && !read_hdc) { /* clear context->draw_... */ } else" be introduced before the "if ((hwnd = WindowFromDC(draw_hdc)))". First, from my reading of the spec, a NULL draw_hdc and a non-NULL read_hdc should still be an error. Second, burying the logic in the pbuffer branch seems obscure. Thanks, Ken

1 0

Re: [PATCH] kernelbase: Add dll with forwards to kernel32.dll
by Nikolay Sivov Feb. 18, 2016

Feb. 18, 2016

On 18.02.2016 19:36, Detlef Riekenberg wrote: > +#@ stdcall CancelThreadpoolIo ntdll.TpCancelAsyncIoOperation > +@ stub CancelThreadpoolIo Hi. I think for new dll it's fine to have a stub, even if kernel32 does not have it. Why did you keep commented forward to ntdll too?

1 0

Re: [PATCH 2/2 v3] user32: Add more dialog creation message tests
by Dmitry Timoshkov Feb. 18, 2016

Feb. 18, 2016

Piotr Caban <piotr(a)codeweavers.com> wrote: > Signed-off-by: Piotr Caban <piotr(a)codeweavers.com> > --- > dlls/user32/tests/msg.c | 72 > +++++++++++++++++++++++++++++++++++++++++++++++++ > 1 file changed, 72 insertions(+) It's getting better, but again doesn't show a lot of details. Please add parameters for most of the recorded messages in the sequence, in particular EM_SETSEL and WM_GETDLGCODE are the mandatory ones. Also, in order to see the origin of EM_SETSEL I'd suggest to call DefDlgProc() manually and use defwndproc_counter around it together with the defwinproc flag. -- Dmitry.

3 10

Re: [PATCH 1/3] msvcp110: Add tr2_sys__Symlink implementation and test.(try 2)
by Piotr Caban Feb. 18, 2016

Feb. 18, 2016

On 02/18/16 10:54, YongHao Hu wrote: > + delete_flag = 1; > + for(i=0; i<sizeof(tests)/sizeof(tests[0]); i++) { > + errno = 0xdeadbeef; > + SetLastError(0xdeadbeef); > + ret = p_tr2_sys__Symlink(tests[i].existing_path, tests[i].new_path); > + if(ret==ERROR_PRIVILEGE_NOT_HELD || ret==ERROR_INVALID_FUNCTION) { > + delete_flag = 0; > + win_skip("Privilege not held or symbolic link not supported, skipping symbolic link tests.\n"); > + break; If second symlink creation fails we'll end up with some not deleted files while all tests are succeeding. I'm not sure if it's a real issue. I also think that it will be better if you remove the delete_flag variable. Something like this will be better in my opinion: ret = p_tr2_sys__Symlink(tests[i].existing_path, tests[i].new_path); if(!i && (ret==ERROR_PRIVILEGE_NOT_HELD || ret==ERROR_INVALID_FUNCTION)) { win_skip(...); ret = p_tr2_sys__Remove_dir("tr2_test_dir"); ok(ret == 1, "tr2_sys__Remove_dir(): expect 1 got %d\n", ret); return; } ok(ret == ERROR_SUCCESS, ...); > + if(tests[i].is_todo) > + todo_wine ok(ret == tests[i].last_error, "tr2_sys__Symlink(): test %d expect: %d, got %d\n", i+1, tests[i].last_error, ret); > + else > + ok(ret == tests[i].last_error, "tr2_sys__Symlink(): test %d expect: %d, got %d\n", i+1, tests[i].last_error, ret); Please use todo_wine_if here. Thanks, Piotr

2 2

Re: mshtml/tests: Trace the buffer when it is unexpectedly changed in mshtml:protocol.
by Jacek Caban Feb. 18, 2016

Feb. 18, 2016

On 02/18/16 03:14, Francois Gouget wrote: > Signed-off-by: Francois Gouget <fgouget(a)free.fr> > --- > dlls/mshtml/tests/protocol.c | 20 ++++++++++---------- > 1 file changed, 10 insertions(+), 10 deletions(-) > > diff --git a/dlls/mshtml/tests/protocol.c b/dlls/mshtml/tests/protocol.c > index f9803df..9a511ba 100644 > --- a/dlls/mshtml/tests/protocol.c > +++ b/dlls/mshtml/tests/protocol.c > @@ -397,7 +397,7 @@ static void test_res_protocol(void) > hres = IInternetProtocolInfo_ParseUrl(protocol_info, blank_url, PARSE_DOMAIN, 0, buf, > sizeof(buf)/sizeof(buf[0]), &size, 0); > ok(hres == S_OK || hres == E_FAIL, "ParseUrl failed: %08x\n", hres); > - ok(buf[0] == '?', "buf changed\n"); > + ok(buf[0] == '?', "buf changed buf[0]='%c'\n", buf[0]); > Why don't you use wine_dbgstr_w here? Thanks, Jacek

1 0

Re: ntdll/tests: Add more tests for calling int $0x2d on x86 and x86_64. (v2)
by Marvin Feb. 18, 2016

Feb. 18, 2016

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://testbot.winehq.org/JobDetails.pl?Key=20657 Your paranoid android. === wvistau64 (32 bit exception) === exception.c:490: Test failed: 42: Unexpected eip 0x2c0000/0x2c0001 exception: unhandled exception 80000004 at 002C0000 exception.c:490: Test failed: 42: Unexpected eip 0x77baed97/0x2c0001 === w7u (32 bit exception) === exception.c:490: Test failed: 42: Unexpected eip 0x2f0000/0x2f0001 exception: unhandled exception 80000004 at 002F0000 exception.c:490: Test failed: 42: Unexpected eip 0x77cde34f/0x2f0001 === w1064 (32 bit exception) === Failure running script in VM: network read timed out (Connect:AgentVersion.h:0/9) === w1064 (64 bit exception) === Failure running script in VM: network read timed out (Connect:AgentVersion.h:0/9)

1 0