May 2021 - Wine-bugs - list.winehq.org

[Bug 48654] New: Automate checking the WTBS patches
by WineHQ Bugzilla 03 May '21

03 May '21

https://bugs.winehq.org/show_bug.cgi?id=48654 Bug ID: 48654 Summary: Automate checking the WTBS patches Product: Wine-Testbot Version: unspecified Hardware: x86 OS: Linux Status: NEW Severity: normal Priority: P2 Component: unknown Assignee: wine-bugs(a)winehq.org Reporter: fgouget(a)codeweavers.com Distribution: --- See bug 48653 for a description of the WTBS and how to check the jobs created when running it on the TestBot. Some patches in the WTBS are not meant to create TestBot jobs but rather to verify that it ignores non-Wine patches, or patches that don't impact the tests or the Wine build, etc. Those still create entries in the 'wine-devel' page, that is in the TestBot's Patches table. Each entry gets the patch subject and a disposition property indicating what the TestBot did with the patch. Unfortunately non-Wine patches are not kept around so there is no way to get at the WTBS fields for those. So a TestWTBS.patches script would need to get the WTBS properties from another source. For instance it could be given the path to the WTBS mbox and get the properties from there. It would not need to understand all the intricacies of the mbox format (though with the appropriate Perl module it may be easy to parse). As a last resort, all we would care about is: * The 'From' lines to delimit emails. * The 'Subject:' lines to match the mbox emails with Patches entries. * The WTBS properties in the commit message following the Subject line. Checking the patch objects would require an extra property: * WTBS.Patch.Disposition This should be set to the expected patch disposition such as 'No patch found', 'Not a Wine patch', etc. A default disposition could be set if a 'WTBS.Job.*' or other task property is specified. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.

1 2

[Bug 48655] New: Automate checking the WTBS emails
by WineHQ Bugzilla 03 May '21

03 May '21

https://bugs.winehq.org/show_bug.cgi?id=48655 Bug ID: 48655 Summary: Automate checking the WTBS emails Product: Wine-Testbot Version: unspecified Hardware: x86 OS: Linux Status: NEW Severity: normal Priority: P2 Component: unknown Assignee: wine-bugs(a)winehq.org Reporter: fgouget(a)codeweavers.com Distribution: --- See bug 48653 for a description of the WTBS and how to check the jobs created when running it on the TestBot. When a 'wine-devel' job completes the TestBot sends one or two emails. * One of the emails simply sends the task reports to the patch author. * The other is only sent if a task generated new errors. The point of the TestWTBS.emails script would be to check that: 1. The TestBot sent the task report email for every Wine patch and not for any patch email. This would be based off of the WTBS.Patch.Disposition field. 2. The TestBot sent a 'found new errors' email only when new errors are expected. This would be based off of the 'WTBS.*.NewFailures' property or the expected errors list. However for both these checks the script would need access to the emails sent by the TestBot. The simplest option would be for the tester to set the WinePatchToOverride property and save the emails to a separate mbox and give the patch to that mbox to the script. As in the TestWTBS.patches case, the script would only care about a few lines: * The 'From' lines to delimit the emails. * The 'Subject' lines to match them to TestBot jobs and thus to the set of WTBS properties. Note that this would not rely on the 'TestBot job xxx results:' part since the new failures email are lacking it. Instead it would match the rest of the subject to the job's Remarks field. * The 'I think I found new failures' line to identify emails reporting new errors. The script could later be expanded to check the content of the email or attachments more in details if needed. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.

1 2

[Bug 32216] New: WienTestBot should skip some tests on some platforms
by wine-bugs＠winehq.org 03 May '21

03 May '21

http://bugs.winehq.org/show_bug.cgi?id=32216 Bug #: 32216 Summary: WienTestBot should skip some tests on some platforms Product: Wine-Testbot Version: unspecified Platform: x86 OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: unknown AssignedTo: wine-bugs(a)winehq.org ReportedBy: fgouget(a)codeweavers.com Classification: Unclassified An example is the ieframe:webbrowser tests which it should skip on NT4 and Windows 2000 platforms. WineTest.exe is said to already have the necessary code somewhere, it's just missing from TestLauncher. See this wine-devel thread for more details: http://www.winehq.org/pipermail/wine-devel/2012-November/097819.html -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.

2 11

[Bug 50719] New: The TestBot mishandles renames
by WineHQ Bugzilla 03 May '21

03 May '21

https://bugs.winehq.org/show_bug.cgi?id=50719 Bug ID: 50719 Summary: The TestBot mishandles renames Product: Wine-Testbot Version: unspecified Hardware: x86-64 OS: Linux Status: NEW Severity: normal Priority: P2 Component: unknown Assignee: wine-bugs(a)winehq.org Reporter: fgouget(a)codeweavers.com Distribution: --- The TestBot analyses patches to determine whether they are Wine patches and which tests are impacted and must be rerun. But when a patch renames a test unit it still schedules a run for the renamed test unit. For instance if apphelp:apphelp is renamed to apphelp:renamed, the TestBot still schedules a run of apphelp:apphelp (in addition to the apphelp:renamed one). This means it does not keep track of Wine's current files list correctly, which also impacts identifying whether a patch is a Wine patch or not, particularly for patch series. This can be tested with the following wtbsuite tests: WTBS Check combined renaming and modifying a file (apphelp:apphelp). [2/2] WTBS S10.2 - ...before renaming it in the next part (amstream:amstream). [7/7] WTBS S5.7 - Check for rename+addition+removal conflicts (msi:suminfo). -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.

1 2

[Bug 48208] New: Detect missing entry point errors
by WineHQ Bugzilla 03 May '21

03 May '21

https://bugs.winehq.org/show_bug.cgi?id=48208 Bug ID: 48208 Summary: Detect missing entry point errors Product: Wine-Testbot Version: unspecified Hardware: x86 OS: Linux Status: NEW Severity: normal Priority: P2 Component: unknown Assignee: wine-bugs(a)winehq.org Reporter: fgouget(a)codeweavers.com Distribution: --- When a Windows test executable cannot run because of a missing test executable the TestBot remains stuck until the 2 minutes timeout. This is because Windows puts up an error dialog waiting for user input. In such a case the test report is almost empty and the issue is not obvious until one looks at the screenshot. The TestBot should detect such a situation and warn the user. One simple approach would be for the TestBot to detect that the test report only contains the TestLauncher's start and done lines and indicates a timeout. In such a case it could add a message to the .err file suggesting to check the screenshot. This would only be a heuristic though. A better approach would be for the TestLauncher to detect the error dialog. Ideally, instead of waiting for the child process for the full 2 minutes it would periodically check for the presence of this dialog, which would also speed up the task run. This approach could also be used by WineTest.exe, though presumably missing entry points would have been detected before making it into the Wine source. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.

1 8

[Bug 50538] New: Wine child process exceptions are miscounted
by WineHQ Bugzilla 03 May '21

03 May '21

https://bugs.winehq.org/show_bug.cgi?id=50538 Bug ID: 50538 Summary: Wine child process exceptions are miscounted Product: Wine-Testbot Version: unspecified Hardware: x86-64 OS: Linux Status: NEW Severity: normal Priority: P2 Component: unknown Assignee: wine-bugs(a)winehq.org Reporter: fgouget(a)codeweavers.com Distribution: --- The "WTBS Crash in a child process (kernel32:toolhelp)." test case shows an inconsistency in the test failures count between the Windows and Wine VMs. On Windows we get: 1bec:toolhelp: unhandled exception c0000005 at 0054273D toolhelp:471: unhandled exception c0000005 in child process 1bec 1750:toolhelp: 3 tests executed (0 marked as todo, 1 failure), 0 skipped. Here there is 1 exception, 1 failure reported by the parent process but 2 error lines, and the task's failure count ends up being 2. But in Wine we get: Unhandled exception: page fault on write access to 0x00000000 in 32-bit code (0x00542c1d). toolhelp:471: unhandled exception c0000005 in child process 0104 00fc:toolhelp: 3 tests executed (0 marked as todo, 1 failure), 0 skipped. So there is still 1 exception, still 1 failure reported by the parent process, still 2 error lines, but the task's failure count ends up being just 1. The issue appears to be that ParseWineTestReport() (and probably dissect too) detects ": unhandled exception...at" lines but not Wine's "Unhandled exception:" ones. This means the TestBot may fail to detect exceptions on Wine. Meanwhile GetReportLineCategory() and _GetLineKey() detect both. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.

1 3

[Bug 50491] New: Detecting whether garbled lines are new
by WineHQ Bugzilla 03 May '21

03 May '21

https://bugs.winehq.org/show_bug.cgi?id=50491 Bug ID: 50491 Summary: Detecting whether garbled lines are new Product: Wine-Testbot Version: unspecified Hardware: x86-64 OS: Linux Status: NEW Severity: normal Priority: P2 Component: unknown Assignee: wine-bugs(a)winehq.org Reporter: fgouget(a)codeweavers.com Distribution: --- Sometimes a failure line is garbled because of race conditions or a missing '\n'. This typically looks something like this: advpack.c:689: WTBS Garble 2advpack.c:691: WTBS Garble 3advpack.c:692: Test failed: WTBS A test failure The TestBot still detects these failures because it knows which test unit is running at that point in the report and thus looks for something like 'advpack.c:\d+: Test failed: ...'. But when it tries to detect whether this is a new error line it cannot compare the whole line with past instances as the garbage part may change or not be present in WineTest runs (particularly if it contains line numbers). That suggests it should extract just the current failure string, that is: advpack.c:692: Test failed: WTBS A test failure This should maybe also be the string stored in the .errors file. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.

1 3

[Bug 27453] New: Lag when loading new sound in source games
by wine-bugs＠winehq.org 03 May '21

03 May '21

http://bugs.winehq.org/show_bug.cgi?id=27453 Summary: Lag when loading new sound in source games Product: Wine Version: 1.3.21 Platform: x86 OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: -unknown AssignedTo: wine-bugs(a)winehq.org ReportedBy: congelli501(a)gmail.com When you start a level in a game powered by valve's source engine, sounds that are played for the first time since the game was launched will cause the game to lag. Ex: open a new portal for the first time in a Portal2 level. Affects: - Portal 2 - Left 4 Dead 2 - Probably more... -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.

2 39

[Bug 7991] Test Drive: Unlimited messed up graphics
by WineHQ Bugzilla 03 May '21

03 May '21

https://bugs.winehq.org/show_bug.cgi?id=7991 Linards <linards.liepins(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |linards.liepins(a)gmail.com --- Comment #36 from Linards <linards.liepins(a)gmail.com> --- Is this bug still present on wine-staging 6.7 ? -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.

1 0

[Bug 51094] New: Battlefield 3 crashes before entering the main menu with Vulkan renderer
by WineHQ Bugzilla 03 May '21

03 May '21

https://bugs.winehq.org/show_bug.cgi?id=51094 Bug ID: 51094 Summary: Battlefield 3 crashes before entering the main menu with Vulkan renderer Product: Wine Version: 6.7 Hardware: x86-64 OS: Linux Status: NEW Severity: normal Priority: P2 Component: -unknown Assignee: wine-bugs(a)winehq.org Reporter: andrey.goosev(a)gmail.com Distribution: --- Created attachment 69954 --> https://bugs.winehq.org/attachment.cgi?id=69954 backtrace 04fc:fixme:d3d:wined3d_device_context_clear_rendertarget_view Not implemented for WINED3D_RTYPE_BUFFER resources. 04fc:err:d3d11:d3d11_immediate_context_ClearRenderTargetView Failed to clear view, hr 0x8876086c. 0518:fixme:d3d:wined3d_buffer_vk_prepare_location Unhandled location 0. 0518:err:d3d:wined3d_buffer_invalidate_range Buffer 0x9fcf0c8 does not have any up to date location. 0518:err:d3d:wined3d_texture_vk_prepare_texture Invalid resource type WINED3D_RTYPE_BUFFER. 0518:err:d3d:wined3d_context_vk_create_image Failed to create image, vr unrecognised(-1000011001). 0518:err:d3d:vk_image_view_type_from_wined3d Unhandled resource type WINED3D_RTYPE_BUFFER. 6.7-64-g4ccf7498aac -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.

1 0