January 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 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 10219] New: Windows Media Player 11 setup fails
by wine-bugs＠winehq.org 03 May '21

03 May '21

http://bugs.winehq.org/show_bug.cgi?id=10219 Summary: Windows Media Player 11 setup fails Product: Wine Version: 0.9.48. Platform: PC URL: http://www.microsoft.com/windows/windowsmedia/download/A llDownloads.aspx?displang=en&qstechnology= OS/Version: Linux Status: NEW Keywords: download, Installer Severity: normal Priority: P2 Component: wine-misc AssignedTo: wine-bugs(a)winehq.org ReportedBy: alex(a)thehandofagony.com The setup program for Windows Media Player 11 tries to validate Windows if the version is set to XP, and this fails. There was talk some time ago about an ntlm_auth patch that would allow the validation, but it fails for me even though I have the correct version of Samba installed. However, the installer does not run the Genuine check if the Windows version is set to Vista. When run thus it does work, however; it tries to create an instance of a wuapi.dll class, which is not registered in Wine. When I use a native dll from XP and register it the console output changes saying no instance could be created. The installer then says no updates could be found. -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.

2 19

[Bug 14547] New: Problem with loading of windows from Java applications
by wine-bugs＠winehq.org 03 May '21

03 May '21

http://bugs.winehq.org/show_bug.cgi?id=14547 Summary: Problem with loading of windows from Java applications Product: Wine Version: 1.1.1 Platform: Other OS/Version: other Status: UNCONFIRMED Severity: normal Priority: P2 Component: -unknown AssignedTo: wine-bugs(a)winehq.org ReportedBy: yolande(a)haneder.biz Created an attachment (id=14894) --> (http://bugs.winehq.org/attachment.cgi?id=14894) Files for the test case We made a test case to show that child windows for Java applications (written in a class are not appearing) are problematic. To start the test, you have to install Java on Wine (I tested with java1.5-10) and start file in the /bin directory of the java installation directory. What is expected (showing during tests on Windows XP): There is a mainframe called "Using a JDesktopPane" Then a row in the main body of the main frame called "Add" Then a child window called "Internal frame" Result in Wine: Totally black screen with only the part with the word "Add" to be visible. Through hovering the mouse over the black part, you can restore most parts of your desktop to access other windows. If you need the source code of any of the file, you just need to ask. -- 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 19

[Bug 21764] New: Sun JRE (jre-6u16-windows-i586-s) installation failure
by wine-bugs＠winehq.org 03 May '21

03 May '21

http://bugs.winehq.org/show_bug.cgi?id=21764 Summary: Sun JRE (jre-6u16-windows-i586-s) installation failure Product: Wine Version: unspecified Platform: x86 URL: http://java.sun.com/products/archive/j2se/6u16/index.h tml OS/Version: Linux Status: NEW Keywords: download, regression Severity: normal Priority: P2 Component: msi AssignedTo: wine-bugs(a)winehq.org ReportedBy: bunglehead(a)gmail.com Sun JRE 6update16 fails to install with 1.1.39. It's a regression and a commit to blame is 1ff992314887d03abeb4098789701ff3bfd5d2d8: msi: Add summary information stream to the streams table. Reverting helps. -- 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 15