Bug 187761

Summary: [GTK] http/tests/security/contentSecurityPolicy/1.1/plugintypes-affects-child.html is flaky
Product: WebKit Reporter: Alicia Boya García <aboya>
Component: WebKitGTKAssignee: Nobody <webkit-unassigned>
Status: RESOLVED INVALID    
Severity: Normal CC: bugs-noreply, cgarcia, mcatanzaro
Priority: P2    
Version: WebKit Nightly Build   
Hardware: Unspecified   
OS: Unspecified   

Description Alicia Boya García 2018-07-18 07:43:01 PDT 
The following test fails on GTK since r233910:

http/tests/security/contentSecurityPolicy/1.1/plugintypes-affects-child.html

--- /home/slave/webkitgtk/gtk-linux-64-release-tests/build/layout-test-results/http/tests/security/contentSecurityPolicy/1.1/plugintypes-affects-child-expected.txt
+++ /home/slave/webkitgtk/gtk-linux-64-release-tests/build/layout-test-results/http/tests/security/contentSecurityPolicy/1.1/plugintypes-affects-child-actual.txt
@@ -1,4 +1,3 @@
-CONSOLE MESSAGE: Refused to load http://127.0.0.1:8000/plugins/resources/mock-plugin.pl because its MIME type does not appear in the plugin-types directive of the Content Security Policy.
 This tests that a plugin-types directive on a top-level page should be inherited by a PluginDocument embedded in the page. This test passes if a console error says that the plugin was blocked.
Comment 1 Alicia Boya García 2018-07-18 08:43:43 PDT 
Regressed by:

[Nicosia] Add debug border, repaint counter state tracking to Nicosia::CompositionLayer
https://bugs.webkit.org/show_bug.cgi?id=187749
Comment 2 Alicia Boya García 2018-07-18 08:50:04 PDT 
Actually, no. That patch is completely unrelated; it just happens that it's the first failing commit, but that's because the bots discard all previous successful builds.
Comment 3 Michael Catanzaro 2018-07-18 09:08:36 PDT 
(In reply to Alicia Boya García from comment #0)
> The following test fails on GTK since r233910:

I see that's the only test failure in recent history. r233912 was successful. I think the test is flaky and should be marked accordingly.

I think you got tricked because r233910-r233911 was the most recent test run at the time you filed the bug. In these cases, it's important to look at the commit in question and guess if it is likely to blame, since it's often the case that the test is flaky and will pass again on the next run (as was the case here).
Comment 4 Alicia Boya García 2018-07-18 09:52:21 PDT 
(In reply to Michael Catanzaro from comment #3)
> (In reply to Alicia Boya García from comment #0)
> > The following test fails on GTK since r233910:
> 
> I see that's the only test failure in recent history. r233912 was
> successful. I think the test is flaky and should be marked accordingly.

Yeah... by the time I downloaded results there were 2 failures in a row, but now it's Pass-Failure-Failure.
Comment 5 Carlos Garcia Campos 2020-08-17 06:36:35 PDT 
Plugins are no longer supported.