Make it possible to debug the inspector code using another instance of the inspector.
Created attachment 216430 [details]
Attachment 216430 [details] did not pass style-queue:
Failed to run "['Tools/Scripts/check-webkit-style', '--diff-files', u'LayoutTests/ChangeLog', u'LayoutTests/inspector-protocol/debugger/nested-inspectors-expected.txt', u'LayoutTests/inspector-protocol/debugger/nested-inspectors.html', u'Source/WebKit2/ChangeLog', u'Source/WebKit2/UIProcess/WebInspectorProxy.cpp', u'Source/WebKit2/UIProcess/WebInspectorProxy.h']" exit_code: 1
Source/WebKit2/UIProcess/WebInspectorProxy.cpp:71: Weird number of spaces at line-start. Are you using a 4-space indent? [whitespace/indent] 
Source/WebKit2/UIProcess/WebInspectorProxy.cpp:87: Weird number of spaces at line-start. Are you using a 4-space indent? [whitespace/indent] 
Total errors found: 2 in 6 files
If any of these errors are false positives, please file a bug against check-webkit-style.
Created attachment 216433 [details]
Comment on attachment 216433 [details]
Attachment 216433 [details] did not pass mac-wk2-ews (mac-wk2):
New failing tests:
Created attachment 216435 [details]
Archive of layout-test-results from webkit-ews-09 for mac-mountainlion-wk2
The attached test failures were seen while running run-webkit-tests on the mac-wk2-ews.
Bot: webkit-ews-09 Port: mac-mountainlion-wk2 Platform: Mac OS X 10.8.5
(In reply to comment #5)
> (From update of attachment 216433 [details])
> Attachment 216433 [details] did not pass mac-wk2-ews (mac-wk2):
> Output: http://webkit-queues.appspot.com/results/22728545
> New failing tests:
This should be unrelated to my C++ change as the test runner will not touch my code. However, the new test has an impact on the others, so I will have to debug that.
(In reply to comment #7)
> (In reply to comment #5)
> > (From update of attachment 216433 [details] [details])
> > Attachment 216433 [details] [details] did not pass mac-wk2-ews (mac-wk2):
> > Output: http://webkit-queues.appspot.com/results/22728545
> > New failing tests:
> > inspector-protocol/debugger/setBreakpoint.html
> This should be unrelated to my C++ change as the test runner will not touch my code. However, the new test has an impact on the others, so I will have to debug that.
Seems like I've uncovered a bug.
The test in this patch is changing the m_pauseOnExceptionsState on the ScriptDebugServer. The state is not reset before the following tests are run.
1. The failing test sets a breakpoint with a condition.
2. The JS in the conditon throws an exception.
3. m_pauseOnExceptionsState that was not reset, so the ScriptDebugServer will notify the inspector about the exception. This will run an event loop and wait.
4. This is unexpected in the test, so it will try to fail early by detaching the inspector window.
5. ScriptDebugServer.m_currentCallFrame is set to 0 and the event loop unrolls.
6. ScriptDebugServer.hasBreakpoint tries to report the exception to the console using ScriptDebugServer.m_currentCallFrame which is 0. <Crash>
I will revert the m_pauseOnExceptionsState in the new test and also add/fix the other bug. I guess that one could reproduce in the inspector by adding a condition that throws an exception. Wait until the debugger stops on the condition. Close the inspector.
Created attachment 216454 [details]
Comment on attachment 216454 [details]
Clearing flags on attachment: 216454
Committed r158976: <http://trac.webkit.org/changeset/158976>
All reviewed patches have been landed. Closing bug.
This broke attaching the inspector, see https://bugs.webkit.org/show_bug.cgi?id=124148