Summary: | [Qt][WK2] Newly added editing/selection/caret-after-keypress.html crashing | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Zoltan Arvai <zarvai> | ||||||||
Component: | Tools / Tests | Assignee: | Eugene Klyuchnikov <eustas> | ||||||||
Status: | RESOLVED INVALID | ||||||||||
Severity: | Normal | CC: | allan.jensen, eustas, jturcotte, kadam, rniwa | ||||||||
Priority: | P2 | Keywords: | Qt | ||||||||
Version: | 528+ (Nightly build) | ||||||||||
Hardware: | Unspecified | ||||||||||
OS: | Unspecified | ||||||||||
Bug Depends on: | 110742 | ||||||||||
Bug Blocks: | 42194, 79668 | ||||||||||
Attachments: |
|
Description
Zoltan Arvai
2013-05-31 06:12:54 PDT
posting messages in WK2 test causes crash. I'll add TextExpectation. It have been landed in http://trac.webkit.org/changeset/151014. :) Created attachment 203440 [details]
Patch
Created attachment 203442 [details]
Patch
Well, I've moved expectation, because there is a meta-bug that is responsible for the whole a group of tests that fail on WK2. (In reply to comment #5) > Well, I've moved expectation, because there is a meta-bug that is responsible > for the whole a group of tests that fail on WK2. Yes, that sounds reasonable. Probably the change can be landed without review like most of the unreviewed gardening patches. Committed r151015: <http://trac.webkit.org/changeset/151015> But why is this causing a crash? In new test we use "eventSender.keyDown". As it is said in https://bugs.webkit.org/show_bug.cgi?id=42194 : "WebKitTestRunner needs an implementation of eventSender. Many tests fail because of this (hundreds)." What I see in stack trace: in WebPageProxy::didReceiveEvent in case of synchronous (test) keyboard event handling we face that type of first message in keyboard event queue is not the same as type of event we just appended to this queue. Assertion that type of last message in queue is the same as first message in queue means that we expect that queue is never longer that 1 item (at least for synchronous case). Stack trace has depth of 206 frames; there are 6 nested key events. 5 events are at "WebPageProxy.cpp:1395", invoked from handleKeyPressEvent. 6-th is at "WebPageProxy.cpp:1396", invoked from handleKeyReleaseEvent. So, when test reaches exit condition and begins unrolling, the problem occurs. I'm not sure, is it OK or NOK to nest eventSender.keyDown invocations. I'll prepare asynchronous test to see if will make Qt5-WK bot happy. (In reply to comment #8) > But why is this causing a crash? Created attachment 203650 [details]
Patch
=== Bulk closing of Qt bugs === If you believe that this bug report is still relevant for a non-Qt port of webkit.org, please re-open it and remove [Qt] from the summary. If you believe that this is still an important QtWebKit bug, please fill a new report at https://bugreports.qt-project.org and add a link to this issue. See http://qt-project.org/wiki/ReportingBugsInQt for additional guidelines. |