[Webkit-unassigned] [Bug 153683] Layout Test fast/parser/external-entities-in-xslt.xml is flaky on El Capitan (but fails most of the time)

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=153683

--- Comment #11 from mitz at webkit.org <mitz at webkit.org> ---
With an archived Debug build of r195795:

$ WebKitTestRunner LayoutTests/fast/parser/external-entities-in-xslt.xml 2> /dev/null
Content-Type: text/plain
CONSOLE MESSAGE: line 1: Document is empty

CONSOLE MESSAGE: line 1: Start tag expected, '<' not found

SUCCESS
#EOF
#EOF

$ DumpRenderTree LayoutTests/fast/parser/external-entities-in-xslt.xml 2> /dev/null
CONSOLE MESSAGE: line 1: Document is empty

CONSOLE MESSAGE: line 1: Start tag expected, '<' not found

Content-Type: text/plain
DumpMalloc: 55205888
SUCCESS
#EOF

--------------------------------------------------------------------------------
With an archived Debug build of r195794:

$ WebKitTestRunner LayoutTests/fast/parser/external-entities-in-xslt.xml 2> /dev/null
Content-Type: text/plain
CONSOLE MESSAGE: line 1: Document is empty

CONSOLE MESSAGE: line 1: Start tag expected, '<' not found

SUCCESS
#EOF
#EOF


$ DumpRenderTree LayoutTests/fast/parser/external-entities-in-xslt.xml 2> /dev/null
CONSOLE MESSAGE: line 1: Document is empty

CONSOLE MESSAGE: line 1: Start tag expected, '<' not found

Content-Type: text/plain
DumpMalloc: 54984704
SUCCESS
#EOF

--------------------------------------------------------------------------------

I don’t know why it’s OK for the ordering of the Content-Type: header with respect to the other output to differ between the tools, and I suspect that the ordering seen in WebKitTestRunner is related to the reported failure, but I don’t see evidence that that has changed recently.

It’s also worth noting that with both test tools, if I run the test a second time in the same instance of the tool, then the CONSOLE MESSAGE lines do not repeat. Consequently, if I use run-webkit-tests to run the test twice, it passes the second time.

Still, even though the WebKitTestRunner output looks the same (and wrong) to me before and after r195795, it looks different to run-webkit-tests, or run-webkit-tests gets it differently.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.webkit.org/pipermail/webkit-unassigned/attachments/20160202/1deee355/attachment.html>