imported/w3c/web-platform-tests/html/semantics/document-metadata/the-base-element/base_href_specified.sub.html causes the next test to be flaky

• Component: DOM
• Status: NEW ---

[Chris Dumez]

imported/w3c/web-platform-tests/html/semantics/document-metadata/the-base-element/base_href_specified.sub.html causes the next test to be flaky.

This is possibly WebKit1-specific.

diff:
--- /Volumes/Data/slave/yosemite-release-tests-wk1/build/layout-test-results/imported/w3c/web-platform-tests/html/semantics/document-metadata/the-base-element/base_href_unspecified-expected.txt
+++ /Volumes/Data/slave/yosemite-release-tests-wk1/build/layout-test-results/imported/w3c/web-platform-tests/html/semantics/document-metadata/the-base-element/base_href_unspecified-actual.txt
@@ -1,3 +1,4 @@
+Blocked access to external URL http://www.localhost:8800/test.ico
 
 PASS The value of the href attribute must be the document's address if it is unspecified 
 PASS The src attribute of the img element must relative to document's address 


Basically, the load error from base_href_specified.sub.html may be displayed on the output of the following test (base_href_unspecified.html).

[Chris Dumez]

Skipped in <http://trac.webkit.org/changeset/189479>.

[Chris Dumez]

Since the load is an image load and it seems to outlive its document loader. This may be related to ping loads. We may need to tweak this test to make sure it waits for the load to finish before completing.

[Alexey Proskuryakov]

We may have other tests skipped because of favicons, I vaguely remember this being a problem before.

[Chris Dumez]

In this case though, I suspect this is not favicon related as the test loads test.ico via an IMG element:
<base id="base" href="http://{{domains[www]}}:{{ports[http][0]}}">
<img id="test" src="test.ico" style="display:none">