1087165 - bug 1084367 disabled many MSE tests

Status: RESOLVED FIXED

(Testing :: web-platform-tests, defect, P5)

Description

[Karl Tomlinson (:karlt)]

+++ This bug was initially created as a clone of Bug #1084367 +++

I guess meta files were accidentally updated with old status after a pull/fetch.

% hg export c27a4a16bf74 | grep disabled:
+  disabled: 1086275
     disabled: true
+    disabled:
+  disabled: https://bugzilla.mozilla.org/show_bug.cgi?id=1066467
+  disabled: https://bugzilla.mozilla.org/show_bug.cgi?id=1066467
+  disabled: https://bugzilla.mozilla.org/show_bug.cgi?id=1066467
+  disabled: https://bugzilla.mozilla.org/show_bug.cgi?id=1066467
+  disabled: https://bugzilla.mozilla.org/show_bug.cgi?id=1066467
-    disabled: TIMEOUT or FAIL https://bugzilla.mozilla.org/show_bug.cgi?id=1066467
+  disabled: https://bugzilla.mozilla.org/show_bug.cgi?id=1066467
+  disabled: https://bugzilla.mozilla.org/show_bug.cgi?id=1066467
+  disabled: https://bugzilla.mozilla.org/show_bug.cgi?id=1066467
-    disabled: https://bugzilla.mozilla.org/show_bug.cgi?id=1066467
+  disabled: https://bugzilla.mozilla.org/show_bug.cgi?id=1066467
+  disabled: https://bugzilla.mozilla.org/show_bug.cgi?id=1066467
+  disabled: https://bugzilla.mozilla.org/show_bug.cgi?id=1066467

Comment 1

[James Graham [:jgraham]]

This wasn't intentional; it must have been a merge error on my part.

Comment 2

[James Graham [:jgraham]]

Although I'm confused; bug 1066467 makes it look like you intentionally disabled these tests?

Comment 3

[James Graham [:jgraham]]

Yeah, so this was a merge error. For tedious reasons I have to run the updates on cedar rather than try, and it got behind m-c. I'm waiting for some try results to reenable the tests.

Comment 4

[James Graham [:jgraham]]

Attachment: Reenable MSE tests accidentially disabled in last update.

I think this is the right set of tests. Expected results set based on a Try run, but please verify that these match what you expect. I wonder if the ones which are different between Linux and Linux 64 are actually unstable?

Comment 5

[Karl Tomlinson (:karlt)]

Comment on attachment 8511104 [details] [diff] [review]
Reenable MSE tests accidentially disabled in last update.

(In reply to James Graham [:jgraham] from comment #4)
> I wonder if the
> ones which are different between Linux and Linux 64 are actually unstable?

Yes, there are some that are not stable yet, and so should not be enabled.

Let's focus here on reverting the accidental changes from bug 1087165.

Comment 6

[Karl Tomlinson (:karlt)]

c27a4a16bf74 didn't change any media-source tests, so expected results should be the same.  I ran hg revert -r c27a4a16bf74\^ testing/web-platform/meta/media-source

Like attachment 8511104 [details] [diff] [review], this reverts the change for NT 5.1.2600, which didn't seem related to the pull from upstream.

https://hg.mozilla.org/integration/mozilla-inbound/rev/bfa9cf004a66

I expect [Test appendBuffer completes previous seek to truncated duration] and
[Test seeking beyond media duration.] are probably still unstable.

I couldn't reproduce the 32-bit failure recorded in attachment 8511104 [details] [diff] [review] for mediasource-config-change-webm-a-bitrate.html.
https://treeherder.mozilla.org/ui/#/jobs?repo=try&revision=a4a14559fb26

Comment 7

[Karl Tomlinson (:karlt)]

https://hg.mozilla.org/mozilla-central/rev/bfa9cf004a66