1515614 - Mozmill failure on 2018-12-20: 15 failing tests due to bug 1508141 removal of tree disableKeyNavigation attribute

Status: RESOLVED FIXED

(Thunderbird :: General, defect)

Description

[Jorg K (CEST = GMT+2)]

TEST-UNEXPECTED-FAIL | [snip]/mozmill/folder-display/test-message-commands.js | test-message-commands.js::test_toggle_read
TEST-UNEXPECTED-FAIL | [snip]/mozmill/folder-display/test-message-commands.js | test-message-commands.js::test_toggle_unread
TEST-UNEXPECTED-FAIL | [snip]/mozmill/folder-display/test-message-commands.js | test-message-commands.js::test_toggle_mixed
TEST-UNEXPECTED-FAIL | [snip]/mozmill/folder-display/test-message-commands.js | test-message-commands.js::test_yearly_archive
TEST-UNEXPECTED-FAIL | [snip]/mozmill/folder-display/test-message-commands.js | test-message-commands.js::test_monthly_archive
TEST-UNEXPECTED-FAIL | [snip]/mozmill/folder-display/test-message-commands.js | test-message-commands.js::test_folder_structure_archiving
TEST-UNEXPECTED-FAIL | [snip]/mozmill/folder-display/test-message-commands.js | test-message-commands.js::test_selection_after_archive
TEST-UNEXPECTED-FAIL | [snip]/mozmill/folder-display/test-message-commands.js | test-message-commands.js::test_tag_keys
TEST-UNEXPECTED-FAIL | [snip]/mozmill/folder-display/test-recent-menu.js | test-recent-menu.js::test_archive_message
TEST-UNEXPECTED-FAIL | [snip]/mozmill/folder-display/test-watch-ignore-thread.js | test-watch-ignore-thread.js::test_ignore_thread
TEST-UNEXPECTED-FAIL | [snip]/mozmill/folder-display/test-watch-ignore-thread.js | test-watch-ignore-thread.js::test_view_threads_ignored_threads
TEST-UNEXPECTED-FAIL | [snip]/mozmill/folder-display/test-watch-ignore-thread.js | test-watch-ignore-thread.js::test_watch_thread
TEST-UNEXPECTED-FAIL | [snip]/mozmill/folder-tree-modes/test-smart-folders.js | test-smart-folders.js::test_folder_flag_changes
TEST-UNEXPECTED-FAIL | [snip]/mozmill/junk-commands/test-junk-commands.js | test-junk-commands.js::test_delete_junk_messages
TEST-UNEXPECTED-FAIL | [snip]/mozmill/message-header/test-message-header.js | test-message-header.js::test_add_tag_with_really_long_label

M-C last good: fa77d367a4b0d4776c0d7fa70c895c9600
M-C first bad: 79cf24341024b84654bb33561e929322af
https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=fa77d367a4b0d4776c0d7fa70c895c9600&tochange=79cf24341024b84654bb33561e929322af

Log https://taskcluster-artifacts.net/KjBU491gTcaC7bgPVRb14w/0/public/logs/live_backing.log says:

EXCEPTION: Message marked as unread, expected read
EXCEPTION: Message marked as read, expected unread
EXCEPTION: Message marked as unread, expected read
EXCEPTION: Timeout waiting for messages to be archived
EXCEPTION: Tag 'undefined' expected on message!
EXCEPTION: Desired selection is: [3] but actual selection is:  [1]
EXCEPTION: Desired selection is: [3] but actual selection is:  [0]
EXCEPTION: Thought we would find row 1 at 126,27 but we found -1
EXCEPTION: Timeout waiting for events: [DeleteOrMoveMsgCompleted,DeleteOrMoveMsgFailed]
EXCEPTION: Header columns didn't change width!  87 == 87

Not sure whether that's related, but I see this a lot in the log:
JavaScript error: chrome://global/content/customElements.js, line 80: TypeError: child is null

Comment 1

[Jorg K (CEST = GMT+2)]

Not sure what it could be, but this one doesn't look good:
adad39927668 Brian Grinstead - Bug 1508141 - Remove support for disableKeyNavigation on richlistbox, autocomplete and tree bindings;r=timdream

We seem to use disableKeyNavigation:
https://searchfox.org/comm-central/search?q=disableKeyNavigation&case=false&regexp=false&path=mail

Here's a run bisecting the range in half at M-C rev 1260c07d4b17:
https://treeherder.mozilla.org/#/jobs?repo=try-comm-central&revision=5db375eeb08b32d8c7c0334c01077ec4ea4d65c6

Comment 2

[Magnus Melin [:mkmelin]]

Pretty sure that would be the one yes.

Comment 3

[Magnus Melin [:mkmelin]]

Discussing in bug 1508141 whether they will keep it for <tree> after all. 
If not, I suppose it's possible to make a binding that extends tree, and override the keypress handler and add back the disableKeyNavigation property.

Comment 4

[Jorg K (CEST = GMT+2)]

(In reply to Jorg K (GMT+1) from comment #1)
> Here's a run bisecting the range in half at M-C rev 1260c07d4b17:
That failed, so
https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=fa77d367a4b0d4776c0d7fa70c895c9600&tochange=1260c07d4b17
hence bug 1508141 as already suggested/confirmed. I could confirm 100% via local backout.

Comment 5

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/comm-central/rev/9fe5af480738
disable 15 failing tests. rs=bustage-fix

Comment 6

[Jorg K (CEST = GMT+2)]

There's another test which intermittently fails on Linux only:

TEST-UNEXPECTED-FAIL | /builds/worker/workspace/build/tests/mozmill/message-header/test-message-header.js | test-message-header.js::test_more_button_with_many_recipients

Log https://taskcluster-artifacts.net/Yk3mrodvTteRZyj2TkXfhA/0/public/logs/live_backing.log says:
INFO -    EXCEPTION: Header Mode didn't change to 'all'!  old=normal, new=normal
INFO -      at: test-message-header.js line 349
INFO -         test_more_button_with_many_recipients test-message-header.js:349 11

First seen after I disabled the 15 failing ones. Could be related or a fallout from disabling the tests.

Comment 7

[Richard Marti (:Paenglab)]

Bug 1515759 is on inbound and should fix the tree failures.

Comment 8

[Jorg K (CEST = GMT+2)]

OK, thanks.

Comment 9

[Jorg K (CEST = GMT+2)]

https://hg.mozilla.org/comm-central/rev/50e06565357c041f6c2ce53bd2efe9bb0f6c15c1
Backed out changeset 9fe5af480738 to re-enable tests after bug 1515759 reverted bug 1508141. a=backout