1355842 - Fix and re-enable layout/base/tests/test_reftests_with_caret.html on Android

Status: RESOLVED FIXED

(Core :: Layout, enhancement, P3)

Description

[Ryan VanderMeulen [:RyanVM]]

Attachment: test log

As part of getting layout/base/tests running on Android again, it was determined that test_reftests_with_caret.html is currently permafailing. In the interests of getting the directory running again, the test has been disabled for now until it can be fixed and re-enabled.

Comment 1

[Ryan VanderMeulen [:RyanVM]]

NI myself to see if we can re-enable these now.

Comment 2

[Ryan VanderMeulen [:RyanVM]]

Lots of failures still on Android:
https://treeherder.mozilla.org/logviewer?job_id=360330118&repo=try&lineNumber=5750

Masayuki, do you have any thoughts on what's going on here?

Comment 3

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Attachment: Result of input-password-RTL-input.html

Comment 4

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Attachment: Reference of input-password-RTL-input.html

Comment 5

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Attachment: Result of bug956530-1.html

Comment 6

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Attachment: Reference of bug956530-1.html

Comment 7

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

About input-password-RTL-input.html: It assumes that editor.password.mask_delay pref is 0 by default, but only on Android, it's not treated as so with the default settings (-1). So, explicitly specifying the mask delay can fix this.

About bug956530-1.html: It checks selection range after selecting all text in <input> from JS and set focus to the <input> after blurring by synthesized mouse clicks. However, only on Android, there is no caret nor selected range. I'm not sure whether this is intentional behavior on Android or not. Ask somebody who is working on Android.

About timeout of bug1097242-1.html: It tries to non-selectable and non-editable <span> element in <div contenteditable> with a click, but the <div> does not get focus with this, that's the reason of the timeout, but I don't know why it fails on Android.

Comment 8

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

My job was here: https://treeherder.mozilla.org/jobs?repo=try&revision=89f135fd40c23a5d15fef7dda70ba859f0041d1f

The patch includes a fix to dump the reftest failure results even in Android.

Comment 9

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

The failure of bug956530-1.html could be caused by failing to click the <input> element if so, it's same as bug1097242-1.html's failure.

Comment 10

[Makoto Kato [:m_kato]]

Does apz.zoom-to-focused-input.enabled=false help this issue?

Comment 11

[Makoto Kato [:m_kato]]

input-password-RTL-input seems to be password mask issue.

Comment 12

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Let's try with fixing the prefs...
https://treeherder.mozilla.org/jobs?repo=try&revision=ff80754a29137a9ffcbaeb807d9af6a9cf083636

Comment 13

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Attachment: Bug 1355842 - Re-enable layout/base/tests/test_reftests_with_caret.html on Android r=emilio!

All of them are still disabled for some test failures. There are 3 patterns
failures:

1. Tests with <input type="password"> assumes that last input character is
   masked immediately.
2. Tests with synthesizing mouse button events in editable element
3. Test with DnD

The first case can be fixed with disabling echoing last input character by the
pref.

The second case can be fixed with disabling "zoom to editable element" by the
pref.

The last case cannot be fixed, therefore I disable it on Android.

Comment 14

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/ba22a155be2e
Re-enable layout/base/tests/test_reftests_with_caret.html on Android r=geckoview-reviewers,emilio,m_kato

Comment 15

[Cristina Cozmuta (:CrissCozmuta)]

https://hg.mozilla.org/mozilla-central/rev/ba22a155be2e