Closed Bug 1493608 Opened 6 years ago Closed 5 years ago

Re-enable test-message-header.js::test_a11y_attrs

Categories

(Thunderbird :: Mail Window Front End, enhancement)

enhancement
Not set
normal

Tracking

(thunderbird68+)

RESOLVED DUPLICATE of bug 1589859
Tracking Status
thunderbird68 + ---

People

(Reporter: jorgk-bmo, Unassigned)

References

Details

(Whiteboard: [Thunderbird-disabled-test])

+++ This bug was initially created as a clone of Bug #1489748 +++ test-message-header.js::test_a11y_attrs was disabled in bug 1489748: https://hg.mozilla.org/comm-central/rev/08bf719c34e9#l7.13 We need to find a way to re-enable it.
Or possibly just remove. It's unclear if there is a real accessibility problem now or not. To test, go to the subject of a mail you're reading. Is it clear that for a subject of "abc", the "abc" is the subject?
7.1--- a/mail/test/mozmill/message-header/test-message-header.js 7.2+++ b/mail/test/mozmill/message-header/test-message-header.js 7.3@@ -67,16 +67,18 @@ function setupModule(module) { 7.12+ 7.13+ test_a11y_attrs.__force_skip__ = true; // disabled for now, see bug 1489748 7.14 } Putting the 'skip' inside setupModule seems unusual to me.
All the other places we have are also inside setupModule. I don't know if it even works reliably from elsewhere.
OK, I mixed it up with test_function.EXCLUDED_PLATFORMS = ["platform name"], which is always below the test function. I'd think test_function.__force_skip__ would also work at the same place.

Just to confirm, the test still fails (as expected). We need to figure out if there is something that needs fixing here or if status quo is acceptable and we can just change the test.

So, this is related to the message header pane: is the subject field working as intended? Do you get a reasonable accessible role there?

Earlier the accessible name used to be prepended with "Subject: " for the actual header value. Is that needed now, or is the dom structure clear enough?

Flags: needinfo?(jpmengual)

(In reply to Magnus Melin [:mkmelin] from comment #5)

Just to confirm, the test still fails (as expected). We need to figure out
if there is something that needs fixing here or if status quo is acceptable
and we can just change the test.

So, this is related to the message header pane: is the subject field working
as intended? Do you get a reasonable accessible role there?

Yes. Subject is good. Unlikely the To field, Subject seems correctly described.

Earlier the accessible name used to be prepended with "Subject: " for the
actual header value. Is that needed now, or is the dom structure clear
enough?

I dont know, maybe Joanie or Samuel know

Flags: needinfo?(samuel.thibault)
Flags: needinfo?(jpmengual)
Flags: needinfo?(jdiggs)

I tried a nightly build, ATM it is messy.

  • For the From/To boxes the situation is the same. In at-spi we are seeing:
  • label From
  • unknown From
    • label MyFromName
      • unknown MyFromName
        • label MyFromName
        • image Add to Address Book

When focus is put by the user on the content of From or To fields, it's the unknown MyFromName object which shows as focused (and indeed the focus rectangle shows up around both the text and the image).

The label From has a label-for relation, but with the unknown From. It is thus not being useful for a screen reader (which shows the content of the focused object, the unknown MyFromName.

What we should ideally have instead is:

  • the focus gets on the internal label MyFromName only, and that object should have role static, not label (since it is not labelling anything else)
  • the label From should have a label-for relation with the two labels MyFromName and the image Add to Address Book
  • the objects with unknown role should have role GROUPING (since AIUI their only purpose is to gather the content)
  • For the Subject box, we are seeing
  • label Subject
  • unknown Subject

i.e. the content of the Subject isn't even exposed! The label has a label-for relation with the unknown Subject.

What we should ideally have instead is:

  • label Subject
  • unknown Subject
    • static Content_of_the_subject

where the label Subject would have a label-for relation with the static Content_of_the_subject

Flags: needinfo?(samuel.thibault)

Ergl, I should have used the Preview. My comment was meant to be separated in two:

  • For the From/To boxes
  • For the Subject box

I used different item characters (* vs -) for these two and for the object hierarchies, but apparently bugzilla mixed them

Status: NEW → RESOLVED
Closed: 5 years ago
Resolution: --- → DUPLICATE
Regressions: 1637536
No longer regressions: 1637536
Flags: needinfo?(jdiggs)
You need to log in before you can comment on or make changes to this bug.