432773 - a few additional key tests

Status: RESOLVED FIXED

(Core :: Widget: Cocoa, defect)

Description

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Attachment: patch

-- Add PR_LOGGING code so that Cocoa can dump the available keyboard layouts on startup. This is helpful for creating tests.

-- Remove the assertion that aNativeKeyboardLayout is nonzero, since zero is a perfectly valid layout ID (it's the "US" layout).

-- Add tests for XUL <key> activation on Mac and Windows.

-- Refactor the test code a bit.

-- Use the "US" layout on Mac instead of "US-Extended" for consistency with Windows.

-- Add test for bugs 359638 and 429160 on Mac and Windows.

-- Add test for bug 432112 on Windows. The test currently does not work on Mac because the existing testing code can't get the layout resource for Swedish.

Comment 1

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

(In reply to comment #0)
> -- Use the "US" layout on Mac instead of "US-Extended" for consistency with
> Windows.

Maybe, we should not use LMGetKbdType API for the testing. Because the testcases assume that the target keyboard type is ANSI for US/US-Extended. See bug 432388 comment 16 to 18.

Comment 2

[Robert O'Callahan (:roc) (email my personal email if necessary)]

I'll switch the tests back to US-Extended. Does US-Extended actually have the same problem --- are there JIS US-Extended layouts?

Comment 3

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Attachment: patch v2

The only change from the previous patch is to leave the Mac US keyboard layout as US-Extended. There may still be a problem with non-ANSI US-Extended layouts, but at least this patch isn't making things worse, so we should still take it.

Comment 4

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

(In reply to comment #2)
> I'll switch the tests back to US-Extended. Does US-Extended actually have the
> same problem --- are there JIS US-Extended layouts?

Yes. But if we are only using alphabes/numerics, then the all keys are same value between ANSI and JIS. If we need to test symbols (e.g., ':'), we should fix it.

Comment 5

[Robert O'Callahan (:roc) (email my personal email if necessary)]

We should, but I'm not sure how, and I'd still like to land this patch since the Tinderboxes will pass.

Comment 6

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

ok. but it may be easy. see |GetUSLayoutCharFromKeyTranslate| of the patch of bug 432389. I just set 40 to mKbType. That works fine for me.

Comment 7

[Josh Aas]

Comment on attachment 319995 [details] [diff] [review]
patch v2

No longer applies.

Comment 8

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Attachment: patch v3

Updated to trunk. This should pass on Tinderbox (10.4 with a real US keyboard). These tests still don't pass on 10.5 and won't pass with a non-ANSI US keyboard. We'll fix those issues separately.

Comment 9

[Josh Aas]

landed

Comment 10

[Karl Tomlinson (:karlt)]

+// test currently does not work, getting the Swedish layout fails

I suspect Matthew's Keyboard Layout Services work could help here.