81365 - Correct behavior of ResetInputState() for mac/

Status: VERIFIED DUPLICATE of bug 81356

(Core :: Internationalization, defect)

Description

[Masaki Katakai]

This is a bug for widget/src/mac of bug 81360.

I'd like to ask Windows and Mac owners of widget/ to check the
bug 53025 still happens with the patch in bug 81356. (or after
the patch is integrated)

On gtk/, some problems happens so I've filed bug bug 81360.
If it works, please mark as WORKSFORME.

Comment 1

[nhottanscp]

Katakai san, is this supposed to to go widget group?

Comment 2

[Masaki Katakai]

Yes, please.

Can anyone verify that that patch

http://bugzilla.mozilla.org/showattachment.cgi?attach_id=34920

can not cause any problem on Mac? I want to know
the results when IME candidate window is invoked.
I understand IME candidate window will never grab input focus
so it should work.

I want the result before bug 81356 is fixed.

Comment 3

[nhottanscp]

I applied the patch on Macintosh, I don't see any problems.
Any specific functionality that I should look at?

Comment 4

[nhottanscp]

Roy's comment in bug 81364.
> Without you patch, the candidate window stays up when user changes the input
> focus but it commits the entry. (which is undesirable)   With your patch, the
> candidate window closes ( which provides a visual clue to user) and the entry 
> is commited.
The same result for Macintosh.

Comment 5

[nhottanscp]

Reassign to [e-mail], let me know if more test is needed.

Comment 6

[Masaki Katakai]

Roy san and Hotta san, can I ask one more testing?

What will happen when you close the dialog with composed text?
Please try the following if you have time,

1. apply the patch
2. start Mozilla
3. Search -> Find In This Page
4. Japanese conversion on on the text field on the dialog
5. input some japanese character
   do not commit characters
6. click cancel button to dismiss the dialog

On UNIX platform, I'm seeing crash problem. It seems that
IME tried to commit the composed text but the window is
being destroyed. I'm not sure it's gtk specific problem.

Comment 7

[nhottanscp]

I did not see the crash on Macintosh. My build is pulled sometime last week (not
remember the exact date).

Comment 8

[Masaki Katakai]

No problem with patch of bug 81356.
Mark as dup of bug 81356.

But still gtk has problem as bug 81360.


*** This bug has been marked as a duplicate of 81356 ***

Comment 9

[Andreas Becker]

Verified as dupe, please reopen in case someone disagrees.