632542 - Text entered with SlideIT VKB sometimes disappears

Status: RESOLVED FIXED

(Firefox for Android Graveyard :: General, defect, P2)

Description

[Naoki Hirata :nhirata (please use needinfo instead of cc)]

1. download slideIT VKB
2. set the system to use SlideIT as a input method
3. open fennec to http://people.mozilla.com/~nhirata/html_tp/formsninput.html
4. click in a field
5. try slide typing or typing regularly

Expected: the typing will show
Actual: field is still blank

Comment 1

[Naoki Hirata :nhirata (please use needinfo instead of cc)]

Mozilla/5.0 (Android; Linux armv71; rv2.0b12pre) Gecko/20110208 Firefox/4.0b12pre Fennec/4.0b5pre

I needed some other steps in here:
1. download slideIT VKB
2. set the system to use SlideIT as a input method
3. open fennec to http://people.mozilla.com/~nhirata/html_tp/formsninput.html
4. click in a field
5. type something
6. submit the form.
7. hit the back button
8. click in a field
9. try slide typing or typing regularly

Expected: the typing will show
Actual: field is still blank

Comment 2

[Alex Pakhotin (:alexp)]

I cannot reproduce the bug as it is described, though there is still a problem with SlideIT keyboard in Fennec. Sometimes the entered word gets deleted when a space is typed after that. The main issue is that it doesn't happen every time. Sometimes the word with a space is added correctly, but most of the time we receive deleteSurroundingText call, but no commitText with the updated word after that, which is weird. I'll do more research to understand why this might happen.

Comment 3

[Alex Pakhotin (:alexp)]

Seems like SlideIT VKB sometimes gets confused on our update of the edit field when it asks to delete the last word, and after this confusion it just stops sending us the new text, which had to replace the deleted word. We wouldn't have this problem if SlideIT would do it properly, like other IMEs do: start a batch operation before actually updating the last word, and not after it has already requested to delete it to rewrite. Some workaround is needed to deal with what we have.

Comment 4

[Alex Pakhotin (:alexp)]

Attachment: Fix

This fixes the issue. Tested with 10 different IMEs - seems to be working.

The patch is for the birch branch.

Comment 5

[Brad Lassey [:blassey] (use needinfo?)]

Comment on attachment 569879 [details] [diff] [review]
Fix

Review of attachment 569879 [details] [diff] [review]:
-----------------------------------------------------------------

Please talk with qa to work out a thorough test plan to find any regressions this may cause

Comment 6

[Alex Pakhotin (:alexp)]

http://hg.mozilla.org/projects/birch/rev/a9a7b2ee14bc

Comment 7

[Alex Pakhotin (:alexp)]

Naoki, Kevin, do you test the birch branch at all?
We need to make sure this change does not break anything.