Bug 44376 – [GTK] onkeyup and onkeydown fires twice in textarea

• Status: RESOLVED
• Resolution: WONTFIX
• Component: New Bugs

[Shellex Wai]

visit http://shellex.soso44.net/test/test_keydown.html for test

= Steps to reproduce =
 - Assign focus to the textarea below.
 - Press a single alphanumeric character on the keyboard. 
 - The event.keyCode of that key will appear in the div.

= Expected behavior =
 onkeyup and onkeydown fires once each time a key is pressed.


= Actual behavior=
 onkeydown fires twice in Chrome and WebkitGtk
 onkeyup fires twice in Chrome

[Martin Robinson]

I cannot reproduce this bug via the latest build of WebKitGTK+ from master or from the version of Chromium in the PPA. Do you have any more information on how to reproduce this?

[Shellex Wai]

I update my debian sid and try 
GtkLauncher in libwebkit (ver 1.2.3-2)
and 
Chrome unstable (ver 6.0.495.0 dev)

then I visit http://lab.shellex.info/test/test_key.html

and I type 'a' in textarea.

the Chrome output as:
[keydown] keyCode:229
[keydown] keyCode:65
[keypress] keyCode:97
[keyup] keyCode:65
[keyup] keyCode:65

GtkLauncher output as:
[keydown] keyCode:65
[keydown] keyCode:65
[keypress] keyCode:97
[keyup] keyCode:65


but my firefox of debian sid and chrome of my android work well, them output as:
[keydown] keyCode:65
[keypress] keyCode:97
[keyup] keyCode:65

Addition, I ask some of my friends to help me verify the problem. 
Some of them can reproduce the same problem and others can not. It's really weird.

Here is my friends' test result:
= Test 1 =
Laptop, keyboard model is acer laptop
OS: Ubuntu 10.04.1
Browser 1: Midori 0.2.7
Result 1:
[keydown] keyCode:65
[keypress] keyCode:97
[keyup] keyCode:65

Browser 2: chromium 7.0.502.0 (57007) 
Result 2:
[keydown] keyCode:65
[keypress] keyCode:97
[keyup] keyCode:65

= Test 2 =
Netbook
OS: Linux netbook 2.6.35-14-generic-pae #20~lucid2-Ubuntu 
Browser 1: Chomium 5.0.375.99 (51029) Ubuntu 10.04
Result 1:
[keydown] keyCode:65
[keypress] keyCode:97
[keyup] keyCode:65

= Test 3 =
Desktop PC
OS: Debian Testing
Browser 1: Midori 0.2.4
Result 1:
[keydown] keyCode:65 
[keypress] keyCode:97
[keyup] keyCode:65

= Test 4 =
Desktop PC
OS: Slackware 13.1 2.6.33.4-smp
Browser 1: Chrome 6.0.495.0 dev
Result 1:
[keydown] keyCode:65 
[keypress] keyCode:97
[keyup] keyCode:65

= Test 5 =
Laptop
OS: ArchLinux
Browser 1: midori 0.2.7
Result 1:
[keydown] keyCode:65 
[keydown] keyCode:65 
[keypress] keyCode:97 
[keyup] keyCode:65

= Test 6 =
Laptop (Hasee HP520)
OS: ArchLnux
DE: Awesome
Browser 1: Chromium 7.0.504.0 (57118)
Result 1:
[keydown] keyCode:229
[keydown] keyCode:65
[keypress] keyCode:97
[keyup] keyCode:65
[keyup] keyCode:65

Browser 2: Google Chrome 5.0.375.126
Result 2:
[keydown] keyCode:229
[keydown] keyCode:65
[keypress] keyCode:97
[keyup] keyCode:65

= Test 7 =
Laptop
OS: Gentoo
Browser 1: Chromium 7.0.504.0 (57042)
Result 1:
[keydown] keyCode:229
[keydown] keyCode:65
[keypress] keyCode:97
[keyup] keyCode:65
[keyup] keyCode:65

I find that this bug only occur to the browsers in laptop in my test case.

And further, the behaviors of chrome and webkitgtk-based browsers are a little different, chrome will fire keyup event twice but webkitgtk doesn't.

I also test webkit browsers in Mac and Windows, they are fine.

[Evan Martin]

I believe keycode 229 is an IME-related notification.  I'd be curious to see what IE does in your test case.  I don't know a lot about this area but I'm not sure Chromium's behavior is necessarily incorrect.  (I think the WebKitGtk behavior of sending the same keycode twice is suspicious, and Chromium's double-send on keyup also looks weird.)

[Martin Robinson]

Shellex, what input method are you using?

[Shellex Wai]

(In reply to comment #4)
> Shellex, what input method are you using?

I use ibus as my input method, although it's not enabled in the test process.

(In reply to comment #3)
> I believe keycode 229 is an IME-related notification.  I'd be curious to see what IE does in your test case.  I don't know a lot about this area but I'm not sure Chromium's behavior is necessarily incorrect.  (I think the WebKitGtk behavior of sending the same keycode twice is suspicious, and Chromium's double-send on keyup also looks weird.)

I quit ibus, then browsers behavior normally. Well, I am not sure is it a webkit bug or not...

[Michael Catanzaro]

Website is no longer available.