Firefox: synthetic keyboard events do not set KeyboardEvent.key property properly about selenium HOT 11 CLOSED

Switching to sendNativeKeyEvent requires significant amount of work to map Selenium key codes to Firefox ones.

from selenium.

This pull request should be for: #588

from selenium.

is this issue dead?
any ETA on that?

from selenium.

is this issue dead?
any ETA on that?

I'm not a java developer, tho if it's a small fix I'll make a PR with some guidance.
I have a semi completed mapping of keycodes to keys

from selenium.

@goldylucks it's kind of dead... the problem is one would need to refactor the code, since sendKeyEvent is deprecated in firefox and it suggest to use this:
https://developer.mozilla.org/en-US/docs/Mozilla/Tech/XPCOM/Reference/Interface/nsITextInputProcessor

I tried to take a stab at that a long time ago, and ended up not having the time or energy to follow through with it. Also since after FF48, this driver will essentially be dead and GeckoDriver will need to be used in the future, I didn't think it seemed worth it to invest the time in it.

Also, the fix would not be in java at all, purely JavaScript.

from selenium.

@lukeis thanks for the response
it's not only firefox
I'm having the same issue in chrome

do u have an estimation on how much time it will take to fix?

and if so, could u point me where to begin with it?

from selenium.

'less easy' aka 'non-trivial', it will take a decent amount of time... where decent is quite a few hours for someone who knows what they are doing.

This bug has been entirely for firefox and references to firefox driver code. Chrome has a different code path and you'd need to first start looking at the chromium source to figure out how it's exactly doing the sendKey event(s).

Here's where I've been able to trace the function call to in chromedriver down to creating events:
https://cs.chromium.org/chromium/src/chrome/test/chromedriver/key_converter.cc?rcl=1468912738&l=193
And then it sends those key events:
https://cs.chromium.org/chromium/src/chrome/test/chromedriver/chrome/web_view_impl.cc?rcl=1468912738&l=400

Which is a remote request to the Chrome Inspector... to which I don't know exactly where that source code is, but somewhere in the chromium repo.

hope you're good at C? :)

from selenium.

btw "quite a few" is a gross understatement, it could take someone days (for firefox), chrome may be less (in the range of hours), since it doesn't likely require the extensive refactor of the mechanism that is creating the key events

from selenium.

Thanks for the pointers.
Sadly I don't know any C

maybe it's better to close this issue?

On Tue, Jul 19, 2016 at 5:58 PM, Luke Inman-Semerau <
[email protected]> wrote:

btw "quite a few" is a gross understatement, it could take someone days
(for firefox), chrome may be less (in the range of hours), since it doesn't
likely require the extensive refactor of the mechanism that is creating the
key events

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#386 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/AGpADD7qAq7Gz9L_FzdNNL2BwVVob4Vvks5qXPQKgaJpZM4D2auh
.

from selenium.

I intend to close this issue (and really all FirefoxDriver issues logged against this repo) when GeckoDriver is the default and Firefox no longer works with the FirefoxDriver here.

from selenium.

It's time to close this issue, geckodriver is the default and it passes this test.

from selenium.