Closed Bug 897467 Opened 11 years ago Closed 11 years ago

Duplicated accesskeys for selectAllCmd and copyCmd which cause misbehavior (opens Printer or Library) instead

Categories

(Mozilla Localizations :: ku / Kurdish, defect)

defect
Not set
normal

Tracking

(firefox25 fixed, firefox26 verified)

RESOLVED FIXED
mozilla26
Tracking Status
firefox25 --- fixed
firefox26 --- verified

People

(Reporter: mario.garbi, Assigned: flod)

References

Details

(Whiteboard: [mozmill])

We caught this with the help of mozmill-tests and it looks like the dtds value for selectAll is set to 'P' which is the same for Printer. Not sure if this intentional for ku localization or a regression but both "selectAllCmd.key" ("P") and "printFrameCmd.accesskey"("p") have the same key.

Here is the changelog responsible of this:
http://hg.mozilla.org/releases/l10n/mozilla-aurora/ku/rev/baf78810a202#l6.266
This is actually a broken behavior. DTD entities do not differentiate between small and capital letters. So both would end up opening the print dialog. Select all should really get a different key assigned to.
OS: Linux → All
Hardware: x86 → All
Whiteboard: [mozmill]
Blocks: 895996
Bug 894513 also fails due to this changeset, at this line: 
http://hg.mozilla.org/releases/l10n/mozilla-aurora/ku/rev/baf78810a202#l6.247

Using "copyCmd.key"("j") it will open Library modal instead of copying selected text.
Summary: DTDS shortkey for "Select All" opens Printer instead → Duplicated accesskeys for selectAllCmd and copyCmd which cause misbehavior (opens Printer or Library) instead
Shouldn't this be a trivial fix we could fix on our own? Do we really need someone from the Kurdish locale for that? I would love to see our Mozmill tests passing for the ku locale.
As I said in bug 894513 comment 5, this is a wide-spread problem in the kurdish localization, 'j' is everywhere.
We could fix at least this single instance so that our existing mozmill tests will pass for this locale. That will help us all.
Fixed in http://hg.mozilla.org/releases/l10n/mozilla-aurora/ku/rev/1179ec7740d5

I still believe that we should start using whitelists/blacklists though.
Thanks Francesco! Lets see what we can do regarding blocklisting specific locales in our mozmill-ci: https://github.com/mozilla/mozmill-ci/issues/309

Mario, can you please test the next Aurora nightly for ku and check if that works? If yes, it would be good to get it backported to beta.
Flags: needinfo?(mario.garbi)
All seems to be ok now, the testrun report show green with Aurora KU on Ubuntu 13.04:

http://mozmill-crowd.blargon7.com/#/functional/report/837c1e0f361ac93453ee697219a1cfb8
http://mozmill-crowd.blargon7.com/#/functional/report/837c1e0f361ac93453ee697219a26f1d
Flags: needinfo?(mario.garbi)
Thanks mario. Francesco, does the patch cleanly apply on beta? I wonder if we can get it backported.
Hurray! So it seems we can close this bug. Mario, please verify with the next beta.
Assignee: nobody → francesco.lodolo
Status: NEW → RESOLVED
Closed: 11 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla26
PS: Beta builds go off of sign-offs, not latest change. I'm not convinced that we should update the sign-off without a native speaker.
(In reply to Axel Hecht [:Pike] from comment #12)
> PS: Beta builds go off of sign-offs, not latest change. I'm not convinced
> that we should update the sign-off without a native speaker.

Oh, that's true. In that case we most likely wont get a signoff in the near future. Anthony, please make sure to stop running ondemand functional tests on ku for now.
You need to log in before you can comment on or make changes to this bug.