We are changing the opt-in process for this final release to include a
testing review. We have asked for this in the past but this is the
first time we'll be requiring that you complete testing before you opt
reason is we want to ensure the standard of quality we've all come to
expect. For new locales this is especially important so please lean on
l10n-drivers and other localizers to help. So, as we get closer to
final release of Firefox 3 all localization teams need to be ready to
start ramping down engineering and translation work, and start ramping
up the amount of testing, watching feedback channels, and also focus
on any remaining critical bug fixing and translation work.
To help support you all l10n-drivers will be working this weekend and
will be hanging out in l10n channel. Any of us can help so feel free
to ping: pascalc, mic, chofmann, Pike (Axel) and from QA: tchung,
Here's what we will be expecting before you opt-in:
1 conduct litmus test: <https://litmus.mozilla.org> or <https://
litmus.mozilla.org/run_tests.cgi?test_run_id=16> look for "Firefox
3.0 l10n Localizer Test Run". For instructions on how
to run these tests go to <http://quality.mozilla.org/node/71>.
2 Reporting your litmus results - you can pass/fail the tests from
the Firefox 3.0 l10n Localizer Test Run, and QA will run a query on
your results. For those that run tests in other areas (like Fx3.0
bfts, ffts), we
would have to know the testgroup, date, and locale used, and QA can
query results. To get this information please open a tracking bug and
comment there or email l10n-d...@mozilla.org.
3 Gain community or beta tester reviews - send us any links to
communities you are monitoring for feedback, for example we've seen
this community in China, <http://hall.sociz.com/index.php?showforum=7>
Here's data that we have about your locale, can you tell us what you
think about it:
* number of untranslated strings <http://l10n.mozilla.org/dashboard/>,
in particular <http://l10n.mozilla.org/dashboard/unchanged.html>
** are the strings that are not translated highly visible in product?
** are there strings that need to be translated before RC1?
** if so, do you think you can you get it done in time?
Here is a minimum list of dialog boxes that are important and highly
visible that must be in good shape for RC1, in addition to those
already covered by the litmus test run:
* help box for updates
* SUMO integration, from preferences, too
* Add-ons manager, suggested add-ons should work
If you would like to run a testing event to get the work done feel
free to email se...@mozilla.com for ideas and potential sponsorship
help. You can also contact me (m...@mozilla.com) directly for feedback,
RTL locales - Please follow us in bug 364536.
beltzner volunteered Ryan Flint for review, but if anybody has skills in
writing and debugging skins, feel free to jump in and help.
This work should be done for RC1, the current deadline for RC1 is
supposed to be Tuesday April 8. I'm watching http://wiki.mozilla.org/Releases
and beltzner for the latest and will post if anything is new
Please make your tests etc and opt in *before* (or by) then. Please do
not hesitate to ping any of l10n-drivers if you need or want help!
Um, I'm confused. What is meant by "help box for updates"? Can I somehow
spawn this box so I can check it? And are localizers now also
responsible for making sure AMO suggested extensions work ... or is the
third point just about "Get Extension" dialog functionality?
Also, does SUMO integration mean that the Help-buttons work and link or
something more. As Francesco notes in another message in this thread,
SUMO currently has a placeholder page for en-US too...
"Help box for updates" is a cryptic way of saying "check for updates in
the help menu".
For the add-ons manager, you should test that the "Get Extension" pane
works. Which it doesn't for localizations with transcribed product
names, so much I can tell.
> Also, does SUMO integration mean that the Help-buttons work and link or
> something more. As Francesco notes in another message in this thread,
> SUMO currently has a placeholder page for en-US too...
That sumo still links to en-US placeholders is darn unfortunate. Is
there a bug on that?
Type about:config and look for network.http.redirection-limit , set it to 0
This will disable the http redirect.
Click on the help menu item and you will no longer be redirected to the
placeholder. The URL should contain your locale and operating system. In
a nightly build I get :
Mac OS X: http://support.mozilla.com/1/firefox/3.0pre/Darwin/it/firefox-help
Win XP: http://support.mozilla.com/1/firefox/3.0pre/WINNT/it/firefox-help
I can't find any bugs about removing those rules from .htaccess and I
don't know if SUMO is ready to take the load (last time it wasn't).
Maybe we should ping also David Tenser or Chris Ilias about this topic
(adding David to cc of this mail).
The French (fr) locale opts-in. We made the Litmus tests for Windows
Vista, Windows XP, GNU/Linux and Mac OS X.
Litmus tests (I run personally tests for Mac OS X, Linux and Windows Vista).
Also: since beta2 I've been using nightly versions on Mac OS X and
Windows (both Vista and Windows XP) for everyday web browsing, and the
person who's updating Italian version of SUMO has been using nightly
builds on Linux in the last weeks.
Since Firefox 3.0b2 we've been using our forum to test, fix and improve
I checked the untranslated strings and I've localized a couple of
strings in the installer; all remaining strings are correct (Italian
word is the same of the English one).
Litmus test completed successfully for Linux, Windows XP/Vista.
Some little testing chatter intermixed with other messages on our forums
at http://www.mozilla.fi/keskustelu/. Untranslated strings ok (generic,
same as en-us, low visibility), other raised issues ok too.
The litmus tests ran on Windows XP, Linux and Windows Vista, but I
couldn't run the help window test since the online help doesn't seem
to work just yet.
The untranslated strings are either the same in Romanian or they
describe security/internet terms.
We don't have a forum just yet, just the old mozilla-ro group on
We have found that two accesskeys are equal so one "eats" the other.
Just reporting the test as Failed and fixing the key on CVS is enough?
Alexander L. Slovesnik a.k.a. Unghost
==>Jabber ID: ung...@mozilla-russia.org
==>Gmail Talk ID: ung...@gmail.com
Most of them are the same in nl as in en-US
Arabic (ar) is opting for RC 1.
Most of them are strings that cannot be translated.
Do you have testers on Mac?
Do you have testers for Mac and Windows?
I see tons of tests submitted on Windows and Linux :-) , do you have
testers on Mac, though?
I see windows testing, how is it looking on Linux? And I know it's
looking bad on Mac, right?
I'm taking notes. I won't be able to organize real testing for you, but
I should be able to get you at least some data.
It would be nice if you could include the frisian version also, because I
cannot seem to find a mac user until now.
Isn't there a way to run a Mac on vmware or something?
Alexandru, doing some mac testing (more on that separately) I found that
the Romanian localization on Mac doesn't quit on apple-Q, but apple-E.
Is that standard for Romanian Mac software, or is that a bug?
Sadly, OSX only runs on mac hardware, IIRC, you're violating the license
if you do anything else.
Included in test (see other thread), but:
There command key for quit is broken. It's set to apple-o, but that's
I guess it should be apple-q, like for all mac apps.
es-ES is opting in to RC1, we have done some test in Linux, Windows
and Mac, and all things we have saw has been fixed.
I did the Litmus tests for Linux and found things to be good. Link is
in the real "opting in to RC1" thread.
I haven't had the chance to test the work done on the Mac theme yet...
There's http://l10n.mozilla.org/~axel/screens/ar/mac.html to scare
children. Though nothing looks quite as bad as the main toolbar, really.
ca opting for RC1
Daily usage of users in different platforms.
Unchanged strings are those which are the same as in en-US.
Some window sizes still need to be corrected in Mac, though. Hope this
might be mostly finished this morning (UTC+2) if l10n tree is not closed
Toni Hermoso Pulido
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
-----END PGP SIGNATURE-----