Action items for l10n release as per discussions during XTech

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view

Action items for l10n release as per discussions during XTech

Axel Hecht

here is my personal list of action items that we need to address for
l10n releases and the surrounding ecosystem.

The basic idea I had, and nobody opposed, was to have an l10n-alpha for
b1, an l10n-beta for b2 and have the RCs together. I want a real alpha
and beta and not just nightly testing to get some attractive builds out
there, and to have milestones at which we can check how far we are in
terms of locales. This should get some tracking into our release process
and help the localizers to get wide feedback and better quality in the
released locales.

In terms of shipped locales, some of us know how many different builds
are coming up for QA and build without l10n. Thus, we won't have a good
chance to release slipped locales of 2.0 post-2.0 and pre-2.0.[0.]x.
Given the Thunderbird experience with 1.5, I'm not sure if x is 1 or 2.

So, for 2.0, we have two scenarios, target shipping up to up to 50
locales at once, or pick a set of locales that we don't intend to ship
at all. We mustn't pick a set of locales at release day and say "oops,
those didn't make it". I, of course, advocate for the 50, so here goes
the proposal:

Get those 45-50 locales for alpha (b1), and beta. Get the RCs with a
small as possible delay to en-US. This is to some extent a build issue,
but build is hopefully scaling pretty well here. What doesn't scale well
is QA. We won't get happy campers if we don't precisely define on which
parts we need moco QA, and on which parts we need community QA.

In order to get a huge bunch of alphas, we need to be feature complete
significantly before b1, I'd say two weeks before code freeze. This
- code of course (installer?)
- trademarks specs
- file formats, in particular search plugins and bookmarks (rip out
Two weeks before freeze looks like May 6th to me. Yes, two weeks may be
more time than you actually need, but there are still things like
vacations and university exams etc, so I think that 2 weeks is really
short timed.

I don't intend to freeze the tree before the RCs in order to get the
best possible quality in the localized builds.

Action items that I kept:

- Trademarks: cbeard to get a list of contractual obligations
               cbeard to finalize search plugin format spec
               gerv to update trademarks policy to be more precise.
               axel to work on automation of bookmarks tests (started).
               cbeard is the puck-stop for trademarks approvals for 2.0,
he'll hopefully only has to do very few passes for those.

- Locales: axel to ping zh-TW and track that down. Wrote mail today.

The action items for cbeard may induce code changes, which makes those
the most pressing ones.

Anything I forgot?

dev-planning mailing list
[hidden email]