« previous week | index | next week »

Meeting Details

• 9:00am Pacific, 12:00pm Eastern
• 650-903-0800 x91 Conf# 206 (US/INTL)
• irc.mozilla.org #l10n-drivers for backchannel
• (guest notetaker: clouserw)

Attendees

Pascalc, Axel, Mic, Seth, Chofmann, Timr, Clint, Wil, Gavin, Schrep

Agenda

• Back-up plan for Axel
  • Gavin focus on coverage for Axel for l10n source verification. Has update script, the moves into reference files and has comparison file script (where shouldn't' run update script anymore).
  • Clint focus on Minotaur and following up with RobC to check in tinderboxes for L10n; will also look at L10n source to see what output should be
  • Chofmann cover on manual tests with Mic et al

• QA needs for Firefox 3
  • Language quality testing: French, German, Polish - Timr would like a quick review to determine if continuing with external review is useful.
  • ACTION: Timr will post results to people and publish link to TomCat, Pascal, Gandalf/Stasz; send feedback to Timr
  • Go forward plan:
    • (1) chofmann download and test for key things tested for fx2 for fx3-b3. do all that testing by hand. will do that for all locales. QA team (tracy or tchung, will depend on who is available) can continue to support this (to run manual tests). do full set of Beta3 builds.
    • (2) timr will figure out if we can get minotaur behaving in reliable and consistent way. what's minimal set of things we need. use it to save human's time. shrink to fit goal of good set of localizations at end. (takes set of preferences, query url from search engines - not search forum or search suggest, bookmarks - compares between old/new builds to determine diff's). Start running minotaur on updates we need to do. AXEL, TIMR, CTALBERT. Call at noon to 1:30 PST, review and write up of what's needed to Schrep et al.
    • (3) have all localization team send us mail when finished litmus test for their locale. get that litmus test plan together and published and point teams to it to do same process (tony should have this - axel needs to do a review of this)

Adendum to notes from follow-on call:

Who: Axel, Clint, Timr, Chofmann

Discussion:

  1. Axel would like to add Search Suggest and Search Form (hitting enter in search box leads to search page).  This seems like an easy addition and add a lot of value.
  2. Finnish (fi) - reporting same things as source verification for ffx 3. Good!
  3. Changes in the builds to bookmarks are seen in Axel's source verifier but not in minotaur results.  Why don' they?
  4. Results from tracy using latest Minotaur show bookmarks are "under-represented" rather than "missing" as the source verifier shows.  The results in the minotaur log should show bookmark hierarchy, titles, number of links, and URLs more clearly.  Results reporting should be cleaned up to be more human readable.
  5. Need to get the build test automation worked out.  We need to get ref files captured and versioned.  Pass these to robcee for automation.  Need to capture ref files from  B1 and B2.  These are preferred over the current state of B3.  This won't have the protocol handler info which is landing after B3, but this is ok for now.  It is critical to get the reference files and automation for FFx3.  FFx2 ref files and automation is not a priority at this point. Clint will work with Robcee on this.
  6. The feature of custom protocol handlers for locales landing after B3.  We need to add tests for this to Minotaur

Next Steps:

   * Add search suggest and search form to minotaur (See #1 above) - Clint
   * Output verification - Create ref files from B1 and B2 to be used for B3 (see #5 above) - Clint/Tracy
   * Human readable comparison files - get results cleaner (See #4) - Clint
   * Run/verify on all FFX3 B3 locales - Clint/Tracy/Axel/Mic
         o This gets us to the important milestone for the tool accurate, complete and reliable for FFx3 -
   * Add protocol handlers to minotaur - Clint
   * Finalize ref files - Clint/Tracy/Mic/Axl
   * Add automation  for FFx3 - clint/robcee
         o 2nd important milestone: automatically testing L10n builds to catch problems earlier

Notes on issues: two problems: what data should look like that gets passed over (tools and run time problem) Which bugs have been written against firefox or tool en-US according minotaur test shipped without any links in bookmarks file - would consider raising QA flags something odd in test run now in situation with 2 days of work, not much Axel can still do in helping to vet out bugs in current situation Axel can't help to ensure the Minotaur is reporting right stuff in right format bookmark issue was another default generation of bookmark regression in Fx2. tried to get places team to fix for B3. gave tracy a simple work around which is ok but not optimal is tracy swamped? We revert to manual process for B3 to use. Maybe we'll be able to do improvements to automation work on Minotaur before B4.

see scaling problem - sheer number 70 languages across 3 platforms is bigger problem

next testing priority is depth testing across all locales. check out bookmakrs, search engines, etc like we did for Fx2. if we can get tools up to help with that, that'll be great

• Tools/Fosdem
• Updates: community, AMO
• Roundtable