The roads I take...

Here's a short summary of Mozilla-related work I've done in week 21/2012 (May 21 - 27, 2012):

• CSI:Mozilla / CrashKill:
  Tested the new URL list tab on crash lists and found a bug it caused, which was subsequently fixed on Friday.
  Tried to test db data for explosiveness but found a permissions error that will get fixed soon.
  Figured out a list of top Flash issues to possibly talk about during the stability work week.
  Kept track of the notification for ol Flash versions going live.
  Continued the discussion of crash reporting for the desktop web app runtime with the developers.
  Made explosiveness reports for FennecAndroid use ADUs data, which is available now that we're using data from the DB.
  Helped triaging topcrashes for the new mobile Firefox beta.
  I also worked a bit on some data I wanted to have around for the stability work week.
  Just like every week, watched new/rising crashes, caring that bugs are filed where needed.
• SeaMonkey:
  Reviewed a Data Manager patch that unfortunately will need some more work.
• Various Discussions/Topics:
  Android XUL Fennec display problems, stability work week planning, SeaMonkey infrastructure, switch to MPL2 headers, community event calendar, Gaia UI units, privacy laws and cookies, LGPL and Mozilla code, etc.

It looks a lot like crash data on the new Firefox Beta for Android is progressing nicely, every beta we ship performs better there, reporting in with lower crash rates. Please continue testing them and send crash reports to us when you encounter such problems!

Also, on Friday I'll be flying off to California to spend two weeks in the Mountain View office (the second of those being the stability work week) and then somewhat short of two weeks of vacation time driving through Northern California. If you're in the Bay Area and want to meet, let me know!

Here's a short summary of Mozilla-related work I've done in week 20/2012 (May 14 - 20, 2012):

• CSI:Mozilla / CrashKill:
  Lots of discussions around blocklisting old Java versions, we'll go for an extra soft approach as a first step due to the high usage of Flash.
  Following an error of Adobe that we found out about, there was an interesting couple of Flash reprocessing events that we needed and I pushed through.
  As we shipped our first beta of the new Firefox for Android, I helped a lot in monitoring and looking into the crashes we are seeing there.
  On the weekend, pinged people on a heavily explosive Nightly crasher introduced by JS code landed on Friday. People were responsive even on a Sunday to look into the issue, review and land a fix.
  Stayed involved in the discussion for Java signature improvements.
  Reworked my Flash and explosiveness reports to fetch their data from the Socorro database instead of the CSV files.
  Just like every week, watched new/rising crashes, caring that bugs are filed where needed.
• German L10n:
  I synched up a few strings in core and DOMi L10n to trunk again.
• Various Discussions/Topics:
  SeaMonkey build system work and website infrastructure problems, Android XUL Fennec display problems, stability work week planning, plans for moving merge/uplift day, HTML forms and mobile, web apps on Linux, HTML5 Mario prototype, etc.

The new Firefox for Android is in Beta now, and for phones it's a really huge step, it launches faster than most other apps, and still has all the Firefox power on websites - even including Flash. Yes, we needed to sacrifice our "webby" technologies for UI design to get there, but we'll go even farther with those on our future phone OS.

If you have an Android phone, help testing this beta (that is, if it has an ARMv7 processor, we're still working on a version for ARMv6). We know about a couple of crashes, but we need more reported, including steps to reproduce if possible, to get this one to be as stable as one would expect from a browser baring the "Firefox" brand. Everything we catch now in beta and we can fix here, will not hit the masses when we go to release.

That said, I hear it's pretty good already (I don't have an Android phone by choice, wanted something even more open), so you can feel good about trusting it with your everyday browsing tasks on your Android phone. And at the same time, you'll be doing good by testing it and - in case you crash it - by sending a report we can then look at and fix!

Here's a short summary of Mozilla-related work I've done in week 19/2012 (May 7 - 13, 2012):

• CSI:Mozilla / CrashKill:
  More discussions and forming of a concrete plan for Rapid Betas on Socorro.
  Found an interesting signature escaping problem in Socorro.
  Did some triage of mobile crashes with Noaki and Sheila, with an eye on the upcoming beta.
  Discussed and made a proposal for Java signature improvements.
  Followed up on WebRT and B2G crash reporting process, will probably need some more poking until things actually happen there.
  I tried to move things forward on potentially blocking old Flash versions with actively exploited security vulnerabilities, and involved the security-group in the discussions. I hope we'll come to a conclusion on this now.
  Just like every week, watched new/rising crashes, caring that bugs are filed where needed.
• SeaMonkey:
  When people saw a strange website reversal to old days, I poked IT people to get this resolved and they found a temporary solution. There's a larger problem behind this which doesn't only affect our site and which they are working on.
• German L10n:
  I synched up core and suite L10n to trunk once again.
• Themes:
  While the 2.9 versions of EarlyBlue and LCARStrek got their AMO reviews, I started working on 2.10 versions fitting with current beta builds.
• Various Discussions/Topics:
  More datacenter move fallout, Android XUL Fennec display problems, stability work week planning, etc.

As we did get the OK for our stability "work week" (actually more of a meeting week) finally, I have now booked my flight for that as well as one more week leading up to that where I can meet with Mozilla folks in Mountain View and possibly San Francisco - and slightly more than a week of vacation following the event. I'm traveling on June 1, if you are in the Bay Area and want to meet between then and June 11, let me know!

Here's a short summary of Mozilla-related work I've done in week 18/2012 (April 30 - May 6, 2012):

• CSI:Mozilla / CrashKill:
  Had tons of discussions, wrote an initial spec for Rapid Betas on Socorro and handled some questions that came up on it.
  This also brought up availability of old graph data, for which I filed a bug.
  Some more poking about the top Firefox 14+ crasher, which ended up in a backout of the cause in Aurora, so we only need to get an actual fix in Nightly 15 now.
  Blogged about The Life Cycle of a Crash.
  Just like every week, watched new/rising crashes, caring that bugs are filed where needed.
• Various Discussions/Topics:
  Ongoing datacenter moves, HTML L10n, getting de L10n ready for Fennec Beta, padlock and favicon, WebAPI security discussions, Persona marketing, stability work week planning etc.

I spent a lot of time this last week at a local FLOSS event called "Linuxwochen", which was coupled with Linux Graphics Meeting (LGM) and BSDDay this time. There, I talked to a lot of people, representing Mozilla (after all, I'm the local Mozilla Rep, and I guess I was the only core Mozillian there - was there any other?) and trying to tell people about all the cool stuff we are doing while at the same time helping them with questions and trying to remotely diagnose why some could have memory and responsiveness problems with Firefox (which is still the largest complaint and a persistent image problem).

It would have been really helpful to have a B2G demo device to show them how capable such a system is, as the reaction was quite split when I mentioned that system - one group cheered, another frowned. A lot of reasons for the frowns are those you expect from that crowd we had there - finding that JS is a bad language (if one at all), that JS should be disabled as much as possible and apps would undermine that by not running without JS, or even the mindset that web technologies should stay restricted to a low-privilege canvas withing a heavy browser UI. I tried to, but it's hard to convince anyone with one of those opinions that we'd be on the right track by pushing web apps. The cheers were mostly for the the very open stack or simply noting that we're absolutely spot-on that this is where things are moving anyhow and Mozilla as a player with the mission we have is important there.

I also could convince at least one LGM attendee of switching to Fennec (Beta) on his Android tablet, saying it's by far the best browser he's seen on a tablet so far (yay for Gecko being fast on complex Slashdot pages he used as test cases) and make him cheer even more for me telling that the native builds coming up will improve speed even more, mostly in terms of UI. I'm convinced a couple of people I talked to will try native Fennec once it's in the store, I got at least one to also take a look at its Nightly version.

All in all, a lot of interesting conversation, I learned a lot of what people are thinking about us, communicated a lot of what's happening in Mozilla, and had a lot of fun with developers from Austria and around the world. After those days, I got back to the routine of work somewhat tired, but satisfied.