The roads I take...

Here's a short summary of Mozilla-related work I've done in week 03/2012 (January 16 - 22, 2012):

• Mozilla work / crash-stats:
  The investigations of the Flash hang drop in Firefox 10 came out differently than what we liked - it looks like a UI change caused either people submitting fewer reports or sending of reports failing - in any case, we didn't improve code, but made us get less data to work with.
  After the last Socorro release and merging erroneous Java crash signatures into one, we finally got traction on the issue of properly reporting Java crashes and we'll have better signatures for them in the future.
  My custom reports for startup and per-device crashes are now working nicely (including weekly reports for devices) and listed on our CrashKill Reports page.
  Alerted people that the Banco do Brazil crash is back.
  Tried to get something moving on blocking Flash 10.0 but without success so far.
  Noticed that the growing amount of "ntdll.dll" crashes must be because of a symbol issue and pushed for getting that resolved.
  Alerted the Socorro team on timezone-related regressions and pushed for getting them resolved for the next update.
  And I helped the team with directions on what we'd like to see for a URL list for signatures.
  As usual, watched new/rising crashes, caring that bugs are filed where needed.
• Firefox cleanup:
  After a long time, I now don't just have reviews on some cleanup patches for search bar and engine manager, but also some time to actually work on getting them pushed. After some try runs, I saw that the former parts cause some problem with accessibility tests, so I only pushed the latter. Even though engine manager might get obsoleted soon, this should improve the code for 12 at least - and I still have one patch open to improve and hopefully also get in. And I need to investigate the test failures of the search bar stuff some more, but there's at least some progress.
• Web apps:
  I fixed a dumb bug in Lantea Maps and it now works fine with touch events.
• German L10n:
  I finally did the rest of the needed L10n for the Mozilla 11 core and SeaMonkey 2.8, uplifted it to l10n-aurora and then did some more trunk L10n.
  While at that, I also did run the script to split Fennec L10n for Android and XUL on both -aurora and -central trees.
  On the translation of the "Mozilla Spaces" page, I chimed in with a few comments.
  And I tried to help with another community member stepping in to localize the next version of Firefox.
• Various Discussions/Topics:
  GC scheduling fixes, SeaMonkey build infrastructure, B2G browser, Fennec UA, binary components, BrowserID and L10n, my tablet being fixed and testing Fennec (XUL for now) on it again, etc.

One thing I noticed this week again wasn't just that I was able to quite productive in the end, but also that people from the community hired by Mozilla (as employees or contractors) often don't work in the same roles as they did before and therefore may end up having less time to still have that community stuff working as well as it should. Also, there's some question on how much you can engage with the community or try out other stuff in Mozilla when you work full-time on one part - and are passionate enough to also get some stuff done in what should be your free time, at times. Some localizations, add-ons, and similar projects have been feeling this in recent years and have a hard(er) time surviving due to it. After all, some people claim that "real life" isn't a myth but does exist somewhere out there and it might be an interesting quest to find out about that as well.
I think, with the kind of organization Mozilla is, we might want to think about maybe giving our staff some possibility even in their paid time to engage in the community in roles that are not strictly their work area. Of course, this can be a hairy topic, so all I'm saying is that we might want to think about that - it's far from sure that there is a good solution to be had there, but maybe there is. In the end, the community with its many aspects is what makes Mozilla great and enables us to create products like Firefox that make the life of millions of people better (yes, Firefox started as a fun side project as well, IIRC).

Here's a short summary of Mozilla-related work I've done in week 02/2012 (January 9 - 15, 2012):

• Mozilla work / crash-stats:
  Filed a bug on investigating why Flash hangs less in Firefox 10. It's awesome that we improved, but we would be happier knowing why we did.
  Filed another bug on some libxul addresses never being resolved as prefixlisting libxul can only be a wallpaper, the real bug lies deeper.
  Chimed in on another skiplist request that has different causes.
  Worked on new custom reports for startup crashes and per-device crash lists, which basically work now.
  In that work, found a configuration change that threatened our custom reports and dealt with it (I'm now setting umask in my reports directly, not in the crontab).
  Followed the Socorro 2.4 release work and deployment, which caused a dip in crash numbers on Friday, but it's better to have that now than somewhere near to a release where we need reliable stats every day.
  As usual, watching new/rising crashes, caring that bugs are filed where needed.
• SeaMonkey:
  Followed our Parallels host going down unexpectedly and doing it again, this might be once again related to this virtualization solution not working well with overcommitting.
• Various Discussions/Topics:
  Fennec UA, "image-suck", testing OpenGL layers again and still seeing problems, my tablet being repaired and on its way to me again, etc.

I've seen that my previous work style created a lot of personal stress so I'm now trying a variation to try and deal with it somewhat (fitting with my post on change) and try to do fixed work times every day to see if that reduces the stress potential. For now, I'll try to work from noon to 9pm every work day (with an hour in between for a meal and grocery shopping), though I'll probably be somewhat flexible about that, so I can make some evening appointments at some days or be able to have some late meetings sometimes. Still, it centers around fixed times and I'm setting an alarm to get up everyday. I hope this more regular schedule helps me somewhat. We'll see.

Here's a short summary of Mozilla-related work I've done in weeks 51/2011-01/2012 (December 19, 2011 - January 8, 2012):

• Mozilla work / crash-stats:
  Adapted all my custom reports scripts for the new versions coming up with the channel uplift on Dec 20.
  More discussions on topcrashers tracking.
  Took part in discussions on shipping a 9.0.1 version to all users already having 9.0 and monitored the release discussions over the holidays on slowly ramping up updates to 9 while we have been understaffed due to holiday season.
  Watched the Socorro team deploy changes to support a separate product for Android Fennec.
  Updated and finalized Q1 priorities the CrashKill team has for Socorro.
  Some discussions on possible skiplist additions and some stack frame addresses not being resolved as they should be.
  Discussed fixing timezones on Socorro (this ended up getting its own release this week).
  Took part in discussions on what Socorro has done in the last quarter and what the plans for the next quarters are.
  More watching of new/rising crashes, caring that bugs are filed where needed.
• Web Apps:
  Updated Mandelbrot and especially Lantea Maps web apps with some additional ideas and fixes, the latter now has the basic functionality, but is still somewhat buggy. If anyone's interested on working to make the Lantea Maps app a really good mapping and tracking app, please contact me, help would be appreciated!
• Various Discussions/Topics:
  Factor prefixes for byte unit, MPL 2.0 release and applying it to Mozilla, Firefox ESR plans, to require or not to require SSE2, warnings as errors, WebAPIs, B2G UI proposals, native Android builds still getting development on Aurora but apparently getting close to being usable, trying to get my tablet problems resolved, etc.

This report is both covering 3 weeks and running late despite that, but it's not due to laziness on my side, it's most due to the holidays, as I took two weeks mostly off and then had a number of backlog stuff to work on this week. I followed bugmail and most planet and newsgroup traffic I'm interesting in even while I was at home with my parents, but I didn't go too deep into anything during that time, and it was good to not strain myself too much for some time. Also, I'm happy I found the time to write about all the change going on in my life but also all around here at Mozilla. This week I'm back to work with full power, and a lot of stuff is happening as a lot of exciting goals have been set at various levels in multiple Mozilla project for this first quarter of 2012. And I have a couple of interesting milestones and events to look forward to myself in this new year!

Like many others, I've been thinking quite a bit these days about what went on last year and what will or might come up in 2012. (And I figure I should bring in a bit more from my overall personality into my future blog posts and mention or quote songs I have in my mind on a particular topic, so I'll start with that here).

One topic that has been with me throughout the year and will probably also continue to be with me is change. A lot of it started with my visit to Mozilla headquarters in Mountain View, CA, in October 2010, actually - I posted about my changing personal priorities back then. And I still remember driving my rental car up to Lake Tahoe, thinking about all those things and listening to the then-just-released Zac Brown Band album "You Get What You Give" and in particular the song "Let It Go", whose lyrics gave me the right mindset for what was I was going through and what 2011 would bring: "Save your strength for things that you can change, forget the ones you can't, you gotta let it go."

Following that, I started 2011 by transferring the vast majority of my responsibilities in SeaMonkey over to other people (we have built up a great team there over the last years, including awesome people like Callek, InvisibleSmiley, etc. - kudos to them to be able to take all that over in their free time) and get the ball rolling on making the project even more sustainable in the future (I hope we'll have news for you on that soon).

Instead, I followed another piece of advice from this song - "When the pony he comes ridin' by, you better sit your sweet ass on it" - and started contracting for Mozilla on the CrashKill team in February, first half-time, finally full-time. With that, my focus changed from SeaMonkey to Firefox and from project management to crash analysis.

For one thing, I ended up growing into that role better than I imagined at first, finding crash analysis more interesting than expected, for the other, this change ended up having more influence on my life than I had imagined. With the need to communicate a lot with different people in this job, from the CrashKill team via the Socorro team that works on the crash-stats server and which I'm coordinating with to various devs, engineering managers or release managers as the need arises in crash analysis.
Unfortunately with me being a "remotie" all communication needs to be online (or via phone) and is stripped down to the essentials needed for the job. Being a very social person, I'm missing the additional nuances that face-to-face communication would bring to the table, and more need for communication as part of the job makes that more obvious to me.
Then, the whole CrashKill team is based in Mountain View, the vast majority of the Socorro team spread across the US, and most engineering or release managers also based in Northern America, so most of that communication as well as all my meetings is happening during US working hours, which from my point of view in Europe is in the evening to night hours, which requires my work time to be mostly at the end of the day. I have been doing work at late hours in the years before, but there was not as much requirement of that before, while now I have to make at least the meetings, and should be available for more conversation on IRC at those times. Making evening appointments becomes quite difficult in that light.
And speaking of requirements, while I could basically completely make my own schedule before, I now should bring in 8 hours of work per day, and with doing that at the end of every day, I need to make all shopping and other private stuff in the afternoon, leaving me all day with "I still have a full work day to deliver today" in mind - until I achieve that and fall into bed. This causes its own share of subconscious stress.
And I'm doing all the work from my own private apartment, not getting out unless I go shopping or take my usual Monday and Tuesday evening off for some Karaoke.
So, I learned that working from home and remotely has its downsides, esp. for the kind of job I'm in there. This is one area I need to work on a lot in 2012 and find solutions that will be connected with another share of change I'm sure.

But not only my role and work life have changed - Mozilla went in a direction I had often spoken for and has changed to a rapid release cycle and started planning for that shortly after I started contracting. I commented in the planning phase and tried to help shape this process and always was convinced it was a good idea, even though we hit more road bumps than expected. I was heavily involved in coordinating to get crash-stats support rapid releases usefully and also laid out publicly how the new process can improve stability.
Mozilla also has revamped its mobile efforts completely - both with a completely new "native UI" version of Firefox for Android, which is in Aurora testing now and with a completely open mobile stack in the form of Boot To Gecko (B2G), a complete "operating system" based on the browser and open web standards (requiring new WebAPIs), which is also coming together piece by piece now.
And next to those changes, we're also working on changing how identity and logins work on the web and changing the current "silo"ed app store model by bringing open concepts for web apps and markets into the fold that easily allow decentralization and users really "owning" their apps.
In the middle of all that, Mozilla has restructured a bit, brought some previously split-off groups back into the common Mozilla fold, hired a lot of new people, lost (as employees but not as community members) a few high-profile ones who were looking for new challenges, worked on the MPL 2.0, founded exciting new initiatives like WebFWD and went stronger on marketing that we are a non-profit - clearly a lot of change happening everywhere, with the mission and the Manifesto standing unchanged and as clear as ever over all of it, though.

All this makes it clear that a lot of change has come in 2011, both to me and Mozilla, and that it's still only the seed for what's to come in the year(s) ahead. The winds of change are still blowing, and I'm excited for what they propel and which interesting experiences they drag in for all of us.

Here's a short summary of Mozilla-related work I've done in weeks 50/2011 (December 12 - 18, 2011):

• Mozilla work / crash-stats:
  Followed the Mozilla datacenter network problems that kept disturbing not just Bugzilla and other services but also our ADU numbers.
  Got my explosiveness reports to use ADU data for 10 and 11, now that it's available to the scripts.
  Triaged more topcrashes and regression crashes.
  Discussed with Sheila, the CrashKill team and with Scoobidiver in bug comments on where we set topcrash keywords or not.
  Participated in discussions on last-minute Firefox 9.0 crash concerns.
  Followed the Socorro team working on Android Fennec support as a separate product.
  Wrote up a proposal of Q1 priorities the CrashKill team has for Socorro.
  More watching of new/rising crashes, caring that bugs are filed where needed.
• Web Apps:
  Following the web apps preview announcement, I decided to put some free time work into some web app experiments myself. I created Mandelbrot and Lantea Maps app prototypes, based on work I had previously and an idea I had been contemplating/investigating for a while. Both still have a number of rough edges and work to be done, but they show promise. I hope to get more out of them when I have time.
• Various Discussions/Topics:
  Repository access levels, Gecko linking RAM size, my tablet problems, etc.

A dozen years of my contributing to Mozilla just became full this Saturday, and that looked like a good point in time to try another new avenue of contribution.
I have been thinking about a web-based OpenStreetMap-using mapping and tracking application for some time, and even created a concept page on my wiki some time ago when a volunteer came up who wanted to contribute something to B2G - and a maps app is still missing there. This weekend, with the web app developer preview being announced, I finally took the chance to dig into it myself. That said, Lantea Maps had its first field test yesterday and I recorded GPX tracks successfully on my N9 and they look reasonable in an application that can show them against a map. I'm looking forward to improving it - and hopefully even someone making it into an app to run and looks really fitting on B2G/Gaia at some point.

Here's a short summary of Mozilla-related work I've done in weeks 49/2011 (December 5 - 11, 2011):

• Mozilla work / crash-stats:
  Communicated with Daniel from metrics because of missing ADU data for my experimental reports.
  Together with other people, tried to keep crash analysis together while Mozilla servers had problems.
  Followed up with the Socorro team on the plan for Android Fennec as a separate product.
  Triaged topcrash bugs with Sheila.
  More watching of new/rising crashes, filing bugs where needed.
• SeaMonkey:
  Created linux64 update snippets for SeaMonkey 2.5.
• Firefox code:
  I did split up my search UI cleanup patches into logical parts, and put the engine manager cleanup in its own bug, also split apart into smaller patches.
• Various Discussions/Topics:
  Tested OpenGL layers and reported its problems, data stores for "native" Fennec, B2G and crashes, Windows update service, my tablet problems, etc.

This week, I was pretty happy not to be in Mozilla IT - they said what they've seen was more or less a "perfect storm" of different problems that made multiple services unavailable for some time and caused a lot of us quite some trouble in work. But still, I don't want to complain, this can happen as we have quite complex systems, hardware problems played into that, and AFAIK we are scaling up some capacities which can lead to bottlenecks and strange problems. The Mozilla IT team did manage to get things under control in a very organized and good manner, and I'm sure some of them didn't sleep well in those days. They put up a monster job and are probably working on making sure such things will not happen again. Their job is somewhat unthankful usually, as you only see when things go wrong, but usually not when things work perfectly, which is how they do most of the time.
We should thank those people at chance we get, as without the infrastructure and services they are managing, our project could not exist and work as well as it does. They are the spine that keeps Mozilla upright and stable on our way into the future of the web. We should be grateful for that.

Here's a short summary of Mozilla-related work I've done in weeks 48/2011 (November 28 - December 4, 2011):

• Mozilla work / crash-stats:
  Contributed to the discussion of where we were with chromehangs and filed the bug to disable the hang monitor for now.
  Got the ball rolling on and stayed in the middle of getting a long-term solution to having different crash-stats reports between Fennec XUL and "native" Android UIs. We have a solution now and are working on implementing it. Thanks to Laura for her game-changing idea.
  Filed a bug to make the Socorro graphs make more sense with zero values.
  Investigated some followup work needed to do for signature summary.
  Did some crash regression bug triage with Sheila and Marcia.
  Continued to watch new/rising crashes closely, filing bugs where needed.
• Themes:
  Did the rest of the porting of Firefox 9 changes to LCARStrek, so both EarlyBlue and LCARStrek 2.6 are basically done and only need some testing before I can send them to AMO.
• Various Discussions/Topics:
  SeaMonkey machines, data stores for "native" Fennec, MXR future, ESR plans, Windows update service, L10n and string typo fixes, Mac OS X 10.5 support, Mozillians "phonebook", tablet hardware problems again, enjoying my Nokia N9, etc.

Just as the last month of 2011 is starting, some discussion and press seems to go out there how this year has been a bad one for Mozilla, but while the media tries to paint the sky black, it looks quite different here inside the project.
It certainly has been and continues to be a year where we face great challenges, but because of that, we have been seeding a few quite great things - delivering new features, performance and memory saving to Firefox users a lot faster than before, a web app framework getting created (at Mozilla Labs) that is open and able to grow to a cross-vendor standard, a web identity mechanism (currently dubbed "BrowserID") coming together that puts users more in control of their logins than before and makes life easier for web developers at the same time, growing a new version of Firefox for Android that is really fast and works smoothly, working on solutions for large deployments (or "Enterprises") again in an own working group, starting work on an open, flexible and fully web-based mobile stack (currently dubbed "B2G") and new standards needed there in the form of the WebAPI project, and much more.
There's surely fallout and a year where a lot of really new stuff is being created usually produces some kind of bottleneck in finished products to show off, so it's no wonder that this one isn't too different. Also, large products in open projects of course have some fallout in terms of discussions, concerns, unclear communications, and slight course corrections along the way. Still, the seeds are sown and great results will grow from them - I'm sure that our awesome community, including paid staff, can take care of those young sprouts in a way that makes them grow into great products.
And then there is the simple fact that the web is a competitive market again, which to a large part is our achievement, and we should be proud of that, but at the same time work hard to stay relevant in that market so we can influence it enough to keep it open for everyone. If the web stays one of openness, opportunity and innovation, we win. If we can help to make the user be more in control of his/her online life, we win. If everyone can put up his/her contents in the way (s)he likes, we win.
Mozilla is the only non-profit in this market and we have an awesome community, so I know we can win. And with your help, we will.

Here's a short summary of Mozilla-related work I've done in weeks 47/2011 (November 21 - 27, 2011):

• Mozilla work / crash-stats:
  Followed the 8.0.1 release that worked around or prevented some crash issues.
  Again provided some feedback on ongoing Socorro work.
  Preliminarily contacted the B2G team about crash reporting.
  Filed bugs for blocking Flash 10.0 and for getting symbols for Android Flash.
  As usual, I also watched new/rising crashes closely and filed bugs for a number of those.
• Themes:
  2.5 versions of EarlyBlue and LCARStrek have been reviewed on AMO now!
  Applied some small fixes to the 2.6 versions I have in development, based on testing and review feedback.
• German L10n:
  Finally applied the central->aurora uplift from last time to the de repo, adding all L10n needed for Aurora in core and SeaMonkey before that.
  Localized all outstanding strings in -central for core and SeaMonkey as well, so that SeaMonkey turned green on all trees.
  Updated DOM Inspector for the German L10n to be in sync with the originals strings.
  Fixed some typos in the German L10n, as well as SeaMonkey addressbook accesskeys once again.
• Various Discussions/Topics:
  L20n patches moving forward, more work on SeaMonkey machines, Fennec native UI project and its fallout, MXR maintenance, ESR plans, getting acquainted with my cool new Nokia N9, etc.

I took a lot of time off work this week, as I'm following US holidays and Thursdays was Thanksgiving. It was good to recharge for a couple of days, given that I started going to bad phrasing again in expressing concerns over some parts of the native UI project for Fennec, but I hope those waters have calmed as much as I myself have now. I'm seeing that the base concerns are shared by others and are seriously being discussed, and that's a good sign. We'll end up with a good offer for Android phone users that represents Mozilla as much as possible given the constraints of the platform, and hopefully will make us credible in the mobile market. If we want complete success on mobile, I still think that B2G is even more compelling, but I think both offers will come out positioned well in their respective space.

I found the following question being raised on German newsgroups, but I guess there's more people interested in this:

Short version:
If you're not using a Mac, switching from 8.0 to 8.0.1 doesn't really help, so we only offer that update for Macs, on all other, 8.0 is just as good, as long as it runs.

Long version:
8.0.1 fixes exactly two things over 8.0:
1) A crash with the newest version of Apple's Java plugin for Mac (the bug is in that new Java version from Apple, but we can work around it with a small patch on our side).
2) On Windows, we're blocking old versions of Roboform that cause 8.0 to crash on startup, we only allow newer versions that don't cause that crash. In case 8.0 is already running, i.e not crashing when starting Firefox, this update doesn't really help anything, and we're avoiding to send people an update when they have no benefit of it, so we don't disturb them. If 8.0 is installed but crashing on startup (because of old Roboform), we don't even get to where an update would be installed. The only thing that helps those people is to manually download 8.0.1 and install it freshly - and actually download a new Roboform version that works with it.
Because of that, 8.0.1 is being offered to all versions (starting with 4.0, incl: 8.0) on Mac OS, but on Windows or Linux only for 4.0-7.0.1, because it doesn't offer any benefit to all the others.

So, in the end, if you're installing a fresh version, 8.0.1 is the right choice, but you don't need to install an update where it's not being offered automatically.