Indistinguishable from Jesse

I recently landed a fix so that when Firefox crashes or hangs on Tinderbox, the error summary shows which test was running.

As we add more tests and platforms to Tinderbox, it's increasingly important for developers to be able to identify each test failure quickly and accurately. Good error summaries make assisted starring of random oranges possible, which greatly reduces the pain induced by intermittent failures. Good error summaries also make it possible to track which failures are most frequent, and therefore concentrate on fixing the most important ones.

Error summaries for crashes could be further improved by showing what kind of process crashed and a signature based on the stack trace.

I'd also like to see better error summaries for memory leaks, compiler errors, python errors, and failures in smaller build steps.

If you see other error summaries on Tinderbox that could be improved, please file bugs. It's an easy way to help Mozilla scale across branches, and it's cheaper than cloning philor.

If you want to stay safe from phishing and other forms of online fraud you need at least a basic understanding of a bewildering array of technologies—URLs, paths, domains, subdomains, ports, DNS, SSL as well as fundamental concepts like browsers, web sites and web servers. Misunderstand any of those concepts and you’ll be an easy target for even the most basic phishing attempts. It almost makes me uncomfortable encouraging regular people to use the web because I know they’ll be at massive risk to online fraud.

- Simon Willison

My latest post on the Mozilla Security Blog:

Fixing security holes without introducing new bugs

Testers can now download debug builds of Firefox rather than going trough the complicated process of building Firefox themselves.

Debug builds have extra code called assertions to inform us when assumptions or surrounding code are incorrect. For example, assertions can reliably tell us when third-party code makes common threading mistakes, even though the crashes from those mistakes appear random. I find about a third of my bugs by noticing assertion failures, and it usually isn't the assertion that's incorrect.

In several bug reports from users who were hitting frequent crashes at random, I suggested that the users compile their own debug builds, hoping that assertion failure messages would lead us to the cause of the crash. Few of the users succeeded, since the compilation tools and instructions were written with programmers in mind. Now that pre-made debug builds are available, we may be able to make more progress on these kinds of crash bug reports.

Big thanks to the release engineering team for making these builds available!

Reed Loden set up bugzil.la as a redirect for Bugzilla bugs and searches. Examples:

• http://bugzil.la/95849
• http://bugzil.la/468562#c12
• http://bugzil.la/firebug+crash
• http://bugzil.la/ALL+flag:in-litmus?

I put together a demo of incrementing numbers in URLs by scrolling.

Hopefully it's not considered a bug that I can use rangeOffset for a text node "inside" a textbox ;) Trying to access rangeParent throws a security exception.

This could be used by an extension like URL Flipper as a shortcut for changing specific numbers in URLs.

Getting Things Done recommends spending an hour or two on a "weekly review" every Friday afternoon. This worked for me at first, but after a month, I found myself procrastinating more often than actually doing a weekly review. Ironically, another concept from the same book helped me understand some of the reasons.

A major problem with my old weekly review checklist was that it included things I had to do in different contexts. For example, I can only check snail mail at home, but I can only check voice mail when I'm not at home. Worse, performing a weekly review seemed to require concentrating on one thing for an hour, and I'd rather spend that concentration on programming.

I decided to split my weekly review into a list of weekly sanity reminders, with each reminder on appropriate day of the week. I chose Monday for reviewing actions assigned to specific contexts, since the contexts "office" and "business-hours" usually become available on Monday. But I review coding projects later in the week, when distractions have died down.

For many of the smaller parts of my old weekly review, such as making sure my iPhone's voicemail list is empty, I chose a day of the week on which a mid-day meeting breaks up my day.

Splitting my weekly review into multiple scheduled items allowed me to put some reminders on the days when they do the most good, and others on the days when distractions do the least harm.

Like the author of GTD, I found that doing each of these things once a week helps me focus on the right projects and avoid some forms of procrastination. They just don't all need to be done on the same day.

Yesterday, I subscribed to the hg feed for reftest.js in Google Reader. Today, Google Reader informed me of a change Mats made. It works!