Tag Archives: firefox

The Joy of Coding (Episode 7): Code review, and a Regression

In this episode, I started with some code review. I was reviewing a patch to make the Findbar (particularly, the Find As You Type feature) e10s-friendly.

With that review out of the way, I had to swap a bunch of information about the plugin crash UI for e10s in my head – and in particular, some non-determinism that we have to handle. I explained that stuff (and hopefully didn’t spend too much time on it).

Then, I showed how far I’d gotten with the plugin crash UI for e10s. I was able to submit a crash report, but I found I wasn’t able to type into the comment text area.

After a while, I noticed that I couldn’t type into the comment text area on Nightly, even without my patch. And then I reproduced it in Aurora. And then in Beta. Luckily, I couldn’t reproduce it in Release – but with Beta transitioning to Release in only a few days, I didn’t have a lot of time to get a bug on file to shine some light on it.

Luckily, our brilliant Steven Michaud was on the case, and has just landed a patch to fix this. Talk about fast work!

Episode Agenda

References:
Bug 1133981 – [e10s] Stop sending unsafe CPOWs after the findbar has been closed in a remote browser

Bug 1110887 – With e10s, plugin crash submit UI is broken – Notes

Bug 1147521 – Cannot type into comment area of plugin crash UI

The Joy of Coding (Episode 6): Plugins!

In this episode, I took the feedback of my audience, and did a bit of code review, but also a little bit of work on a bug. Specifically, I was figuring out the relationship between NPAPI plugins and Gecko Media Plugins, and how to crash the latter type (which is necessary for me in order to work on the crash report submission UI).

A minor goof – for the first few minutes, I forgot to switch my camera to my desktop, so you get prolonged exposure to my mug as I figure out how I’m going to review a patch. I eventually figured it out though. Phew!

Episode Agenda

References:
Bug 1134222 – [e10s] “Save Link As…”/”Bookmark This Link” in remote browser causes unsafe CPOW usage warning

Bug 1110887 – With e10s, plugin crash submit UI is broken – Notes

The Joy of Coding (Episode 5): Much Code Review

In this fifth episode, I didn’t work on any bugs. Instead, I did a bunch of code review. Ever wanted to know what a Firefox engineer does to review a patch? If so, then this episode is for you!

Episode Agenda

References:
Bug 1140898 – [e10s] “View” > “Switch Page Direction” doesn’t work in e10s

Bug 1140878 – [e10s] “Switch Page Direction” in remote browser causes unsafe CPOW usage warnings

Bug 1066531 – [e10s] Switching tabs can result in old content being displayed for a split second after the tab bar is updated

The Joy of Coding (Episode 4)

The fourth episode is up! Richard Milewski and I found the right settings to get OBS working properly on my machine, so this weeks episode is super-readable! If you’ve been annoyed with the poor resolution for past episodes, rejoice!

In this fourth episode, I solve a few things – I clean up a busted rebase, I figure out how I’d accidentally broken Linux printing, I think through a patch to make sure it does what I need it to do, and I review some code!

Episode Agenda

References:
Bug 1136855 – Print settings are not saved from print job to print job
Notes

Bug 1088070 – Instantiate print settings from the content process instead of the parent
Notes

Bug 1090448 – Make e10s printing work on Linux
Notes

Bug 1133577 – [e10s] “Open Link in New Tab” in remote browser causes unsafe CPOW usage warning
Notes

Bug 1133981 – [e10s] Stop sending unsafe CPOWs after the findbar has been closed in a remote browser
Notes

The Joy of Coding (Episode 3)

The third episode is up! My machine was a little sluggish this time, since I had OBS chugging in the background attempting to do a hi-res screen recording simultaneously.

Richard Milewski and I are going to try an experiment where I try to stream with OBS next week, which should result in a much higher-resolution stream. We’re also thinking about having recording occur on a separate machine, so that it doesn’t bog me down while I’m working. Hopefully we’ll have that set up for next week.

So this third episode was pretty interesting. Probably the most interesting part was when I discovered in the last quarter that I’d accidentally shipped a regression in Firefox 36. Luckily, I’ve got a patch that fixes the problem that has been approved for uplift to Aurora and Beta. A point release is also planned for 36, so I’ve got approval to get the fix in there too. \o/

Here are the notes for the bug I was working on. The review feedback from karlt is in this bug, since I kinda screwed up where I posted the review request with MozReview.