Karl Dubost: [worklog] Outreach is hard, Webkit aliasing big progress

Пятница, 05 Февраля 2016 г. 11:15 + в цитатник

Tunes of the week: Earth, Wind and Fire. Maurice White, the founder, died at 74.

WebCompat Bugs

A Web compat bug which can be fixed by adding a CSS property, but which led also to the opening of a bug for Gecko about max-width.
The sad comment of the day was in that bug about fonts on Nokia site: Forgive me to say this, but... Sadly, FF's market has been dropping and keeps dropping, losing to Chrome. At this point "works in Chrome, used to work in FF" should mean "fix it in FF asap". An end user only sees FF doesn't work, Chrome does. They don't care about the technical reasons.
GetPocket displays an interesting issue on Firefox for Android. The bug had been opened a little while ago, but I took time to create an account to investigate. It sounded familiar to me. It may be related to Bug 1244192
Two weeks ago, there was an issue about a BOM breaking a templating system in Dojo. Henri Sivonen looked at it and indeed found a bug in XHR implementation in Firefox.
Sometimes another team of Mozilla will ask help from Webcompat team for contacting site owners to fix an issue on their Web site which hinders the user experience on Firefox. Let's go through some tips to maximize the chances of getting results when we outreach.

WebKit aliasing

When looking for usage of -webkit-mask-*, I remembered that Google Image was a big offender. So I tested again this morning and… delight! They now use SVG. So now, I need to test extensively Google search and check if they can just send us the version they send to Chrome.
Testing Google Calendar again on Gecko with Chrome user agent to see how far we are to receive a better user experience. We can't really yet ask Google to send us the same thing they send to Chrome. A couple of glitches here and there. But we are very close. The better would be for Google to fix their CSS, specifically to make flexbox and gradients standard-compatible.
The code for the max-width issue (not a bug but implementation differences due to an undefined scenario in the CSS specification) is being worked on by David Baron and reviewed by Daniel Holbert. And this makes me happy, it should solve a lot of the Webcompat bugs reports. Look at the list of SeeAlso in that bug.

Webcompat Life and Working with Developer Tools

Changing preferences all the time through "about:config" is multiple step. I liked in Opera Presto how you could link to a specific preference, so I filed a bug for Firefox. RESOLVED. It exists: about:config?filter=webkit and it's bookmark-able.
Bug 1245365 - searching attribute values through CSS selectors override the search terms
A discussion has been started on improving the Responsive Design Mode of Firefox Developer Tools. I suggested a (too big) list of features that would make my life easier.

Firefox OS Bugs to Firefox Android Bugs

Web Compatibility Mozilla employees reduced their support for solving Firefox OS bugs to its minimum. The community is welcome to continue to work on them. But some of these bugs have still an impact on Firefox Android. One good example of this is Bug 959137. Let's come up with a process to deal with those.
Another last week todo. I have been closing a lot of old bugs (around 600 in a couple of days) in Firefox OS and Firefox Android in Tech Evangelism product. The reasons for closing them are mostly:
- the site doesn't exist anymore. (This goes into my list of Web Compatibility axioms: "Wait long enough, every bug disappears.")
- the site fixed the initial issue
- the layout.css.prefixes.webkit; true fixes it (see Bug 1213126)
- the site has moved to a responsive design

Bug 812899 - absolutely positioned element should be vertically centered even if the height is bigger than that of the containing block

This bug was simple at the beginning, but when providing the fix, it broke other tests. It's normal. Boris explained which parts of the code was impacted. But I don't feel I'm good enough yet for touching this. Or it would require patience and step by step guidance. It could be interesting though. I have the feeling I have too much on my plate right now. So a bug to take over!

Testing Google Search On Gecko With Different UA Strings

So last week, I gave myself a Todo "testing Google search properties and see if we can find a version which is working better on Firefox Android than the current default version sent by Google. Maybe testing with Chrome UA and Iphone UA." My preliminary tests sound pretty good.

Reading List

Frameworks vs libraries (or: process shifts at Mozilla).
Mozilla's internal dialog on HTTPS by Dave Winner, who's not happy about the move to HTTPS in general. He is worried and it's understandable because the announcements on "burning down everything non HTTP" were not followed-up by regular status on a milder position. Google has also a worrying Marking HTTP As Non-Secure page. That said Mozilla bug had no activity for more than one year, which in bug tracker standards may mean "wait long enough it will die by itself". The push from HTTP to HTTPS should not be valid by itself or more exactly, we should not equate a technology with user security and choices. Maybe there are better and more accessible ways for both Web developers and users to create the same secure environment. Let's Encrypt is a very interesting initiative, but has still too many implications down the road on the development work.
Discussion in the CSS Working Group about Web Compatibility for 0 vs 0deg for angles. The resolution: "Angles can drop unit when value is 0"
An email conversation summary visualization. Mail is something I like very much. But I manage it in a way which is a bit different than many people. I also think that the email apps on Mobile are just dumb and it might be a part of the hate that people have for mails. If you send notifications mail to your email address, give yourself a gift… use filtering.
mike published about A quiz about ES2015 block-scoped function declarations (in a with block statement)