What could be wrong?
Another first: I’ve never before debugged a script that depends on what font a string is rendered with. Twitter somehow wrote one.
It’s good to know that the Chrome team is willing to break Mobile GMail to improve cross-browser interoperability and spec compliance. I’m sure it helps that both products come from the same company, of course..
I finally figured out the rest of the story (from JS point of view) in the bug the values of a Google 3d graph can’t be modified and it’s not too interesting - Firefox fires an extra and unwanted blur event at an input type=number. The real reason I’m mentioning this bug is: how often do we get to debug an issue on the Google search results page?. For me it’s a rare occasion, though YMMV.
(“Bug safari”: linking to oddities in open bug trackers, mainly Mozilla’s and webcompat.com - meant to be a regular feature. We’ll see about the frequency..)
Save web developers from pre-, post-, instead-, and uturn-events - Bugzilla rant about how slowly we’ve moved the discussion forward regarding the finer details of clipboard event support. Yours truly is not not guilty of failing to hurry this up..
If you pretend to support -webkit- prefixed CSS, you must also relax your cross-origin security policy - obviously..