On 02/16/2017 07:20 PM, Benjamin Smedberg wrote:
It has happened repeatedly over the past two years that we discovered critical issues that affect websites only after we ship to release.
I don't think removing -moz-appearance even has the potential of being "critical". All that happens is that you get native styling instead (at worst). There shouldn't be any loss of function. I don't think an average web user will even notice it. In the unlikely event that we do want to roll this back, all we need to do is toggle the preference values, which I believe we can do even in release builds nowadays by pushing a restart-less system add-on.
It is not hard nowadays to add metrics (telemetry) to measure web feature usage.
Getting the data we would need in this case is actually a lot harder than you may think. It's probably easy to measure "does this web page have a style rule that declares a valid -moz-appearance value?", but that's useless in this case, since we really need to know if there's an 'appearance/-webkit-appearance' declaration that "wins" in this style rule, or in some other rule. Or if there's a 'border' declaration etc that also turns off native styling in some cases, or if the element is a <div> or something that has no support for native styling to begin with, etc. Adding that telemetry machinery to the style system seems rather tricky to me. And given the low risk with removing -moz-appearance - doesn't seem worth implementing to me. (I think our goal is to remove all -moz-prefixed CSS properties eventually anyway, at which point this machinery would have little value.)
In this case, I understand the advantage of shipping CSS 'appearance'. I'm less sure about what it would cost us to keep supporting -moz-appearance: none, perhaps indefinitely.
It seems you haven't considered that *removing* a -moz-prefixed property could actually *improve* web compatibility. We actually have some evidence of that for -moz-appearance:none in bug 1328474. I suspect this is the case for -moz-appearance. If the web author also wrote -webkit-appearance/appearance:none (as they should) then there is no change, and if they didn't they might not have intended to have 'none' styling in the first place. It's just that they never tested it in Firefox. I get that impression after studying the usage of these properties in some github projects. AFAICT, IE11 (on a Win7 desktop) has no support for 'appearance' at all (or any prefixed variant thereof), and Edge specifically added support only for '-webkit-appearance:none'. And I assume the web still works in those browsers. /Mats _______________________________________________ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
