Our Google Docs efforts have stalled for a number of reasons. Besides
cross-origin wrappers, there aren’t any big missing pieces that we can
point at as being a clear problem in our current Servo builds. Efforts
like the Chromium UseCounter data and Manish’s feature-detector user
script show us a list of features that we don’t implement yet that
Google Docs may care about, but it’s very difficult to tell whether
we’re receiving fallback behaviour that works around them.
I propose a new focus for our efforts rather than guessing which DOM
features we should implement:
1) Prioritize fixing assertions and crashes that are triggered by Google
Docs so we can actually experiment with it.
2) Un-minify JS/CSS/HTML that Servo encounters before evaluating it, and
store local copies for later investigation. [0]
3) Build more tools and features that expose Google Docs’ behaviour. For
example, if retrieving JS backtraces from Rust code would be useful,
that’s something we can enable without much effort [1].
One specific piece of investigation is the missing text editing. If
you’re lucky enough to reach a blank document, a text cursor appears but
pressing keys does not do anything. We need to figure out why that’s the
case, and we also need to figure out how to do that investigation
efficiently. I would love to compile a wiki page with a list of
strategies for investigating web compatibility problems.
Cheers,
Josh
[0] https://github.com/servo/servo/issues/14824 (unminify all JS for
more readable exceptions, and store local copies for later investigation)
[1] https://github.com/servo/servo/issues/14987
_______________________________________________
dev-servo mailing list
dev-servo@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-servo
