User Agent: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/55.0.2883.87 Safari/537.36 Steps to reproduce: Attempt to parse dom/webidl/Window.webidl Actual results: Failed to parse invalid identifier "__content" at https://github.com/mozilla/gecko-dev/blob/60785fce6d52289c2e147364863f9d42207c5f91/dom/webidl/Window.webidl#L379 Expected results: Parse success (well-formed WebIDL file); identifier should match https://heycam.github.io/webidl/#prod-identifier. To be honest, I'm not sure what code depends on this identifier name, and therefore how straightforward it would be to fix.
I don't believe that parsing our WebIDL files with any parser other than our own is a goal.
The context here is that Mark is trying to ingest all of Blink's IDL and the specs it links to into a tool that identifies differences, the kind of thing we all love looking for manually. It would be useful to also compare against Gecko and WebKit, so fewer differences in the dialects is better. But I see there are also ifdefs in there, so some workarounds will be needed, and allowing leading underscores seems like a simple case to deal with.