The HTML parser should handle inline tags independently of their ids

[ymeine]

Currently the HTML parser (either on its own or embedded in the AT parser) handle inline tags in a very restrictive way. When encountering the beginning of a tag, there is a specific rule that checks if the following id is among a hard-coded set of inline ids to determine if the current tag can be inline, otherwise the id is parsed generically - it can be anything.

First, the list shouldn't be hard-coded: using a predicate in the grammar to hack the rule and taking the set of ids from an external input would be far better.

Then, it would be great, even if we agree this is difficult, to be able to parse without ambiguity inline tags with any id.