support for reStructuredText

[jaguarfi]

I was wondering whether or not it is possible to support reStructuredText.

[valentjn]

@jaguarfi Thanks for the suggestion. For this, we'd need a Java parser for reStructuredText. The only one that I can find is JRst (see this answer on Stack Overflow and this fork on GitHub), but that hasn't been maintained in ten years. More convoluted solutions would involve calling external programs such as Pandoc, but that would be pretty slow in an editing context.

So unless you know a parser or a better way, the answer to your question is "probably, but it would be pretty difficult."

[valentjn]

Closing as this is currently out of scope.

[valentjn]

There's another constraint: We need not only a parser, but one that outputs an AST (abstract syntax tree) with corresponding source code positions. Otherwise, we can't map positions of the LanguageTool error messages back to positions in the original text. Both JRst and Pandoc (see #91) fail to satisfy this constraint.

[valentjn]

I'm currently writing a very basic converter/filter myself. It basically strips off all the markup to reduce the number of false positives LanguageTool produces. It's not a parser (reStructuredText is very complex), but I guess it's better than nothing.

Don't get your hopes too high; for example, tables are currently ignored as a whole. Stuff like multi-line cells would lead to non-monotonous text position mapping (text of one multi-line cell continues "after" the text of another multi-line cell), and that would screw up the locations of all the diagnostic messages.

[valentjn]

This should be fixed in the next release.

[valentjn]

Fix released in 10.1.0.