Add support for org-mode files

[sirno]

Describe the bug
Ltex will not start with org files although enabled.

Steps to reproduce

1. install vscode-org-mode extension
2. Add setting:

{
"ltex.enabled": [
    "bibtex",
    "latex",
    "markdown",
    "rsweave",
    "org"
  ]
}

3. Open a file with *.org

Expected behavior
Ltex starts and checks document

Sample document
Any org document.

LTeX configuration

{
"ltex.enabled": [
    "bibtex",
    "latex",
    "markdown",
    "rsweave",
    "org"
  ]
}

[neoscaler]

Can confirm this, seems like a general problem. I am trying to check Asciidoc (identifier 'asciidoc') files with Ltex. Same as described above. Checking MD works.

Settings:

    "ltex.enabled": [
        "bibtex",
        "latex",
        "markdown",
        "rsweave",
        "asciidoc"
    ],

[valentjn]

Maybe the docs of ltex.enabled are a bit misleading/incomplete here. Of course that setting only works for languages that are actually supported by LT_EX. You can't just supply an arbitrary language and expect LT_EX to be able to parse it out-of-the-box. All currently supported languages are already given in the default of ltex.enabled: bibtex, latex, markdown, and rsweave. See also the feature list in the README: “Checks L^AT_EX and Markdown documents.” So, this issue is mostly a feature request, and partly a documentation bug.

To be able to work with org-mode files, we need an org-mode parser written in Java; the parser needs to preserve source code positions. I'll look into this. @neoscaler Please open a separate feature request for Asciidoc.

[valentjn]

The available options of org-mode parsers written in Java and in active development seem to be orgzly/org-java and 200ok-ch/org-parser. The former is licensed under GPLv3 and the latter under AGPLv3, both without linking exceptions (such as the Classpath exception). Unfortunately, these licenses are incompatible with the MPL-2.0 LT_EX is using. I'm not sure how you can possibly license a library (it's a whole different story for standalone programs) under the GPL and not under the LGPL, as that essentially means that only GPL libraries/programs can use that library.

So, until somebody comes up with a different parser we can use (and which preserves source code positions), this feature is going to be out-of-scope.

Mentioning #91 as fixing that would help here as well.

[valentjn]

FWIW, it's possible to check arbitrary files as plain text by adding the code language ID such as "org" to ltex.enabled and activating the extension by running LTeX: Activate Extension. However, without a parser, this will yield some false positives regarding org-mode syntax.

[sirno]

Thanks @valentjn for the effort!

[valentjn]

If the result doesn't have to be 100% accurate (regarding false positives), I'll try writing a parser myself. It doesn't have to be an actual parser; it suffices to strip all the markup away. If the syntax of org-mode is reasonable (which I don't know), this might be within my reach.

[sirno]

Since it is very similar to markdown, it may even be possible to adjust a md parser to make it work. For me, being able to activate the extension to quickly check text in all sorts of places is already great.

[valentjn]

I finished the Org parser. Now I know enough Org to be able to say that Org isn't sufficiently similar to Markdown. Reusing the Markdown parser was not an option, unfortunately.

This is fixed in 10.1.0.