Feature Request: Create tests for syntax highlighter YAML files

[px-64]

When creating or updating syntax files for languages that have complex syntax rules (such as nested formatting types in Markdown, $\LaTeX$, or AsciiDoc), it may be helpful to have syntax examples and their correct parsings to compare changes against.

For example, I am working on updating the Markdown syntax file to fix some style nesting bugs, such as the following bug that incorrectly renders bold text inside an italicized sentence:

In this case, once a solution has been found, it would be quick to test a few strings of italicized and bold text and verify that a fix works. However, when combined with other rules, this case would become more time-consuming to manually verify that fixes have been made and existing cases have not been broken.

Would it be possible to implement some sort of parsing validation system to ensure code is broken down into the correct "pieces"?

One related example that comes to mind is CommonMark's spec tests that include pairs of Markdown source strings and their proper HTML renderings. I'm not sure how Micro handles parsed code or if a system like this would be possible, but I wanted to bring up the idea in case others may find it useful as well.

[px-64]

Another idea might be including example code files to accompany each syntax file. For example, a file named python3-example.py to accompany python3.yaml in the runtime/syntax/ directory, where the example file contains samples of all the syntax rules provided by python3.yaml.

I have found myself repeatedly creating short sample code files for syntax highlighting tests, so this would save development time and provide a common base for everyone to compare across revisions.

[Andriamanitra]

I think the actual issue is (once again) that the syntax highlighting system is too limited to properly handle nesting so you're forced into writing finicky hacks to work around it. I've expanded on this before in #3705 (comment).

A testing system could help a little bit, but if it takes significant effort to implement we may want to consider replacing the syntax highlighting system with something more robust first (eg. zyedidia/flare).

One related example that comes to mind is CommonMark's spec tests that include pairs of Markdown source strings and their proper HTML renderings. I'm not sure how Micro handles parsed code or if a system like this would be possible, but I wanted to bring up the idea in case others may find it useful as well.

Sublime Text also has a really convenient syntax testing system that could serve as inspiration.