Syntax highlighting for ucode

[eric-j-ason]

I'd like to get support for ucode in syntax highlighters such as Chroma (which is used by Gitea) and Pygments, perhaps by starting with their Javascript definitions and modifying them. Is there a specification of what keywords and other token types exist in ucode?

For reference:

• Javascript in Chroma
• Javascript in Pygments

[eric-j-ason]

And in nano, of course, included in OpenWrt!

[eric-j-ason]

A pull request for introducing ucode in Chroma has been created.

[jow-]

I added a pull request to include syntax highlighting in OpenWrt's nano package:
openwrt/packages#22623

[jow-]

@eric-j-ason - I suggest to use the nano highlighting I submitted in the other PR as authoritative source for the supported syntax keywords.

See also https://github.com/jow-/ucode/blob/master/lexer.c#L47

There's some special cases like as and from which are syntax elements in import statements but otherwise no reserved identifier (see recently fixed bug at 1468cc4).

Another complication is that ucode may run in template mode where by default the file source is uninterpreted plain text and actual ucode logic only appears in {{ ... }} or {% ... %} expression blocks. (See the grey text in screenshot https://user-images.githubusercontent.com/2528802/281428096-478dddef-fe44-4c01-ac3e-63fa6c605f6e.png for illustration)

[eric-j-ason]

@eric-j-ason - I suggest to use the nano highlighting I submitted in the other PR as authoritative source for the supported syntax keywords.

See also https://github.com/jow-/ucode/blob/master/lexer.c#L47

Yes, good idea. Those lines in lexer.c was actually what I used to figure out the list of keywords.

There's some special cases like as and from which are syntax elements in import statements but otherwise no reserved identifier (see recently fixed bug at 1468cc4).

Oh, I see. Does it work the same in JavaScript, or is it different? In what I did for Chroma, I just took JavaScript as the starting point, and adjusted the list of keywords. I'm afraid I don't know how to do anything more complicated in Chroma or any other highlighter.

Another complication is that ucode may run in template mode where by default the file source is uninterpreted plain text and actual ucode logic only appears in {{ ... }} or {% ... %} expression blocks. (See the grey text in screenshot https://user-images.githubusercontent.com/2528802/281428096-478dddef-fe44-4c01-ac3e-63fa6c605f6e.png for illustration)

Yes, also this I didn't know how to do, so I had to leave it out. It's nice that you did it for nano.

I haven't quite figured out how highlighting works on GitHub. Perhaps they use a complete parser, enabling additional features such as getting the definition of a variable by clicking it. This is what they say: "We use Linguist to perform language detection and to select third-party grammars for syntax highlighting. You can find out which keywords are valid in the languages YAML file."

[eric-j-ason]

I added a pull request to include syntax highlighting in OpenWrt's nano package: openwrt/packages#22623

I have noticed that these highlighting rules assume that the file is a template, and breaks on a pure ucode file.

[stokito]

For most developers like one evening contributors to Luci the "ucode" is an additional confusion. It would be very helpful to use the .js as extension to make IDE and linters happy. Some syntax discrepancies aren't critical but if we can avoid them that would be great.

There is a similar project https://github.com/cesanta/mjs with a rule:

Any valid mJS code is a valid ES6 code.
Any valid ES6 code is not necessarily a valid mJS code.

The ucode has own global functions that will be highlighted by IDE in red. We may workaround this by some import of ucode.js that will have declarations of all functions.

The main problem is that the .uc files can be handled by an interpreter and used as cgi-bin.
Maybe this can be supported with a double extension .uc.js.