Nano Highlighter initial implementation

[jspenguin2017]

Update: Initial implementation finished. Please open new issues for other problems.

[jspenguin2017]

Nano's current profile:

[jspenguin2017]

uBO's profile:

[jspenguin2017]

I might have to use a negative look-ahead, but I'm a bit worried about performance as it will have to scan the text twice.

[jspenguin2017]

Negated domains are not properly handled @ameshkov Shouldn't that be | instead of ,?

Test cases:

||coinhive.com^$third-party,domain=~cnhv.co,~coinhive.com

Another one:

[ameshkov]

Your syntax highlighter is awesome!

[jspenguin2017]

Old opening:

RegExp rules are not properly handled when the RegExp contains $.

Test cases:

/\.download\/[0-9]{2,9}\/$/$script,stylesheet,third-party,xmlhttprequest
/\.faith\/[0-9]{2,9}\/$/$script,stylesheet,third-party,xmlhttprequest
/\.link\/[0-9]{2,9}\/$/$script,stylesheet,third-party,xmlhttprequest
/\.loan\/[0-9]{2,9}\/$/$script,stylesheet,third-party,xmlhttprequest
/\.lol\/[0-9]{2,9}\/$/$script,stylesheet,third-party,xmlhttprequest
/\.men\/[0-9]{2,9}\/$/$script,stylesheet,third-party,xmlhttprequest
/\.online\/[0-9]{2,9}\/$/$script,stylesheet,third-party,xmlhttprequest
/\.party\/[0-9]{2,9}\/$/$script,stylesheet,third-party,xmlhttprequest
/\.racing\/[0-9]{2,9}\/$/$script,stylesheet,third-party,xmlhttprequest

/http://[a-zA-Z0-9]+\.[a-z]+\/.*(?:[!"#$%&'()*+,:;<=>?@/\^_`{|}~-]).*[a-zA-Z0-9]+/$script,third-party,domain=keezmovies.com|redtube.com|tube8.com|tube8.es|tube8.fr|www.pornhub.com|youporn.com

---

Also, these rules are in EasyList but Nano Highlighter marks them as invalid: Those options are not accepted by Nano.

||ngohq.com/images/ad.jpg$~collapse
@@||oload.tv^$genericblock,generichide
@@||openload.co^$genericblock,generichide
@@||streamango.com^$genericblock,generichide
@@||spiegel.de^$genericblock,generichide

---

, in script:inject arguments should be treated as operator just like other ,.
That's how it's parsed.

---

Handle folding of !#if preprocessor directive. AdguardTeam/AdguardBrowserExtension#917
Ace's doc on this is pretty non-existent... Might have to read deep into their code.

---

Something is not right with Traditional Chinese locale:
#40 (comment)

[jspenguin2017]

This is fixed:

RegExp rule highlighting is not properly fixed:

Test rules:

/test$/

[ameshkov]

JFYI,

I've just figured that Ace provides an easy conversion from TextMate language to Ace rules:
https://github.com/ajaxorg/ace/wiki/Importing-.tmtheme-and-.tmlanguage-Files-into-Ace

The thing is that a lot of filters maintainers use external text editors (besides the filters editor inside of a blocker), and most of the popular text editors (Sublime, Atom, VS Code) support TextMate language extensions, so it might useful to use a TM language as the main source for the rules.

I'm planning to experiment with this during the holidays, but anyway, I guess this might be useful to you if you decide to do the same.

Here's a simple VS code language plugin:
https://github.com/ameshkov/VscodeAdblockSyntax

TM language file:
https://github.com/ameshkov/VscodeAdblockSyntax/blob/master/syntaxes/adblock.tmLanguage.json

It's rather simple (as I've just made it) and can only highlight classic elemhide rules:

[jspenguin2017]

Filter rules are different than other programming language, it's not something you'd code 100 lines and test them together, I find myself code them in Nano's IDE, test them right away, then paste them into my filter list. Nano will validate and lint it right away with the existing filter parser, which can be a bit difficult to implement in Atom plugin.

At first, I tried to make a plugin for Atom, but then I couldn't find the proper documentation. Maybe I was missing something...

[ameshkov]

Filter rules are different than other programming language

Take a look at all the languages VS code supports, some of them are very different from imperative languages we are used to:
https://github.com/Microsoft/vscode/tree/master/extensions

Nano will validate and lint it right away with the existing filter parser, which can be a bit difficult to implement in Atom plugin.

I suppose, that validation and syntax highlighting are different types of tasks indeed. Every editor provides its own way for that. For instance, VS code does it via "language server" extensions. Some obvious mistakes can be highlighted using regular expressions, though.

Regarding Atom, it also supports the TextMate theme conversion:
http://flight-manual.atom.io/hacking-atom/sections/converting-from-textmate/

[jspenguin2017]

Unless you guys have constructive comments, just don't post it. It's not getting anywhere. I can't change gorhill's decisions, no matter I agree with him or not.

[jspenguin2017]

I don't mind civil discussions, but the discussion is getting not that civil and off topic from development of Nano.

[jspenguin2017]

Added preprocessor highlighting: AdguardTeam/AdguardBrowserExtension#917

[jspenguin2017]

I'll leave this thread open for quick comments, but please report more complicated problems in new issues.