🐛 Fix yaml number rendering

[homersimpsons]

Fixes #177

I copied the yaml.json file from https://github.com/scrivo/highlight.php/blob/d1c6b0956a2b0d3efc137e4d8c5bf5c5e220bac1/src/Highlight/languages/yaml.json#L84 and updated the "number" regex to add _ where I think the symfony parser allow them:

- "begin": "(-?)(\\b0[xX][a-fA-F0-9]+|(\\b\\d+(\\.\\d*)?|\\.\\d+)([eE][-+]?\\d+)?)\\b"
+ "begin": "(-?)(\\b0[xX][a-fA-F0-9_]+|(\\b\\d[\\d_]*(\\.[\\d_]*)?|\\.\\d[\\d_]*)([eE][-+]?\\d[\\d_]*)?)\\b"

There may be additional places where the symfony parser allow the _ for numbers.

I also updated the tests to add the example from the documentation.

[wouterj]

Thanks for the fix!

Overwriting the syntax file in this project is indeed the most sensible option (see also my more lengthy response on this matter: #176 (comment)).

The highlight.js syntax file has support for some more number types (especially timestamps, datetimes, etc.). Maybe we can also add those regexes (as a separate number item)? (given the Yaml component supports date strings)

[homersimpsons]

The highlight.js syntax file has support for some more number types (especially timestamps, datetimes, etc.). Maybe we can also add those regexes (as a separate number item)? (given the Yaml component supports date strings)

I can add those, but are they supported by symfony/yaml's parser? I saw we were parsing dates but not sure about the exact syntax supported.

[homersimpsons]

Just tested it. Their regex is pretty eager:

• It allows any year of 4 digit, I sticked with 2024 (I may have found a symfony bug without this constraint)
• It allows having a fractional time part without having a time

var_dump(Yaml::parse(<<<YAML
v0: 2024-12-07 23:44:39.35677451141299410 	    	    		   		 	 	 	  	 	    		   			  		 			 			 				  	 		
v1: 2024-02-3.671	-11
v2: 2024-82T3:66:07		+36
v3: 2024-07-05t6:39:20	 	   		 	        	
v4: 2024-01-30.000				 	 	  	  				 							 	 	 	  	 	 			      			 				  	 	     	   	   	 		    	
v5: 2024-09-03 			 					  	   			    	   	  	    		  	
YAML
, Yaml::PARSE_DATETIME));

will give the following, we can see that symfony won't handle all possible "dates":

array(6) {
  'v0' =>
  class DateTimeImmutable#4 (3) {
    public $date =>
    string(26) "2024-12-07 23:44:39.356774"
    public $timezone_type =>
    int(3)
    public $timezone =>
    string(3) "UTC"
  }
  'v1' =>
  string(17) "2024-02-3.671     -11"
  'v2' =>
  string(20) "2024-82T3:66:07           +36"
  'v3' =>
  class DateTimeImmutable#5 (3) {
    public $date =>
    string(26) "2024-07-05 06:39:20.000000"
    public $timezone_type =>
    int(3)
    public $timezone =>
    string(3) "UTC"
  }
  'v4' =>
  string(14) "2024-01-30.000"
  'v5' =>
  class DateTimeImmutable#6 (3) {
    public $date =>
    string(26) "2024-09-03 00:00:00.000000"
    public $timezone_type =>
    int(3)
    public $timezone =>
    string(3) "UTC"
  }
}

@wouterj Should I still use their regex or use the one from symfony/yaml?

[homersimpsons]

@wouterj I used the regex from symfony/yaml. I had to clean it up so it works in JS (removed the capture groups), and I also removed the non-capture groups that are unused.

The failure is probably not linked to my changes. I rebased on main, but I still have the error.

[javiereguiluz]

Nice work! Thanks @homersimpsons

[wouterj]

Nice, this is looking great!

Failure is indeed not related to this PR.

[javiereguiluz]

Merged! Thanks