[CssSelector] Fix escape patterns #48771
Conversation
This comment was marked as resolved.
This comment was marked as resolved.
| @@ -49,22 +49,22 @@ public function __construct() | |||
| $this->identifierPattern = '-?(?:'.$this->nmStartPattern.')(?:'.$this->nmCharPattern.')*'; | |||
| $this->hashPattern = '#((?:'.$this->nmCharPattern.')+)'; | |||
| $this->numberPattern = '[+-]?(?:[0-9]*\.[0-9]+|[0-9]+)'; | |||
| $this->quotedStringPattern = '([^\n\r\f%s]|'.$this->stringEscapePattern.')*'; | |||
| $this->quotedStringPattern = '([^\n\r\f\\\\%s]|'.$this->stringEscapePattern.')*'; | |||
There was a problem hiding this comment.
https://github.com/scrapy/cssselect/blob/ddd9784977fca6e7da4439d837e4f510f1f10638/cssselect/parser.py#L926 (We are missing \\ before the quote)
| } | ||
|
|
||
| public function getNewLineEscapePattern(): string | ||
| { | ||
| return '~^'.$this->newLineEscapePattern.'~'; | ||
| return '~'.$this->newLineEscapePattern.'~'; |
There was a problem hiding this comment.
Check the 3 escape"_sub_* "functions" and how they are used. All occurrences are replaced.
| ['#test\:colon', ['Hash[Element[*]#test:colon]']], | ||
| [".a\xc1b", ["Class[Element[*].a\xc1b]"]], | ||
| // unicode escape: \22 == " | ||
| ['*[aval="\'\22\'"]', ['Attribute[Element[*][aval = \'\'"\'\']]']], |
There was a problem hiding this comment.
I borrowed those tests from https://github.com/scrapy/cssselect/blob/ddd9784977fca6e7da4439d837e4f510f1f10638/tests/test_cssselect.py#L550. They highlighted the "quotedStringPattern" problem.
|
@fancyweb if you have some availability and are keen on checking tat python code, you might want to work on porting the missing features to Symfony 6.3 😄 I create a few issues about them (but I'm quite sure the list is still incomplete for now) |
|
Thank you @fancyweb. |
I rechecked the original Python source code we "borrowed" (https://github.com/scrapy/cssselect/blob/master/cssselect/parser.py) and fixed some issues.