[DomCrawler] Use the native HTML5 parser on PHP 8.4 #61475
Conversation
90067d6 to
12d3f8c
Compare
f963618 to
a6a0033
Compare
a6a0033 to
bd7fb51
Compare
…nks to the native DOM parser (nicolas-grekas) This PR was merged into the 8.0 branch. Discussion ---------- [DomCrawler] Always parse according to HTML5 rules thanks to the native DOM parser | Q | A | ------------- | --- | Branch? | 8.0 | Bug fix? | no | New feature? | yes | Deprecations? | no | Issues | - | License | MIT Follows #61475 Commits ------- 0425b2a [DomCrawler] Always parse according to HTML5 rules thanks to the native DOM parser
|
Just a note/heads up for others that find themselves here... I upgraded to v7.4.0-BETA1 - ran my unit tests, which worked in 7.3.x - and now they fail :) Messages like: A reproducer is here https://github.com/PhilETaylor/symfony-7.4-alpine-reproducer I think this is a |
|
the fact that Alpine.js makes you write code that reports warnings in spec-compliant HTML parsers looks like an issue for alpine.js instead. However, it looks like the new code path does not use |
Totally agree. Just mentioning it as Im probably the first to see this since upgrading/testing the beta.
I only mentioned it as the OP mentioned this should remove b/c breaks, and then my tests broke up upgrade. Im not expecting a fix, unless you think one is needed. |
|
I suggest opening an issue instead of discussing that on a merged PR, to give it more visibility. |
This PR keeps the
DOM*-based API but uses the native HTML5 parser on PHP 8.4 instead of masterminds/html5.This works by parsing HTML strings using
Dom\HTMLDocumentthen serializing to XML, and loading again usingDOMDocument::loadXML().This basically replaces #61356 since it removes any BC breaks.
The drawback compared to a more native approach is the double-parsing that happens.
This could be worked on later by providing a way to leverage the new
Dom\*API directly.To be proved worth it before.