Hi Christoph, Dennis,
Well, I don't think it would be a big deal to move the bundled lexbor to
somewhere where it is always available. I mean, so far it's only used
by ext/dom so it's bundled there, but if other parts of the php-src code
base would use it, we could put it elsewhere.Exactly. You might be aware that I'm working on an "uri" extension (https://externals.io/message/123997)
Yes, and I only briefly saw that before, but I’m excited, because I’ve wanted very much to be able to properly parse URLs within PHP. Myself, I was also interested in seeing if we could get Ada into the language.
As with HTML parsing, I see much value in having additional interfaces that aren’t a DOM interface but which are designed for specific software purposes.
and it also needs some parts of lexbor. My implementation currently depends on ext/dom
for simplicity's sake, however if the vote once passes, this temporary solution has to be changed.
Therefore we previously agreed with Niels that we would make lexbor an "internal extension" (similar to mysqlnd), or
at least we would somehow find a way for it to be always available, just like how Christoph said.
With all the improvements going around PHP these days, I find it extremely important to finally be able to reliably and safety understand some of the most basic content that we produce and parse: HTML and URLs.
Although the user-space libraries are of varying completion and quality, all of them suffer from the fact that it’s so challenging to efficiently parse most content using PHP. Getting these things baked into the language of the web will bring a potent uplift to the entire ecosystem, both because there will be less corruption, but also because performance won’t suffer in getting there.
Regards,
Máté