perf: improve URL parsing performance #9694
Conversation
anonrig
requested review from
a team,
stephanie-anderson and
ale-cota
and removed request for
a team
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
| } | ||
| } | ||
|
|
||
| const match = url.match(urlRegex); |
Check failure
Code scanning / CodeQL
Polynomial regular expression used on uncontrolled data
anonrig
force-pushed
the
improve-url-parsing
branch
from
1dcba82 to
67676d6
Compare
size-limit report 📦
|
anonrig
force-pushed
the
improve-url-parsing
branch
from
67676d6 to
07f0572
Compare
@mitsuhiko The current implementation is prone to bugs. Regexp approach is wrong and results in bugs. For example currently we don't handle |
| urlInstance: URL; | ||
| }>; | ||
|
|
||
| const urlRegex = /^(([^:/?#]+):)?(\/\/([^/?#]*))?([^?#]*)(\?([^#]*))?(#(.*))?$/; |
There was a problem hiding this comment.
Is there an external reference to the source of this regexp that would make sense to add here? I'm asking because these look daunting (at least to me), but I find that they can usually be a good way to educate, for example maybe by linking to the standard that this is supposed to support.
|
@anonrig I think we could use |
Follow up for #9676.
Without introducing any breaking changes, this is what I came up with that is in par with current regexp parsing. I'm not quite happy with it, but I think it's better than the current approach.
cc @AbhiPrasad