You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The URL https://www.42x.42/ parses in version 2.2 of the crate but in 2.3 fails with InvalidIpv4Address. Reading https://url.spec.whatwg.org/#concept-ipv4-parser, my naive reading implies parsing this as ipv4 is correct as the domain ends with numbers. However, curl 8 doesn't agree (it attempts to resolve the host) and up until recently Firefox didn't either (it now does because it's on url 2.5 :)
What I'm less clear on is whether this change was intentional in a minor version bump - the fact this URL no longer works in Firefox was surprising and has a number of implications (eg, databases which carefully store only validated URLs now need to deal with the possibility some stored URLs are actually invalid, etc)
If it was intentional, is there any policy for such changes? I can't see where changes to parsing semantics was called out in either the docs or the upgrading notes, so I'm wondering if there's some way we should have know to expect this kind of change in this kind of version bump?
The text was updated successfully, but these errors were encountered:
What I'm less clear on is whether this change was intentional in a minor version bump - the fact this URL no longer works in Firefox was surprising and has a number of implications (eg, databases which carefully store only validated URLs now need to deal with the possibility some stored URLs are actually invalid, etc)
That is an unfortunate side effect. Thankfully this kind of URLs are pretty rare on the web.
The URL
https://www.42x.42/
parses in version 2.2 of the crate but in 2.3 fails withInvalidIpv4Address
. Reading https://url.spec.whatwg.org/#concept-ipv4-parser, my naive reading implies parsing this as ipv4 is correct as the domain ends with numbers. However, curl 8 doesn't agree (it attempts to resolve the host) and up until recently Firefox didn't either (it now does because it's on url 2.5 :)What I'm less clear on is whether this change was intentional in a minor version bump - the fact this URL no longer works in Firefox was surprising and has a number of implications (eg, databases which carefully store only validated URLs now need to deal with the possibility some stored URLs are actually invalid, etc)
If it was intentional, is there any policy for such changes? I can't see where changes to parsing semantics was called out in either the docs or the upgrading notes, so I'm wondering if there's some way we should have know to expect this kind of change in this kind of version bump?
The text was updated successfully, but these errors were encountered: