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
It seems we had a restriction that a URL can't have a fragment nor credentials parts,
however, it seems these restriction were removed from the format at this commit. WICG/webpackage@826cadf
For each (url, responses) entry in the index map:
Let parsedUrl be the result of parsing ({{URL}}) url with
no base URL.
If parsedUrl is a failure, its fragment is not null, or it includes
credentials, return an error.
It is a bit unclear whether this removal was intentionally done or not.
If this was intentionally done, that might be fine.
We don't have any restriction on the format itself. Instead, each use case can introduce such a restriction.
e.g.
Subresource loading will ignore a URL which has a fragment.
I didn't intend that change to make the subresource loading use case more lenient, but I think the WG could decide either way on whether the format itself forbids URLs with fragments or credentials.
I suggest we say that index entries MUST NOT include fragments or credentials, but of course the use cases will have to explicitly specify what to do when bundles violate that requirement.
suggest we say that index entries MUST NOT include fragments or credentials
That sounds reasonable to me, and matches the current chromium's bundle parser's implementation, and the behavior of CLI tools, I think, though I don't have a strong preference here.
It seems we had a restriction that a URL can't have a fragment nor credentials parts,
however, it seems these restriction were removed from the format at this commit. WICG/webpackage@826cadf
It is a bit unclear whether this removal was intentionally done or not.
If this was intentionally done, that might be fine.
We don't have any restriction on the format itself. Instead, each use case can introduce such a restriction.
e.g.
Subresource loading will ignore a URL which has a fragment.
cc: @jyasskin
It would be better to clarify our intention in this issue.
The text was updated successfully, but these errors were encountered: