92747 - URL - http: with @ in request (HTTP error 404)

Status: RESOLVED WORKSFORME

(Core :: Networking, defect, P2)

Description

[Robert Grampp]

From Bugzilla Helper:
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:0.9.2+)
Gecko/20010729
BuildID:    2001072908

Mozilla cannot find an URL that contains "@" (see URL field!). Mozilla
transposes '@' to Hex %40 after entering the URL - but this should work as well.


Reproducible: Always
Steps to Reproduce:
1.enter URL
2.
3.

Expected Results:  show existing page

Comment 1

[Bradley Baetz (:bbaetz)]

Hmm. This is interesting. There is no need for us to escape the @, since it is
not reserved in this part of the uri. But rfc2616, section 3.2.3, says:

"Characters other than those in the "reserved" and "unsafe" sets (see RFC 2396
[42]) are equivalent to their ""%" HEX HEX" encoding."

Now, @ is in the reserved set in rfc2396, but section 2.2 of that says that:

"Characters in the "reserved" set are not reserved in all contexts. The set of
characters actually reserved within any given URI component is defined by that
component. In general, a character is reserved if the semantics of the URI
changes if the character is replaced with its escaped US-ASCII encoding."

Section 3.3 of that spec has grammar allowing escaped characters in the path.

However, since either behaviour is correct, we should do what ns4 and
(presumably) ie do, and not escape it.

So I don't think that we're doing anything wrong, but we probably shouldn't be
escaping, for compatability.

darin? benc? comments?

Comment 2

[John Taylor]

-> darin

Comment 3

[benc]

If someone types a URL from a user-entry point, I think we should escape it.

If an author posts something (links in page), my general take is to transmit in
a way the current browsers do this.

I've only red up to RFC 1738, so I am listening to anyone that quotes a newer
document...

Comment 4

[Darin Fisher]

investigating for 0.9.6, this works fine from NS 4.75 (tested under linux).
i think we should strive for compatibility in this case since the spec would
permit it.

Comment 5

[benc]

summary more consistent w/ other summaries now...

Comment 6

[Darin Fisher]

-> 0.9.8

Comment 7

[Darin Fisher]

not critical for mozilla1.0

-> future

Comment 8

[Thomas]

The url is no longer valid. I tried to google for other sites with "@" in the
url, but I couldn't find one that easy...
Anyone has another example?

Comment 9

[Darin Fisher]

-> default owner

Comment 10

[Gordon P. Hemsley [:GPHemsley]]

http://mozilla.gphemsley.org/bugs/bug_092747/[e-mail]/test.html

WFM.