92776 - URL is missparsed on this JIS page

Status: VERIFIED FIXED

(Core :: Internationalization, defect)

Description

[Katsuhiko Momoi]

** Observed with 7/27/2001 Win32 branch build **

On the above page, there is an HREF link which contains JIS-encoded
characters as part of the URL. Mozilla missparses them and
cannot find the intended page. Neither IE 5.5 nor Comm 4.x has 
this problem.

The problem part contains Japanese characaters "1", "Space" and "1".
We missparse these JIS sequences.

To look for the problem part, search for "1998" on the above
page. It is the 1st link below "1998" that has this problem.

I will attach a mimum test case extracted from the above page 
below. Just click on it with various browsers and see that only
Mozilla fails.

Comment 1

[Katsuhiko Momoi]

Attachment: A minimum test case. Just try to click on the URL link, you will see the problem. The file is encoded in ISO-2022-JP.

Comment 2

[nhottanscp]

Attachment: Patch, avoid Finish to overwrite a result of Convert.

Comment 3

[Roy Yokoyama]

looks good.  /r=yokoyama

Comment 4

[Johnny Stenback (:jst)]

spec += p; could be changed to spec.Append(p); to make it clearer what it does,
but either way, sr=jst

Comment 5

[nhottanscp]

checked in

Comment 6

[Yuying Long]

Fixed verified on 08-06 trunk build on Win2k-CN and Linux RH6.2-Ja:
Can find the right page now, but there is a problem that can not point to right
anchor place, I already filed bug 94017 for that.

Mark verified for this one here.