Re: [tlug] Re: WWWJDIC backdoor issue

["Josh Glover" <jmglov@example.com>]

On 16/06/07, Jim Breen <jimbreen@example.com> wrote:

> > But I wonder why the UTF-8 method doesn't work with Japanese text any
> > more. Am I doing something wrong, or is there a recent bug in WWWJDIC?
>
> The former. You have a space in front of the åã (see the %20 in the URL
> you quoted.) Any "Japanese" key starting with a non-Japanese character
> (e.g. a space) is assumed to be in romaji.

/me smacks his forehead

Thanks, Jim. I retried my query, sans the space:

http://www.aa.tufs.ac.jp/~jwb/cgi-bin/wwwjdic.cgi?1MUJ%B9%A5%A4%AF

but I still get no matches. I get an odd error message that leads me
to believe maybe I am having an encoding-related problem:

"Search Key: ïï (longest match - original key: ïïïï"

Obviously, using the WWWJDICT interface directly to search on "åã"
gives me the expected results.

The odd thing is, this used to work perfectly in Firefox. I wonder if
something changed in the quicksearch implementation between 1.5 and
2.0 that broke this.

But a quicksearch on Japanese Wikipedia works fine:

http://ja.wikipedia.org/wiki/%E9%88%B4%E6%9C%A8%E4%B8%80%E9%83%8E

That is the disambiguation page for "éæäé".

Ah... just found an interesting clew. Doing the following quicksearch:

http://www.aa.tufs.ac.jp/~jwb/cgi-bin/wwwjdic.cgi?1MUJ%s

for "éãã" results in the no match error as described above, and the
following in the URI bar:

http://www.aa.tufs.ac.jp/~jwb/cgi-bin/wwwjdic.cgi?1MUJ%BF%A9%A4%D9%A4%EB

Stripping everything after the ?1MUJ from the URI, typing in "éãã",
and submitting works, with the following appearing in the URI bar:

http://www.aa.tufs.ac.jp/~jwb/cgi-bin/wwwjdic.cgi?1MUJ%E9%A3%9F%E3%81%B9%E3%82%8B

I am now wondering if the first is not UTF-8? Can an encoding expert
step in and tell me for sure?

WTF is Firefox doing?

--
Cheers,
Josh