bug#16733: messed up unicode chars in package description

[Glenn Morris]

Juanma Barranquero wrote:

> that if the response header does not include a coding, url-insert
> doesn't know what to do. Ironically, that's exactly the case with
> Glenn's original report:
>
> C:\> lwp-request -m HEAD
> http://elpa.gnu.org/packages/ascii-art-to-unicode-readme.txt
> 200 OK
> [...]
> Content-Length: 1255
> Content-Type: text/plain
> [...]

IMO this could be a bug in whatever script generates elpa webpages.
Seems like it should specify a charset, based on coding:, if present.

> So, at this point, I see the following alternatives:
>
> 1) Leave it as it is now, with
> detect-coding-string/decode-coding-string (or perhaps
> decode-coding-(inserted-)region).

Seems fine to me for 24.4. It's a minor issue.
Maybe reopen this as a reminder to revisit it afterwards (or new report).

(BTW, I'm surprised if url.el does not have a way to handle 404s.)