Re: bug not found. Encoding problem?
On Thu, 31 Jul 2003, Colin Watson wrote:
> On Thu, Jul 31, 2003 at 11:39:11AM -0500, Drew Scott Daniels wrote:
> > When looking at http://bugs.debian.org/cgi-bin/pkgreport.cgi?src=ppmd
> > thismorning (11:20AM CDT) I found that the link to bug #195016 [1]
> > didn't work and said ..."No bug number". Typing bugs.debian.org/195016
> > created a different URL [2] which worked. What's happening here?
>
> It works for me here with w3m.
>
> > I'm trying to find the encoding that didn't work and I can't...
> > perhaps my browser autodetected an unselectable encoding... A browser
> > bug? Can anyone check the logs to see what's actually being submitted?
>
> I'm not really sure what kind of "encoding" you might be referring to,
> but it certainly looks like a browser bug. Here are some matching log
> entries:
>
> (IP address elided) - - [31/Jul/2003:11:37:31 -0500] "GET /cgi-bin/bugreport.cgi?bug=195016\\ HTTP/1.1" 200 792 "-" "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; .NET CLR 1.1.4322)"
> (IP address elided) - - [31/Jul/2003:11:37:39 -0500] "GET /cgi-bin/bugreport.cgi?bug+AD0-195016 HTTP/1.1" 200 121 "http://bugs.debian.org/cgi-bin/pkgreport.cgi?src=ppmd"; "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; .NET CLR 1.1.4322)"
>
> The stray backslashes and that "+AD0-" business are just complete
> rubbish, not any kind of transfer encoding I'm familiar with (and in any
> case a transfer encoding is not allowed to affect the text of links as
> interpreted by the browser). Try looking at the source of the referring
> page to see if it really says that; if so, it's either a browser bug or
> network corruption, perhaps in a broken transparent proxy somewhere.
>
Fwiw, I'd guess it's a browser bug since I've seen it before. Thanks for
the logs. The text looks fine when I do a copy link location and even in
the Address bar. Somehow some funky encoding gets selected when I go from
the pts to the bts to a bug for certain packages and in some other
situations.
There must be a few ways to work around this. The BTS including some
encoding information in it's header might be helpful? The PTS has <meta
content="text/html; charset=ISO-8859-1" http-equiv="Content-Type"> in it's
header... I think I have trouble with the PTS too ocationally, but I can't
rember for sure. Microsoft suggests [1] that this workaround would help
people like me who are stuck using broken browsers (although, pesonaly,
this is only sometimes).
[1] http://msdn.microsoft.com/library/default.asp?url=/workshop/author/dhtml/reference/charsets/charset4.asp
Note the encoding on this page is set wrong, it should be UTF-8 and not
windows-1252.
Drew Daniels
Reply to: