borked character encoding, firefox

September 9, 2007 5:22 PM

Did I do something to mess up character encoding in firefox?

Comments like this one don't seem to be rendering correctly no matter what encoding I pick, and auto-detect just goes with utf8. For someone that sees it properly, what encoding do you have for it?

The uber bit in "Ein Realplayer. DRM, DRM, über alles!" bit gives me an A with the curly accent over it and a 1/4 symbol instead of what I'm guessing to be an accented U, which ain't right. Help!

Running FF2.0.0.6
addons:
adblock plus
colorzilla
customizegoogle
download statusbar
fasterfox
fontfinder
text/plain
long titles
locationbar^2
minimize to tray
tab mix plus
measureit
linkification
tiny menu
greasemonkey: mefinavigator, mefi comment dividers, gmailto

The problem is not specific to mefi, I've seen it on other sites, it's just the latest instance that prompted me to ask.
This is another one that I can't seem to get to render right (the accented character in flaneur is messed up) despite fiddling about with encoding types. I distinctly remember that page, in particular, looking fine at some point in the past.

posted by juv3nal to computers & internet (5 comments total)

oh..btw, I'm on xp if that has any relevance
posted by juv3nal at 5:23 PM on September 9, 2007

I'm seeing gibberish as well.
Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8.1.6) Gecko/20061201 Firefox/2.0.0.6 (Ubuntu-feisty)
posted by flabdablet at 5:46 PM on September 9, 2007

No, you're seeing it correctly, at least on that MeTa page. The actual byte sequence being sent by the server is C3 83 C2 BC, which is UTF-8 for U+00C3 "A WITH TILDE" followed by U+00BC "VULGAR FRACTION ONE QUARTER".

What clearly happened is that someone else applied the characters-to-UTF8 transformation twice, because if the bytes had been C3 BC you would have gotten a lower-case u with dots (U+00FC). So yes, there's an encoding problem somewhere, but in this case your browser is accurately displaying what Metafilter is sending.

(This is visually indistinguishable from the misinterpretation that happens if the server sends the right thing but the browser interprets it in 8859-1. But that's not what's happening in this particular case.)
posted by hattifattener at 5:50 PM on September 9, 2007

Right, the encoding mismatch has already happened and has been set in stone by copy and paste.
posted by Rhomboid at 6:55 PM on September 9, 2007

Wow, never occurred to me that the page might already be messed. Thanks.
posted by juv3nal at 8:30 PM on September 9, 2007