Forum character encoding

[JamesF]

This post contains a hyperlink with Greek characters in it: http://www.rib.net/forum/f29/greek-r...s-29548-7.html

They render as gibberish in my browser because the page's Content-Type header and meta tags declare the character encoding as ISO-8859-1, but those characters are in UTF-8. If I view the page as Unicode, they look mostly correct.

There's something similar happening here: http://www.rib.net/forum/f8/free-bac...tml#post667564

Viewed using the declared (ISO-8859-1) encoding, the hyperlink in that post contains two capital letters A with circumflex; viewed as Unicode, they disappear, and some of the other characters are rendered wrongly instead. (Because, although the multi-byte character » in UTF-8 is coincidentally the same as the two characters » in Latin-1, that coincidence doesn't extend to the apostrophe and the ellipsis, because they then get rendered with the Unicode replacement character: �).

Oddly, if I put multi-byte Unicode characters like arrows into a post, they seem to work: ←→ ←link test→. If I look at the page source, all of the arrows use HTML character entities. If I put Greek characters into a post (π σ κ), they also use character entities, so they also always render correctly.

So, because my posts all seem to work, I'm not sure if it's just a problem for hyperlinks, or whole posts, or something else.

For reference, I'm viewing and posting this using Firefox 35 on a Linux desktop.

[Janet H]

When did this begin to occur?

[Poly]

James - I *think* the issue arrises when someone puts in a hyperlink in their post. The forum software intelligently tries to use the title tag of the destination page to create the link text. It seems if the title tag contains complex character sets this confuses things.

[A1an]

Whoooosh.......

[JamesF]

When did this begin to occur?

Unfortunately I don't know. I've definitely seen it before, but I can't remember when — it just looks like noise and one tends to ignore it until the programmer within eventually puts two and two together and gets 0b100. Not terribly helpful in a bug report, I know.

e-ENPLO: Viper Ribs 2015: Το Î*ανελλήνιο Î*Ï�ωτάθλημα Φουσκωτών Σκαφών - The Hellenic Rib Racing championship

Poly: you're right. If I copy-and-paste the same URL from the Greek RIB Championship thread post #62, my post has the problem.

The pasted URL is sent like this in the POST request body:

http%3A%2F%2Fwww.e-enplo.com%2F2015%2F02%2Fviper-ribs-2015-hellenic-rib-racing.html

So it appears to be a server-side thing, not the post editor.

Edit: I wondered if the RIBnet forum software was behaving correctly but being given an incorrect content-type. But the response from that URL properly declares itself as UTF-8 in its HTTP Content-Type header and a meta tag.

[willk]

Originally Posted by A1an

Whoooosh.......

"Whoosh - right over your head"
or
"Whoosh - SuperPoly to the rescue"?