dont now, must be added at the last update - maybe (it is not in the old antville skins)

eg: balabuja.antville.org source starts with (replace # with >)

#?xml version="1.0" encoding="UTF-8"?# #!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "www.w3.org"#

#html xmlns="www.w3.org" xml:lang="en" # #head# ...

xml, encoding and the namespace declaration are new

see message below. The xml declaration is OK, but the encoding="UTF-8" is bogus. I removed it, and it should now be regarded as what it is, i.e. ISO-8859-1 encoded. Let me know if it isn't.

i suspect, that you are translating to utf-8 literally when you parse the skins at save time

adding a "ü" to a skin results in a "ü" at first save and a "ü" at the second save...

[Mozilla 1.1 // Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.1) Gecko/20020826]

Please go to the skin editor page and open the View -> Character Coding menu of your Mozilla browser. What does it say? It should say Western (ISO-8859-1). Is it possible you manually set this to UTF-8 somewhere on that specific antville weblog? Does manually setting it to "Western" work? (It should.)

It's possible that the existence of the declaration makes Moz think it's UTF-8, because that's infact the standard encoding for xml, but the web server explictly sends Content-Type: text/html; charset=iso-8859-1 in the response header. Heck, maybe it's the discrepance between text/html and the decl that's causing the problem.

Anyway, this is all due to exagerated standards compliancy, both from our side and from Mozilla.

PS: everything works right for me with Moz 1.1 b on Linux.

character settings already is at "Western (ISO-8859-1)".

i remember that encoding defaults to utf-8, but that depends on your parser software. again, i suspect that your parser converts (skins only) to utf-8 internally.

i understand, that i should have used html-entities, anyway. that is no problem to do and the most easy solution for both of us.

No, you definitely don't have to use HTML entities. Yes, you can use HTML entities.

But if the setting is to Western and it sends/receives ü as ü, then this is a bug in Mozilla 1.1, I can say that with 100% certainty. But try restarting Mozilla, maybe it has the UTF-8 encoding memorized from previous requests.

After that, I'll add an explicit encoding="ISO-8859-1" to the page so there should be no more problems.

the character "ü" as displayed in the form/textarea is a "ü" in the source - funny farm!

once it's been corrupted, it is correct that it remains corrupted. The interesting thing would be if it shows correctly when freshly entered.

sorry hns, i think it is my browser

it's gone crazy now - replacing umlauts with a "?" sign... on certain sites.

thank you for your patience.

update: i did a reinstall of mozilla. everything is fine now. hns, i owe you a coffee.