Try charset="utf=8" and an acronym topic.

BTW, why use utf-8? To support copy-past from websites into TWiki topics without having curly quotes converted to question marks.

AC

I don't understand the first line, or what the bug is - where are you setting this, in the GET URL or something? Please give an example of the problem, ideally attaching the topic .txt to this report - see TWiki:Support.SupportGuidelines for hints on reporting I18N bugs.

Generally, setting %CHARSET% is a mistake and should not be necessary. You just need to set a locale ending in .utf8 if that's what you want to use - however, as discussed in the TWiki web docs, it's really only a good idea to use UTF-8 if you are using Chinese, Japanese or some other double-byte language, where UTF-8 is the only safe encoding and the loss of WikiWord I18N is not an issue.

Enabling copy-paste from other websites is a good idea but it means forgoing a lot of I18N support if you need I18N of WikiWords.

RD

Do you mean I should set this: en_US.utf8 ?

AC

You should set {Site}{Locale} in configure to en_US.utf8 - the {Site}{CharSet} is only there as an override for some unusual cases (e.g. charset is spelt differently in the locale from the spelling acceptable in HTTP headers for use by browser).

See TWiki:TWiki.InstallationWithI18N and please update that if parts are unclear - there's a bit under troubleshooting saying not to set the CHARSET var, maybe you could do a similar bit for {Site}{CharSet}?

Also, curly quotes are often a Windows-1252 character set peculiarity - going UTF-8 is a bit radical just for that.

• UPDATE: For your site, you could just set the locale to use ISO-8859-1 and then set {Site}{CharSet} to Windows-1252 so that Windows-based browsers don't convert the Windows-1252 only characters (e.g. quotes) to NCRs. This is probably the simplest solution all round.

If it would be OK to convert them to normal ASCII double/single quotes (i.e. " and '), it would be a useful feature to recognise the numeric character reference and convert them. There's some code out there that's similar - maybe Google:demoroniser?

TWiki support for UTF-8 is very limited, no WikiWord autolinking of I18N characters for example - however, if you are using it in English only you will be OK.

Have set this to Discarded since this is not a bug (except perhaps in docs).

RD