Issue: Import from Mediawiki - Stop at step 2

After exhaustive testing, I have determined that iconv() is useless and I'm not sure why it exists - you essentially have two choices when characters aren't the same: let it crash, or strip the characters. It looks like mb_convert_encoding() works as expected for this situation. I am working on incorporating it into the importer for the next build - the problem is figuring out when it is actually needed, since default vBulletin 4 would be UTF-8 -> UTF-8, and otherwise vBulletin's character set isn't easily readable any more.

Ok - it's not possible to convert my forum to UTF-8.
I think i'll do the import of the important sites manually. by the way best testing of the vaultwiki ;-)

The encoding issues have been fixed for the next build - it will now attempt to convert both text AND titles. It's possible some conversions will still be incorrect, but more common ones like UTF-8 -> ISO-8859-1 should be handled correctly. It attempts to determine the existing encoding by analyzing the input string, and converts these to the detected encoding used by vBulletin - so make sure you are importing while viewing the AdminCP with the same language you want the imported wiki pages in.

The previous implementation used iconv, which did not work if the locale was set incorrectly, or if MediaWiki did not flag its data as UTF-8. Either way, more character set combinations should now be supported.

For further issues with the importer, please open a new report.

OK, thanks. Will try it...