Comment 0 for bug 769554

The patch that closed bug #629996 causes a serious regression for latin-1 (non-utf-8) characters. Specifically, all string are truncated after the first non-ascii character.

I opening this bug because I don't have access to re-open bug #629996.

While the original patch was indeed accepted upstream months ago, it was changed hours later when the regression was found. Only the original patch was manually applied in Ubuntu.

Proposed merge above re-fixes the bug without the regression, as well as several other segfaults by upgrading to upstream.