Re: [Bug-readline] certain wide characters mess up readline

[Chet Ramey]

On 5/24/16 1:42 PM, Grisha Levit wrote:
> This was in en_US.UTF8.  But I see now that the issue must be dependent on
> the underlying Unicode library support.

Not exactly.  It's not specific to Unicode per se.  It's the locale
definition and how the mb* functions and wcwidth() behave.

> 
> For example this works fine on Ubuntu 16.04, but not on CentOS 7 or OSX
> 10.11.  What I'm surprised about is that in either case the character *is*
> recognized as taking up 2 columns for the purposes of printing but not for
> the purposes of cursor movement during line editing.

That's not surprising.  The line is a buffer of bytes, and editing is
performed on potentially-multibyte characters.  If, for instance, mbrtowc
reports a sequence of bytes as composing an invalid multibyte character,
readline will treat them as a sequence of single-byte characters, and
output them to the terminal.  If the terminal emulator chooses to display
those bytes as a Unicode character, readline can't know that.

It may also be a problem with the redisplay code.  I can't tell yet.

-- 
``The lyf so short, the craft so long to lerne.'' - Chaucer
                 ``Ars longa, vita brevis'' - Hippocrates
Chet Ramey, ITS, CWRU    address@hidden    http://cnswww.cns.cwru.edu/~chet/