[Bug-readline] UTF-8 positioning errors

[Thomas . Wolff]

Hello,
In UTF-8 mode, I see two cursor positioning problems in combination 
with a complex PS1 setting and one independent logical positioning 
problem.

input: aübx

^A:
visually:  cursor on x
internal: cursor on b
-> logical error: cursor should actually move to a since ü is a letter
-> layout error: cursor should be placed on b

input i:
visually: aübibx, cursor on x
internal: aüibx, cursor on b (can be seen after ^R)
-> layout error: previous error propagates to bogus display, cursor 
   again misplaced

input j:
visually: aübibjbx, cursor on x
internal: aüijbx, cursor on b (can be seen after ^R)

PS1 value literally (likely to be mangled in mail):
\[]10;black;address@hidden:\[\]$(dirsuf)\[\]: 
\[]2;address@hidden:\w]1;\W\]
PS1 value, using ^[ and ^G for ESC and BEL:
\[^[]10;black;address@hidden:\[^[[7m\]$(dirsuf)\[^[[0m\]: 
\[^[]2;address@hidden:\w^G^[]1;\W^G\]


In cygwin, I see additional problems in combination with bash filename 
completion, same PS1 setting:

> ls x*
x.exe  xchars  xpc
> ls xTAB (using TAB for filename completion):
x.exe   xchars  xpc
> ls x x (produced by readline)
-> layout error: extra space and x, not fixed by ^R

> echo xTAB (using TAB for filename completion):
x.exe   xchars  xpc
> echo xx (produced by readline)
-> layout error: extra x, this time no extra space


Kind regards,
Thomas Wolff