[Wtr-general] OT: strange UTF-8 phenomenon

[Chris McMahon]

Hi...

I'm having a (slow) conversation about UTF8 in soap4r and http-access2
with Hiroshi Nakamura, and in the course of submitting a problem
report ran across the following very weird behavior:

cut/paste to normal CMD prompt:
>str = 'è or à'

cut/paste to IRB prompt:
irb(main):002:0* >str = 'ach or acc'

for those of you following along with the saga of UTF8, this might be
a clue of some sort.  Any conjectures as to why IRB behaves like that
would be welcome.