On Wed, 12 Oct 2016 23:27:28 +0000, Lindy Mayfield <lindy.mayfi...@sas.com> 
> ISPF-L, no, MVS-OE list, yes.  But since it's often a system wide setting 
> that IBM may or may not recommend, that’s why I chose IBM-Main first to ask.  
> It affect the entire machine, not just OMVS guys.
This motivates several questions:

o On editing a Classic file or a new or untagged UNIX file,
  should the programmer be allowed to select a CCSID?  The
  only options at present are UTF-8 and ASCII (I assume this
  means ISO8859-1).

o Should the default be the CCSID of the terminal?

o Should it be possible to change the CCSID of the Edit
  session in progress?

o When a new UNIX file is saved, should it automatically be
  tagged with the CCSID of the Edit session?

o On a COPY of a file with a tagged CCSID different from the
  Edit session, what should happen?  Should the COPY dialog
  allow the programmer to specify or override the file's CCSID?

Edit does a very good job of dealing with a file CCSID different
from the terminal's CCSID, for example 1208 and 1047.  But if I
change a character to (nondisplayable) x'00'; HEX OFF; FIND P'.';
Edit places the cursor at an incorrect location and tells me
"x'20' found."  That's a (displayable) ASCII blank that is present
at that incorrect cursor location.

UTF-8 introduces several varieties of "nondisplayable":

o A valid UTF-8/Unicode representation of a character not
  in the terminal's repertoire.

o A valid UTF-8 representation of an unassigned Unicode point.
  (How badly do we actually need a brontosaurus emoji!?)

o An overlong UTF-8 code.

o An undecipherable UTF-8 code.

Should the diagnostics differentiate among these?

> -----Original Message-----
> From: Paul Gilmartin
> Sent: keskiviikkona 12. lokakuuta 2016 23.28
Hmmm... You asked about Danish, but your Mail Agent seems to
be speaking Finnish.

-- gil

For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Reply via email to