Craig Vespe asked me at the SIGSOFTVEND audio conference this morning to
describe some of the redo enhancements coming in Express 3.
 
The CI, Linkeditor, Volutil, Debug and probably some other subsystems all
call the same redo code wrt editing the command line.  (Unfortunately all
subsystems manage their own redo stack their own way.)
 
I have enhanced the redo procedure as follows:
 
^     means to upshift a character
v     means to downshift a character
 
dw    means to delete a word
^w    means to upshift a word
vw    means volkswagon or word downshift
 
d<delim> means to delete to delimiter
^<delim> means to upshift to delimiter
v<delim> means to downshift to delimiter
 
>^    means to upshift at end-of-line
>v    means to downshift at end-of-line
>^w   means to upshift the last word
>vw   means to downshift the last word
>dw   means to delete the last word
 
A word is currently defined as anything delimited by a space, comma or
semicolon (or end-of-line, beginning-of-line).  Leading word delimiters
are skipped when determining the end of the word.  The word delimiter
itself is not acted upon.
 
Examples:
---------
  abc.def ghi;;jk,lmn=op q/r/s/t
  dw              (deletes abc.def)
      ^w          (upshifts def)
  d.              (deletes abc)
             dwd  (deletes ;;jk,)
          d=dd    (deletes ghi;;jk,lmn=op)
 
  abc.def ghi;;jk,lmn=op q/r/s/t
  >^w             (upshifts q/r/s/t)
  >d=             (deletes op q/r/s/t)
  >dwdw           (deletes lmn=op q/r/s/t)
 
 
I've thought about making the word delimiters for redo the same as the word
delimiters for the new WORD() extraction evaluator function (which after
using both I find more useful), namely: space, comma, semicolon, tab, quotes,
equalsign, brackets and parenthesis....opinions??
 
regards,
Jeff Vance, MPE Lab
 
--