Hi Donna,

On Apr 14,  9:51am, Donna Garverick wrote:
> how to nomalize or fix hppath to take out duplicate
> references.  i believe someone posted a bit of script to do
> this but i'm not having any luck finding it on raven.
> perhaps the better question is -- is it even worthwhile to do?

I imagine it is worthwhile for long HPPATHs with lots of dups.
Anyways here is a script I wrote that should work.  Note, it handles
duplicate MPE and POSIX elements in the path variable.

Jeff Vance, CSY

------------------cut here---------------
PARM path=HPPATH
# This script removes all duplicate entries in the "path" parm.  MPE
# and POSIX path elements are accounted for.  The "path" parm names
# a CI variable that contains HPPATH-like elements.
#
setvar i 0
while setvar(tok,word(!path,,setvar(i,i+1),n)) <> "" and &
      n < len(!path) do
   setvar mpe_tok (fsyntax(tok) = "MPE")
   # do a token replace rather than a character replace, as repl() does,
   # so that substrings of path elements aren't incorrectly replaced
   setvar j i
   while setvar(next_tok,word(!path,,setvar(j,j+1),m)) <> "" do
      setvar mpe_next_tok (fsyntax(next_tok) = "MPE")
      if (mpe_tok and mpe_next_tok and dwns(tok) = dwns(next_tok)) or &
         (not mpe_tok and not mpe_next_tok and tok = next_tok) then
         # remove matching element
         setvar !path lft(!path,n)+rht(!path,-m-1)
         # decrement target word (j) due to match
         setvar j j-1
      else
         # adjust ending index of the last unique path token
         setvar n m
      endif
   endwhile
endwhile

--