Re: alternation option end of string

What you really want is to _remove_ whatever comes after either "cats" or,
if the thing after "cat" is not an s, "cat".

But if you're reading comma-separated fields, you can just use `read`:

    IFS=, read base idx comment <<<"$INPUT"

e.g.

for INPUT in file file, file,8 file,88 file,8,comment file,88,comment \
             file,8a file,a8 file,a8,comment file,8a,comment
file,8a,com,ment \
             file,8,com,ment file,8,comm,9 file,8,comm,9a file,8,comm,a9; do
  IFS=, read base idx comment <<<"$INPUT"
  printf '\n%8s: %s\n' INPUT "$INPUT"
  printf '%8s:\t%s\n' base "$base"
  printf '%8s:\t%s\n' idx "$idx"
  printf '%8s:\t%s\n' comment "$comment"
done

   INPUT: file
    base: file
     idx:
 comment:

   INPUT: file,
    base: file
     idx:
 comment:

   INPUT: file,8
    base: file
     idx: 8
 comment:

   INPUT: file,88
    base: file
     idx: 88
 comment:

   INPUT: file,8,comment
    base: file
     idx: 8
 comment: comment

   INPUT: file,88,comment
    base: file
     idx: 88
 comment: comment

   INPUT: file,8a
    base: file
     idx: 8a
 comment:

   INPUT: file,a8
    base: file
     idx: a8
 comment:

   INPUT: file,a8,comment
    base: file
     idx: a8
 comment: comment

   INPUT: file,8a,comment
    base: file
     idx: 8a
 comment: comment

   INPUT: file,8a,com,ment
    base: file
     idx: 8a
 comment: com,ment

   INPUT: file,8,com,ment
    base: file
     idx: 8
 comment: com,ment

   INPUT: file,8,comm,9
    base: file
     idx: 8
 comment: comm,9

   INPUT: file,8,comm,9a
    base: file
     idx: 8
 comment: comm,9a

   INPUT: file,8,comm,a9
    base: file
     idx: 8
 comment: comm,a9


On Sun, May 3, 2026 at 4:23 PM Ray Andrews <[email protected]> wrote:

>
>
> On 2026-05-03 11:45, Bart Schaefer wrote:
> > On Sun, May 3, 2026 at 6:56 AM Ray Andrews <[email protected]>
> wrote:
> >> It might not actually be the '$' just some mechanism for matching a
> >> string that must end after some substring.
> > That's not what you're asking for, tho.  You're asking for the entire
> > string to fail to match if there's anything after the substring, but
> > in the same pattern you want the substring itself to match so you can
> > extract it.  You can't simultaneously have a failure and a success.
> > That's why Roman's solution has (compare && out=success ||
> > out=failure) -- there's no single pattern expression that can do both
> > branches.  (There might be a perl expression that works, but I'm not
> > confident pcre is perl-ish enough to do it.)
> I thought the thing was logically possible because it can be
> articulated: 'cat' then either nothing || an 's' which can itself be
> followed by anything.  But if it's not doable, as I said, no matter cuz
> the functionality is there and I'd do it Roman's way anyway cuz it's
> self-explanatory.  Too much terseness already, so I wish for more
> terseness very lightly.  Sorta like a while back I was wondering if:
> (( x > 5 )) && x=5
> ... had a more compact syntax -- if not so what, it's not a problem,
> still, sometimes zsh offers these little shortcuts.
>
>
>
>
>

-- 
Mark J. Reed <[email protected]>