The following command sequence (slightly altered for clarity) was
entered while investigating the discussion in
http://lists.research.att.com/pipermail/ast-users/2013q3/004188.html.
$ print ${.sh.version}
Version AJM 93u+ 2012-06-12
$ index=1
$ print ${!not_an_array[index]}
not_an_array[index]
$ set | grep array
_='not_an_array[index]'
$ typeset -n ref=not_an_array[index]
$ set | grep array
not_an_array=([0]=)
ref='not_an_array[1]'[1]
$ print ${!not_an_array[index]}
not_an_array[1]
$ unset not_an_array
$ set | grep array
_=not_an_array
not_an_array=([0]=)
ref='not_an_array[1]'[1]
$ unset -n ref
$ set | grep array
not_an_array=([0]=)
$ unset not_an_array
$ set | grep array
_=not_an_array
not_an_array=([0]=)
$
I note the following bugs therein:
1. "typeset -n ref=not_an_array[a]" magically creates the associative
array, not_an_array, with empty element "0". Before that
not_an_array didn't exist.
2. After that the output of "print ${!not_an_array[index]}" changes
from "not_an_array[index]" to "not_an_array[1]".
3. The magically created not_an_array cannot be unset even after ref
is unset.
Note that this occurred in ksh-u. You guys have made a lot of
corrections to name references and the output of the set command since
ksh-u. Maybe these bugs don't even exist anymore.
Terrence Doyle
_______________________________________________
ast-users mailing list
[email protected]
http://lists.research.att.com/mailman/listinfo/ast-users
