On Thu, 08 Sep 2016 10:08:23 -0400, Anthony Coulter wrote:
> We can fix either the manual or ksh itself; this diff takes the latter
> approach. It is tempting to do this with "typeset -ir PPID" but that
> actually doesn't work:
>
> $ FOO=123
> $ typeset -ir FOO
> ksh: FOO: is read only
Perhaps we should fix the above issue as well, other ksh versions
don't throw an error there. Below is a simple fix, though a bit
gross due to the magic 0x4.
- todd
Index: bin/ksh/var.c
===================================================================
RCS file: /cvs/src/bin/ksh/var.c,v
retrieving revision 1.55
diff -u -p -u -r1.55 var.c
--- bin/ksh/var.c 30 Dec 2015 09:07:00 -0000 1.55
+++ bin/ksh/var.c 8 Sep 2016 15:20:20 -0000
@@ -661,6 +661,7 @@ typeset(const char *var, int set, int cl
*/
for (t = vpbase; t; t = t->u.array) {
int fake_assign;
+ int error_ok = KSH_RETURN_ERROR;
char *s = NULL;
char *free_me = NULL;
@@ -683,6 +684,10 @@ typeset(const char *var, int set, int cl
t->type = 0;
t->flag &= ~ALLOC;
}
+ if (!(t->flag & RDONLY) && (set & RDONLY)) {
+ /* allow var to be initialized read-only */
+ error_ok |= 0x4;
+ }
t->flag = (t->flag | set) & ~clr;
/* Don't change base if assignment is to be done,
* in case assignment fails.
@@ -692,7 +697,7 @@ typeset(const char *var, int set, int cl
if (set & (LJUST|RJUST|ZEROFIL))
t->u2.field = field;
if (fake_assign) {
- if (!setstr(t, s, KSH_RETURN_ERROR)) {
+ if (!setstr(t, s, error_ok)) {
/* Somewhat arbitrary action here:
* zap contents of variable, but keep
* the flag settings.
