On Tue, Apr 29, 2014 at 12:07:52PM +0200, Werner Fink wrote:
> The alias problem is a side effect of
>
> --- ksh93/src/cmd/ksh93/sh/parse.c.alias
> +++ ksh93/src/cmd/ksh93/sh/parse.c
> @@ -1545,7 +1545,7 @@ static Shnode_t *simple(Lex_t *lexp,int
> {
> if(nv_isattr(np,BLT_DCL))
> {
> - assignment =
> 1+(*argp->argval=='a');
> + assignment = 1;
> if(np==SYSTYPESET)
> lexp->intypeset = 1;
> else if(np==SYSENUM)
>
> without this change the alias error is gone. Could be related to the
> release entry
>
> 14-04-05 A bug in which type name starting with the letter a, did not
> allow instances to be created has been fixed.
Just re-enabled my already reported patch which now includes
+++ ksh93-crashes.dif (working copy)
@@ -64,6 +64,15 @@
if(flag<=0)
stakdelete(slpold->slptr);
else
+@@ -1548,6 +1551,8 @@ static Shnode_t *simple(Lex_t *lexp,int
+ else if(np==SYSENUM)
+ lexp->intypeset = 2;
+ key_on = 1;
++ if
(strcmp(argp->argval,"alias")==0)
++ assignment++;
+ }
+ else if(np==SYSCOMMAND)
+ cmdarg++;
and also seen that the error detected in builtins.sh is the new check
added with 2014-04-15
+cd "$tmp"
+if mkdir -p f1
+then redirect {d}<f1
+ pwd=$(pwd)
+ ( cd -f $d && [[ $(pwd) == "$pwd/f1" ]]) || err_exit '$(pwd) does not
show new directory'
+ [[ $(pwd) == "$pwd" ]] || err_exit '$(pwd) is not $pwd'
+ [[ $(/bin/pwd) == "$pwd" ]] || err_exit '/bin/pwd is not "$pwd"'
+ [[ $(/bin/pwd) == "$(pwd)" ]] || err_exit '/bin/pwd is not pwd'
+ cd "$pwd"
+ rmdir "$pwd/f1"
+fi
I've tried this on the command line
mkdir f1
redirect {d}<f1
pwd=$(pwd)
echo $pwd
/proc/17529/fd/12/.
ls -l /proc/17529/fd/12
lr-x------ 1 abuild abuild 64 Apr 29 11:11 /proc/17529/fd/12 ->
/usr/src/packages/BUILD/ksh93/f1
the directory is identical but not the path string ... maybe it is better to
use
pwd -P
/usr/src/packages/BUILD/ksh93/f1
or
pwd -f $d
/usr/src/packages/BUILD/ksh93/f1
--
"Having a smoking section in a restaurant is like having
a peeing section in a swimming pool." -- Edward Burr
pgppGcXAEuC_k.pgp
Description: PGP signature
_______________________________________________ ast-users mailing list [email protected] http://lists.research.att.com/mailman/listinfo/ast-users
