Re: PATCH: fix for SEGV

[Sven Wischnowsky <wischnow@xxxxxxxxxxxxxxxxxxxxxxx>]

Bart Schaefer wrote:

> On Aug 24,  2:16pm, Sven Wischnowsky wrote:
> } Subject: PATCH: fix for SEGV
> }
> }   ...) ./zsh -f
> }   % setopt autocd; hash wc=/bin/wc; echo|wc
> } 
> } It'll give you a SEGV in `isreallycom()', trying to look at
> } `cn->u.name' which isn't set because for hashed commands `cn->u.cmd'
> } is used.
> } 
> } +    if (cn->flags & HASHED)
> } +	return 1;
> 
> Is that really the right fix?  I thought the whole point of isreallycom()
> was to find cases where the hash table contains a mapping that doesn't
> refer to a real executable.  I'm not sure it should succeed just because
> the mapping was added explicitly.

I wasn't sure either and was hoping for a reply. isreallycom() is used
only in one place, so we could easily move the extra test there.

Bye
 Sven

--- os/exec.c	Wed Aug 25 16:21:16 1999
+++ Src/exec.c	Thu Aug 26 09:14:13 1999
@@ -615,16 +615,12 @@
 int
 isreallycom(Cmdnam cn)
 {
-    if (cn->flags & HASHED)
-	return 1;
-    else {
-	char fullnam[MAXCMDLEN];
+    char fullnam[MAXCMDLEN];
 
-	strcpy(fullnam, cn->u.name ? *(cn->u.name) : "");
-	strcat(fullnam, "/");
-	strcat(fullnam, cn->nam);
-	return iscom(fullnam);
-    }
+    strcpy(fullnam, cn->u.name ? *(cn->u.name) : "");
+    strcat(fullnam, "/");
+    strcat(fullnam, cn->nam);
+    return iscom(fullnam);
 }
 
 /**/
@@ -1689,7 +1685,8 @@
 	    char *cmdarg = (char *) peekfirst(args);
 
 	    hn = cmdnamtab->getnode(cmdnamtab, cmdarg);
-	    if (hn && trycd && !isreallycom((Cmdnam)hn)) {
+	    if (hn && trycd && !(((Cmdnam)hn)->flags & HASHED) &&
+		!isreallycom((Cmdnam)hn)) {
 		cmdnamtab->removenode(cmdnamtab, cmdarg);
 		cmdnamtab->freenode(hn);
 		hn = NULL;

--
Sven Wischnowsky                         wischnow@xxxxxxxxxxxxxxxxxxxxxxx