Re: The pattern containing NUL.

[Peter Stephenson <pws@xxxxxxxxxxxxxxxxx>]

Tanaka Akira wrote:
> zsh-3.1.6-test-1 dumps core as following on NetBSD 1.4.
> 
> Z(2):akr@localhost% Src/zsh -f
> localhost% print $OSTYPE 
> netbsd1.4
> localhost% print $ZSH_VERSION 
> 3.1.6-test-1
> localhost% tmp=a
> localhost% print "${line#$'\0'}"
> zsh: segmentation fault (core dumped)  Src/zsh -f
> Z(2):akr@localhost% 
> 
> The backtrace was follows.

Well, I can't get the core dump and don't entirely believe the backtrace,
but there does seem to be a bug... does this fix it?

We really need a pattern matching expert to fix up the globbing code.  It's
slow and complicated (all right, it would look out of place if wasn't).

--- Src/glob.c.null	Sun Jul 18 17:37:35 1999
+++ Src/glob.c	Mon Jul 19 10:03:42 1999
@@ -2724,6 +2724,9 @@
      * Here we bypass tulower() and tuupper() for speed.
      */
     int xi = (STOUC(UNMETA(x)) & 0xff), yi = (STOUC(UNMETA(y)) & 0xff);
+    /* A NULL is a real null, since a \000 would be metafied. */
+    if (!*x || !*y)
+	return 0;
     return xi == yi ||
 	(((c->stat & C_IGNCASE) ?
 	  ((isupper(xi) ? tolower(xi) : xi) ==

-- 
Peter Stephenson <pws@xxxxxxxxxxxxxxxxx>       Tel: +39 050 844536
WWW:  http://www.ifh.de/~pws/
Dipartimento di Fisica, Via Buonarroti 2, 56127 Pisa, Italy