> Date: Mon, 01 Oct 2007 22:34:30 +0200
> From: Eli Zaretskii <[EMAIL PROTECTED]>
> Cc: [email protected]
>
> It's valid C, allright, but the code clearly wants not to modify the
> `name' argument to the function, and the function's prototype promises
> that it won't be modified. So modifying it would be a breach of
> contract between the function and its callers. If, for example, at
> some point, `find_directory' is passed a constant string literal as
> its argument, Make could crash at run time.
>
> I will suggest an alternative patch in a while.
Here's my suggestion. Could Yongwei Wu please test this and see if it
works?
2007-10-01 Eli Zaretskii <[EMAIL PROTECTED]>
* dir.c (find_directory) [WINDOWS32]: Don't modify the const
string pointed by p.
--- dir.c~0 2007-07-21 19:06:58.251125000 +0300
+++ dir.c 2007-10-01 22:57:29.531250000 +0200
@@ -453,27 +453,24 @@ find_directory (const char *name)
hash_insert_at (&directories, dir, dir_slot);
/* The directory is not in the name hash table.
Find its device and inode numbers, and look it up by them. */
-
-#ifdef WINDOWS32
- /* Remove any trailing '\'. Windows32 stat fails even on valid
- directories if they end in '\'. */
- if (p[-1] == '\\')
- p[-1] = '\0';
-#endif
-
#ifdef VMS
r = vmsstat_dir (name, &st);
+#elif defined(WINDOWS32)
+ {
+ char tem[MAXPATHLEN], *t;
+
+ /* Remove any trailing slashes. Windows32 stat fails even on
+ valid directories if they end in a slash. */
+ memcpy (tem, name, p - name + 1);
+ t = tem + (p - name - 1);
+ if (*t == '/' || *t == '\\')
+ *t = '\0';
+ r = stat (tem, &st);
+ }
#else
EINTRLOOP (r, stat (name, &st));
#endif
-#ifdef WINDOWS32
- /* Put back the trailing '\'. If we don't, we're permanently
- truncating the value! */
- if (p[-1] == '\0')
- p[-1] = '\\';
-#endif
-
if (r < 0)
{
/* Couldn't stat the directory. Mark this by
_______________________________________________
Make-w32 mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/make-w32
