I've checked in the attached patch which fixes a core dump (in the
child process) when a zero length command array is passed.
2005-03-11 Archie Cobbs <[EMAIL PROTECTED]>
* native/jni/java-lang/java_lang_VMProcess.c: Fix segfault
in the case that a zero length command array is passed.
-Archie
__________________________________________________________________________
Archie Cobbs * CTO, Awarix * http://www.awarix.com
Index: native/jni/java-lang/java_lang_VMProcess.c
===================================================================
RCS file:
/cvsroot/classpath/classpath/native/jni/java-lang/java_lang_VMProcess.c,v
retrieving revision 1.3
diff -u -r1.3 java_lang_VMProcess.c
--- native/jni/java-lang/java_lang_VMProcess.c 21 Dec 2004 16:17:07 -0000
1.3
+++ native/jni/java-lang/java_lang_VMProcess.c 12 Mar 2005 03:47:22 -0000
@@ -177,6 +177,8 @@
* directory into a single array for simplicity of (de)allocation.
*/
cmdArrayLen = (*env)->GetArrayLength(env, cmdArray);
+ if (cmdArrayLen == 0)
+ goto null_pointer_exception;
if (envArray != NULL)
envArrayLen = (*env)->GetArrayLength(env, envArray);
if ((strings = malloc(((cmdArrayLen + 1)
_______________________________________________
Classpath-patches mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/classpath-patches
