From: Archie Cobbs (archie@whistle.com)
Date: Fri Oct 09 1998 - 13:49:55 EDT
Tim,
Does this patch make sense to you to address James' point?
Index: external.c
===================================================================
RCS file: /home/cvspublic/kaffe/kaffe/kaffevm/external.c,v
retrieving revision 1.6
diff -u -r1.6 external.c
--- external.c 1998/09/11 15:38:33 1.6
+++ external.c 1998/10/09 17:51:12
@@ -88,6 +88,9 @@
int len;
lpath = (char*)Kaffe_JavaVMArgs[0].librarypath;
+ if (lpath == 0) {
+ lpath = getenv(LIBRARYPATH);
+ }
home = (char*)Kaffe_JavaVMArgs[0].classhome;
len = 0;
Thanks,
-Archie
> From kaffe@rufus.w3.org Thu Oct 8 16:23:44 1998
> Date: Thu, 08 Oct 1998 16:11:24 -0700
> From: James Lundblad <jamesl@eecs.berkeley.edu>
> Organization: EECS Dept., UC Berkeley
> To: "kaffe@rufus.w3.org" <kaffe@rufus.w3.org>
> CC: Christopher Hylands <cxh@eecs.berkeley.edu>
> Subject: initNative() kaffe 1.0b2
>
> I think it would be nice for compatibility reasons if initNative() (
> kaffe/kaffevm/external.c ) would use the LIBRARYPATH ( aka
> LD_LIBRARY_PATH ) ala 1.0b1 when both Kaffe_JavaVMArgs[0].librarypath
> and Kaffe_JavaVMArgs[0].classhome are uninitialize ( == 0 ).
> This allows JNI code that worked for 1.0b1 and the JDK to work without
> modification.
>
> Here are the diffs for kaffe/kaffevm/external.c (1.0b2):
>
> 99a100,103
> > if ((home == 0) && (lpath == 0)) {
> > ptr = getenv(LIBRARYPATH);
> > len += strlen(ptr);
> > }
> 116a121,123
> > if ((home == 0) && (lpath == 0) && (ptr != 0)) {
> > strcat(libraryPath,ptr);
> > }
>
> --
> James A. Lundblad Visiting Scholar?
> phone: 510-643-7612 558 Cory Hall
> fax: 510-642-2739 EECS Department
> email: jamesl@eecs.berkeley.edu University of California
> http://www.eecs.berkeley.edu/~jamesl Berkeley CA 94720-1772
___________________________________________________________________________
Archie Cobbs * Whistle Communications, Inc. * http://www.whistle.com
This archive was generated by hypermail 2b29 : Sat Sep 23 2000 - 19:57:00 EDT