Thanks for checking that Sian!
The issue seems to be that the GetSystemProperty() call on J9 returns a
0 length string when the bootclasspath property is not set. For some
reason when this gets concatenated to the rest of the bootclasspath
properties J9 cannot parse it properly (Im yet to understand why as the
property looks ok once the concatenation is done) and falls over.
A simple fix could be something like [1]. Does anyone object to me
applying this patch just to get J9 working again immediately? I assume
this doesn't break DRLVM?
Regards,
Oliver
[1]
Index: luniglob.c
===================================================================
--- luniglob.c (revision 608423)
+++ luniglob.c (working copy)
@@ -358,7 +358,7 @@
if (tokensScanned == 1)
{
char *oldPath = bootstrapClassPath;
- if (!oldPath) {
+ if ((!oldPath) || (strlen(bootstrapClassPath) == 0)) {
bootstrapClassPath = str_concat (PORTLIB,
bootDirectory, props[i].value, NULL);
} else {
Sian January wrote:
Hi Oliver,
I'm seeing this on Windows too with J9, but not with DRLVM. I get the
following error on the command line:
Unhandled exception
Type=Segmentation error vmState=0x0005ffff
Regards,
Sian
On 03/01/2008, Oliver Deakin <[EMAIL PROTECTED]> wrote:
Im currently seeing a crash running anything on classlib+J9 under
Windows (I havn't tried Linux). If I revert the patch applied to
luniglob.c at revision r607409 the issue is fixed. Is anyone else seeing
this?
Regards,
Oliver
--
Oliver Deakin
Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number
741598.
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU
--
Oliver Deakin
Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number 741598.
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU
