Comment 2 for bug 308949
Revision history for this message
| Alastair Bridgewater (alastair-bridgewater) wrote | #2 |
© 2004
Canonical Ltd.
•
Terms of use
•
Data privacy
•
Contact Launchpad Support
•
Blog
•
Careers
•
System status
•
f33e652
(Get the code!)
Juho Snellman committed a fix for this on x86oid (gencgc) systems as 1.0.3.11.
I have tried the "new test case", with random interruptions, on both an x86-linux and an x86-64-linux build post-1.0.27, and it was perfectly reliable, suggesting to me that the test case is more of a threading stress test than anything else.
On unicode builds, we will usually have to allocate a fresh copy of the string anyway just for external-format handling. I was unable to determine if this is non-heap space or not through a quick browse of the source, but it could be a usable approach.