Excellent!
Thank you for originally raising this issue; I think you were first to
stumble on this unfortunate bug. Because you raised it here, and
iterated like crazy to narrow it down, we were able to eventually
track it down & resolve it. Lucene progress would not happen without
such active/involved users like you!
Mike
Ian Lea wrote:
Excellent. And to confirm - my always-fails index load runs to
successful completion on this release.
Thanks Mike.
--
Ian.
On Wed, Jul 30, 2008 at 7:09 PM, Michael McCandless
<[EMAIL PROTECTED]> wrote:
FYI -- there is a nasty bug that affects Lucene in Sun's 1.6 hotspot
compiler, starting with 1.6.0_04. At least 3 known cases have been
seen on this list.
Details are here:
https://issues.apache.org/jira/browse/LUCENE-1282
The bug causes silent index corruption during merging, such that the
merge completes successfully but in fact has introduced corruption
into the index that goes undetected until you open an IndexReader on
it.
Sun has now fixed the bug as of build 28 of the 1.6 JRE update 10
beta
(1.6.0_10-rc-b28), here:
http://download.java.net/jdk6/binaries
If you are using 1.6.0_04 or higher, you should consider upgrading to
this JRE release, or, upgrading once it's an official release (it's
beta
now)!
Mike
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
