Here's Nigel's original blog post describing the automatic patch testing setup for Hadoop: http://developer.yahoo.net/blogs/hadoop/2007/12/if_it_hurts_automate_it_1.html
All the Jira state changes are sent to the Hadoop mailing list. A daemon on the build machine is subscribed to the mailing list and processes the mail using the processHadoopPatchEmail.sh script. The patch "queue" created by that script is administered by the Hadoop-Patch-Admin build which runs the hudsonPatchQueueAdmin.sh script every few minutes. When there is a patch to test, the Hadoop-Patch build is kicked off running the hudsonBuildHadoopPatch.sh script.
I'm planning to adapt Pig's test-patch.sh script and their test-patch specific ant targets to Hive, but it's clear we'll need to add some new Hudson jobs similar to the ones that Pig has, e.g. Pig-Patch-Admin and Pig-Patch-h7.grid.sp2. Can someone with Hudson admin access please take a look at these jobs and try to determine whether or not it's feasible to set this up for Hive?