| Release notes for Gerrit 2.4.1 | |
| ============================== | |
| Gerrit 2.4.1 is now available: | |
| link:http://code.google.com/p/gerrit/downloads/detail?name=gerrit-2.4.1.war[http://code.google.com/p/gerrit/downloads/detail?name=gerrit-2.4.1.war] | |
| There are no schema changes from 2.4. However, if upgrading from | |
| anything but 2.4, follow the upgrade procedure in the 2.4 | |
| link:ReleaseNotes-2.4.html[ReleaseNotes]. | |
| Bug Fixes | |
| --------- | |
| * Catch all exceptions when async emailing | |
| + | |
| This fixes email notification issues reported | |
| link:https://groups.google.com/group/repo-discuss/browse_thread/thread/dd157ebc55b962ef/652822d6fbe61e71[here]. | |
| * Fixed cleanup of propagated SshScopes | |
| + | |
| This improves error reporting in case of email notification errors. | |
| * issue 1394 Fix lookup of the 'Commit Message' file in patch set | |
| + | |
| There is an assumption that the commit message is always first in the list of | |
| files of a patch set. However, there was another place in Gerrit code, which | |
| did binary search through the list of the files, without taking this assumption | |
| into account. In case when a patch set contained a file which lexicographically | |
| sorted before '/COMMIT_MSG' (like '.gitignore' for example) it could have | |
| happened that the commit message was not found and, as a side effect, it wasn't | |
| possible to review it. | |
| * issue 1162 Fix deadlock on destroy of CommandFactoryProvider | |
| * Honor the sendmail.smtpUser from gerrit.config on upgrade | |
| + | |
| If sendmail.smtpUser was not present in the gerrit.config then don't set it in | |
| site upgrade. | |
| * issue 1420 Forge committer bypassed | |
| + | |
| It was possible to forge committer even without having permission for that. | |
| This was a regression from 2.3. | |
| * Make sure the "Object too large..." error message is printed when an object | |
| larger than receive.maxObjectSizeLimit is rejected by Gerrit | |
| * Display proper error if file diff fails because content is too large | |
| * Get around a log4j bug that causes AsyncAppender-Dispatcher thread to die and | |
| block other threads | |
| ** Make async logging buffer size configurable | |
| ** Make logging events discardable, prevent NPE in AsyncAppender-Dispatcher thread |