The 2.6 series release notes contain important changes in this release series.
- HIGH Due to the way that email addresses with Unicode in the 'local part' are handled, it was possible to generate a password reset token for an email address and have it delivered to a separate email address with Unicode homoglyphs that normalized to the original email address.
- LOW Admin users could still access user reports after being suspended.
- Packages have been updated to the latest security versions.
- Migration data exported from GitHub Enterprise with
ghe-migrator did not include issue file attachments, which could cause imports to another server to fail.
- SAML reauthentication could fail if the SAML identity provider returned large headers in the authentication response.
- LDAP sync could fail on suspended users if restricted groups are not configured.
- Pushing Git LFS objects to a fork of a repository the user only has read access to would fail.
- PSD files stored in LFS failed to render.
- The settings would fail to be copied to the high availability replica if NTP has not been configured.
- SSH keys added or removed via the management console after high availability replication has started could fail to be copied to the replica.
- Hostnames that contain hyphens could not be used in the proxy exclusion list in the management console settings.
- Alambic services would not run on job-server cluster nodes.
- ElasticSearch on cluster nodes could enter a split-brain state in the event of a network partition or failure.
- Pre-receive hook environment variables were not all set on repository initialization. This could lead to pre-receive hooks running incorrectly on the first commit that takes place when creating a repository via a web browser. (updated 2016-07-13)
- Downloading identical user or repository reports in quick succession could lead to a build up in duplicate jobs that could affect the performance of the appliance.
- The automatic update check will only download the latest release that the appliance can upgrade directly to.
- We incorrectly redirect to the dashboard if you accessed GitHub Enterprise using an alias while in private mode. This might happen if you set a fully qualified domain name but the subdomain resolves correctly.
- Images uploaded to issues save with an absolute URL, so they can be broken if the hostname changes.
- On a freshly set up GitHub Enterprise without any users, an attacker could create the first admin user.
- Custom firewall rules aren't maintained during an upgrade.
- Enqueued background jobs are sometimes not purged when a repository is deleted.
- On instances upgraded from 2.3 and earlier, restoring a protected branch archived whilst running 2.3, will not restore all the settings correctly. This does not affect new instances or protected branches archived on later releases.
- Editing custom messages in the Admin Center doesn't provide emoji suggestions.
- Native emoji are lost when saving custom messages in the Admin Center.
- Repository push logs don't record whether a push was forced.
svn checkout may timeout while the repository data cache is being built. In most cases, subsequent
svn checkout attempts will succeed.
- Git LFS tracked files uploaded through the web interface are incorrectly added directly to the repository.
- Uploading PNG images with through the web interface can fail with the error 'Something went really wrong, and we can't process that file.'
- GitHub Enterprise clustering can not be configured without https. (updated 2016-08-01)
- Console text is difficult to read on OpenStack KVM. (updated 2016-08-03)
- The initial import of the VMware OVA image may fail when deployed via vCenter Server 6.0 or 6.5. The import will succeed when performed directly on an ESXi host. (updated 2017-02-23)
- Git LFS objects may take up to an hour to replicate in a High Availability configuration. (updated 2017-02-23)
The GitHub Team