- May 12, 2016
-
-
Oliver Gondža authored
[JENKINS-32777] Cover 'delete-builds' CLI by tests
-
- May 11, 2016
-
-
Daniel Beck authored
-
Kohsuke Kawaguchi authored
Conflicts: changelog.html
-
Kohsuke Kawaguchi authored
-
Kohsuke Kawaguchi authored
-
Kohsuke Kawaguchi authored
-
Jesse Glick authored
-
Ing. Pavel Janousek authored
-
Ing. Pavel Janousek authored
Slighty optimized and better legible asserts
-
Ing. Pavel Janousek authored
-
Oleg Nenashev authored
[JENKINS-34710] - PluginWrapper should not throw IOException if somebody enables the enabled plugin (#2327)
-
Alex Earl authored
* Remove CVS historical code - JENKINS-4610 * Fix whitespace issue
-
Pavel Janousek authored
reload-configuration covered by test-cases
-
pagio81 authored
More details: https://issues.jboss.org/browse/WFLY-4782
-
Baptiste Mathus authored
-
- May 10, 2016
-
-
Oliver Gondža authored
-
Oleg Nenashev authored
[JENKINS-34674,JENKINS-34675] - Handling of the default update site ID
-
Baptiste Mathus authored
As detailed by @stephenc: > So there is `getInstanceOrNull()` which is for use in Jenkins core or > in code that plugins add that may escape the Jenkins singleton > lifecycle. > > In regular plugins `getInstance()` will never return null as the > lifecycle does not instantiate them until after there is a Jenkins > singleton and the plugins will be stopped before there is no > singleton. > > The `getActiveInstance()` is therefore equivalent to `getInstance()` > and should never have been born. > > The only way a plugin can get a `null` from `Jenkins.getInstance()` is > if you install an `atExit()` handler, use a `PhantomReference` type > thing, or directly manipulate the servlet container in some way or > other. A correctly written plugin should unwind any of those things > when it is stopped or a method bound to the termination lifecycle. If > there are some cases (for which I cannot anticipate a good coder being > able to write a correct termination lifecycle method to fix) where the > code cannot be unhooked... then in those cases the plugin should use > `getInstanceOrNull()` and know how to respond to the `null` that could > not be avoided.
-
- May 09, 2016
-
-
Daniel Beck authored
[JENKINS-34328] CLI command disconnect-node extracted from core to CLI
-
Carlos Sanchez authored
[JENKINS-34683] - Do not print stack trace if plugin is missing dependencies
-
Kanstantsin Shautsou authored
Workflow users may want to catch specific Exceptions without stracktrace in future.
-
Manuel Recena authored
-
Daniel Beck authored
-
- May 08, 2016
-
-
Kohsuke Kawaguchi authored
-
Kohsuke Kawaguchi authored
-
Kohsuke Kawaguchi authored
-
Jean-Philippe Briend authored
* Add French translation on New Item screen * Add French translation to Manage Jenkins page * Add French translation to GlobalToolConfiguration page * Add missing French translation to Manage Jenkins page * Correct translations typos and PR remarks
-
Oleg Nenashev authored
-
Tom Fennelly authored
HttpSession listener ExtensionPoint
-
Oleg Nenashev authored
-
Oleg Nenashev authored
[FIXED JENKINS-34675] - Prevent hanging of the WebUI if the default Update Site ID cannot be resolved. Also hardens the Java code a bit.
-
- May 06, 2016
-
-
Daniel Beck authored
[JENKINS-34601] Fix help file for node name attribute
-
Daniel Beck authored
[FIX JENKINS-31611] Restrict access to plugin pages
-
Daniel Beck authored
-
- May 05, 2016
-
-
Daniel Beck authored
Lithuanian translation update
-
Pavel Janousek authored
-
Ing. Pavel Janousek authored
-
Manuel Recena authored
* [JENKINS-32936] Initial source code modifications * [JENKINS-32936] Three views are using this approach * [JENKINS-32936] Conflicts were resolved in a wrong way
-
Ing. Pavel Janousek authored
disconnect-node extended to accept multiple node names disconnect-node covered by test-cases
-