In other areas of Stapler you can separately specify the 'class' that the field
should be searched, so you might want to see if there is an annotation that
would do that.
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
At: Apr 22 2014
I'd be quite surprised to find out that nothing like this is already available
in the vast pool of Java libraries on the Internet :-)
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
At: Mar 27 2014 10:25:42
Jerome,
Not sure about a
No, you need to do validation while the user is changing the configuration
itself.
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
At: Mar 7 2014 12:48:11
Hello,
Is there a way to cancel the save of a job configuration and output a
Compile (build) errors and test errors are already distinguished. A build
failure produces a red (failure) result, where a test failure produces a yellow
(unstable) result.
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
At: Mar 6 2014
Isn't this a question for the people who make Cobertura, not Jenkins?
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com, jenkinsci-us...@googlegroups.com
At: Feb 25 2014 23:53:43
hi all,
just wanted to know , does cobertura latest version
regards
irfan
On Wed, Feb 26, 2014 at 8:09 PM, Kevin Fleming (BLOOMBERG/ 731 LEXIN)
kpflem...@bloomberg.net wrote:
Isn't this a question for the people who make Cobertura, not Jenkins?
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
In Git, tags are in no way associated with branches, so there's no such thing
as 'tags from a specific branch'. It's possible to analyze all new tags that
are seen and determine if every commit reachable from a tag is also reachable
from that specific branch, but this is certainly not something
snide mode
For a continuous integration project, Jenkins seems to have quite a few
error-prone manual processes.
/snide mode
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
At: Jan 7 2014 14:11:44
1.546 has been released yesterday, but
You need world-level transactions :-)
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
At: Jan 7 2014 15:42:50
On 01/07/2014 11:14 AM, Kevin Fleming (BLOOMBERG/ 731 LEXIN) wrote:
snide mode
For a continuous integration project, Jenkins seems
When you say 'canonical' in this proposal, do you mean the repositories used
for making releases, or the repositories where development (and especially,
pull requests) would be handled?
If it's the former, I could see that being worthwhile, especially if *nobody*
has permissions to push to the
Gerrit.
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
At: Nov 12 2013 10:13:13
On 12 November 2013 14:40, Kevin Fleming (BLOOMBERG/ 731 LEXIN)
kpflem...@bloomberg.net wrote:
When you say 'canonical' in this proposal, do you mean
I have not deployed a DEB build system similar to Koji at this point, I'm still
investigating to decide what I want to use.
- Original Message -
From: vtu...@redhat.com
To: Kevin Fleming (BLOOMBERG/ 731 LEXIN)
Cc: jenkinsci-dev@googlegroups.com
At: Nov 5 2013 07:53:10
I am not very
I'd be interested in Jenkins/Koji integration, but I need to build DEBs as well
as RPMs, and there doesn't seem to be any support for DEBs in Koji.
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
At: Oct 31 2013 10:03:10
Hi,
if you don't
This might also be possible:
st:include page=whatever.jelly class=hudson.model.AbstractProject /
This insulates you from filesystem/URL path issues.
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
At: Oct 31 2013 11:30:17
On Thu, Oct 31,
Your message is a bit confusing, but it appears that you are asking for a
parameter option screen to be display after the job has started; this doesn't
really fit the Jenkins model, as the request to run a job is placed in a queue,
and when an executor becomes available the job is started.
Another issue: I converted a very simple class (four fields) over to using only
@DataBoundSetter, with no @DataBoundConstructor at all. When Jenkins tries to
instantiate this class, I get an error from Stapler:
NoStaplerConstructorException: There's no @DataBoundConstructor on any
constructor
://issues.jenkins-ci.org/browse/JENKINS-18629
/Domi
On 04.10.2013, at 20:15, Kohsuke Kawaguchi kkawagu...@cloudbees.com wrote:
On 10/04/2013 07:00 AM, nicolas de loof wrote:
maybe using @PostConstruct ?
2013/10/4 Kevin Fleming (BLOOMBERG/ 731 LEXIN) kpflem...@bloomberg.net
mailto:kpflem
How will the object know that Stapler has completed calling DataBoundSetters
(that all configuration data has been applied)?
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
At: Oct 4 2013 00:38:55
great,
so next step is to be able to
Nothing in the EC2 plugin sends credentials to the slave instances; they are
not needed for any normal slave operations. I cannot speak to the S3 plugin as
I don't use it.
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
At: Sep 17 2013
I reported this weeks ago, it's a simple thing to fix and has already been
fixed in later releases. The root cause is that the license for a package that
the POM pulls in has changed, and it has changed from 'nothing' to 'CDDL'. The
plugin is attempting to change the license *to* CDDL, and
Re-executing the JUnitParser is going to be very expensive for builds with a
large number of unit tests. When I've considered this sort of thing in the
past, I was pretty certain that it would require 'help' from the project
itself, so that the parser can be made aware of newly-available test
Could this be implemented on top of the Credentials plugin, rather than
standalone?
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
At: Aug 2 2013 06:31:33
Hi,
I have written a Jenkins plugin to manage keychains and provisioning
You've taken on a rather large task :-)
Your approach of providing a mock AmazonEC2 seems sound, but there shouldn't be
any need to mock AmazonEC2Cloud as well; instead, AmazonEC2Cloud (or EC2Cloud
underneath it) should have a way for the test harness to override the class
that is used to
The best approach for supporting Windows slaves on EC2 is indeed to use SSH;
it's the simplest method of launching slaves. If you don't want to go that
route, the next option would be to use JNLP to launch the slave, and pass the
Jenkins node name for the slave through the EC2 'user data'
Today the ec2-plugin has no way to set environment variables on slaves that it
launches; in my usage of it, I do that sort of thing in the jobs that are run
on the slaves, not globally on the slave itself.
I can see how what you are proposing would be useful, but I'd be hesitant to
create an
It's certainly possible, yes. That's a good strategy anyway, if the
functionality becomes popular then it can be merged into the core if desired.
Whether it can be done by a plugin will depend primarily on whether the Node
management pages in the core were designed with extension points for
Oh, one more thought: it looks like it would be possible to use WMI (from Java
using the J-Interop DCOM library) to make connections to Windows computers and
start Jenkins slaves on them. This would allow master-initiated connections
(which are better than JNLP connections as nothing needs to
Interesting! I hadn't gotten far enough in my exploration of Windows on EC2 to
have discovered that :-)
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
At: Jul 24 2013 17:23:11
On Wed, Jul 24, 2013 at 5:18 PM, Kevin Fleming (BLOOMBERG/ 731
In the plugin I'm working on, I need to present a list of choices to the user.
The user will be able to select one or more of them (thus checkboxes would be
appropriate, not radio buttons). Most likely I'll store the data in a
ListString or something similar.
The tricky part is that the list
All of your class fields on the form either need to be public or have getFoo()
methods to retrieve them, or the form building code won't be able to read them.
---
Sent from Bloomberg Anywhere for Android
- Original Message -
From: Adam Retter jenkinsci-dev@googlegroups.com
At: Tuesday,
on the doFill... Methods
On Friday, 12 July 2013, Kevin Fleming (BLOOMBERG/ 731 LEXIN) wrote:
Continuing my development of a replacement AWS plugin...
I've got a class that extends Cloud (called EC2), so it can be added and
configured from the global config page (it appears in an f:hetero-list). I
I'm sure I'm missing something obvious, but I can't figure out how to generate
the Javadoc for the Jenkins API from a Git clone of the core repository. I'd
like to have an offline-readable copy of the Javadoc, and the Jenkins site
doesn't appear to have a ZIP or JAR of it available, so I'd be
I'm working on a rewrite of the EC2 plugin, and as part of it I'm going to
leverage the SSH Slaves plugin rather than having SSH connection and management
logic directly in the new plugin.
To do this, I've got a class, SSHConnectionMethod, which has a field that is of
type
.
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
At: Jul 12 2013 09:51:54
On 07/12/2013 09:41 AM, Kevin Fleming (BLOOMBERG/ 731 LEXIN) wrote:
st:include class=hudson.plugins.sshslaves.SSHConnector page=config.jelly
it=${it.connector}/
Unfortunately
-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
At: Jul 12 2013 10:47:28
On 07/12/2013 09:53 AM, Kevin Fleming (BLOOMBERG/ 731 LEXIN) wrote:
Interesting idea, but I'm not sure that will work when 'connector' is null
because the object is being created by the user of the config form
instance itself.
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
At: Jul 12 2013 10:47:28
On 07/12/2013 09:53 AM, Kevin Fleming (BLOOMBERG/ 731 LEXIN) wrote:
Interesting idea, but I'm not sure that will work when 'connector' is null
because
:53 AM, Kevin Fleming (BLOOMBERG/ 731 LEXIN) wrote:
Interesting idea, but I'm not sure that will work when 'connector' is null
because the object is being created by the user of the config form.
Usually in this context ${descriptor} is already set by the control that lets
the user add an item
Continuing my development of a replacement AWS plugin...
I've got a class that extends Cloud (called EC2), so it can be added and
configured from the global config page (it appears in an f:hetero-list). I have
another class which is Describable, that EC2 keeps a list of in a
DescribableList,
Assuming the regularly scheduled meeting is going to happen today, can we add
my changes to ReconfigurableDescribable (link below) to the agenda? I'd like to
get some idea whether this is an appropriate approach or not, as I'm building a
plugin that needs this mechanism.
Nobody on this mailing list is under any obligation to help you, nearly
everyone here participates as a volunteer to the project.
If you need a stable, production ready version of Jenkins, you should be using
an LTS release, not a weekly release. If you want to run weekly releases, you
need to
{https://github.com/kpfleming/jenkins-core/tree/reconfigurables}
After taking suggestions from Jesse and Stephen, I've put together a small
patch that allows DescribableList to properly reconfigure instances of
ReconfigurableDescribable, even if the hetero-list contains multiple instances
from
You could also build a local version of Jenkins that doesn't bundle the
Subversion plugin, then reference that version (using a special version string)
when bundling the custom plugin.
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
At: Jul
Since I have a production system that relies heavily on the EC2 plugin, I'm
considering a massive cleanup/reorganization of it, to take advantage of new
things that Jenkins offers now (SSH Credentials plugin and other stuff). Along
the way, I plan to introduce, where needed, some locking to
OpenJDK 6 is fine, that's not the issue.
Running the Jenkins source code tests is entirely independent of having an
installed Jenkins available. I would start by removing all the stuff you have
added to the PATH, removing the extra environment variables (other than
JAVA_HOME), and trying
Prior to OpenJDK 7, the Oracle and Open JDKs differed substantially; as of
OpenJDK 7, they are essentially the same thing. Still, I haven't had any
trouble with Jenkins-related stuff on OpenJDK 6.
- Original Message -
From: jenkinsci-dev@googlegroups.com
To:
Following up on the thread from last week (or the week before), I'm still
investigating how to properly implement ReconfigurableDescribable objects (that
live in DescribableLists) so that they can actually be reconfigured properly.
After a tip from Jesse, I'm looking at hetero-list.jelly; it
Not really, no. If your tests need to wait for an external event, you should
break up the job into multiple phases, and trigger the later phases when you
know the external event has occurred.
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
You can also do something like this:
st:include page=ajaxBackupResults.jelly class=BackupLink
You'll probably need to specify the fully-qualified name of the class that owns
the Jelly file.
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
At:
After Stephen's hint yesterday, I've been exploring implementing this interface
in ec2-plugin, because many of the Describable objects in that plugin have
transient fields that contain references to data in EC2 (open sessions,
references to other objects, etc.) and replacing those objects when
:57, Kevin Fleming (BLOOMBERG/ 731 LEXIN)
kpflem...@bloomberg.net wrote:
After Stephen's hint yesterday, I've been exploring implementing this interface
in ec2-plugin, because many of the Describable objects in that plugin have
transient fields that contain references to data in EC2 (open
:-)
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
At: Jun 18 2013 11:37:57
On 18 June 2013 15:57, Kevin Fleming (BLOOMBERG/ 731 LEXIN)
kpflem...@bloomberg.net wrote:
After Stephen's hint yesterday, I've been exploring implementing this interface
in ec2
that actually renders the instances in
Jenkins.clouds into the hetero-list... more exploration ahead I guess.
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
At: Jun 18 2013 16:14:12
On 06/18/2013 02:24 PM, Kevin Fleming (BLOOMBERG/ 731 LEXIN
2013, Kevin Fleming (BLOOMBERG/ 731 LEXIN) wrote:
(sorry in advance for the length of this message, but there are a lot of
details)
Francis Upton and I have been trying to work through some data persistence
issues in the EC2 plugin; while some of what we need to know can be determined
through
salvation)
On Friday, 14 June 2013, Kevin Fleming (BLOOMBERG/ 731 LEXIN) wrote:
(sorry in advance for the length of this message, but there are a lot of
details)
Francis Upton and I have been trying to work through some data persistence
issues in the EC2 plugin; while some of what we need to know
Ahh... makes sense.
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
At: Jun 3 2013 13:41:17
On 06/03/2013 01:27 PM, Kevin Fleming (BLOOMBERG/ 731 LEXIN) wrote:
Any configuration data persisted under the old class names would need
The data, at that point, exists only in memory in the user's browser. Any
validation prior to form submission would need to be done by JavaScript
(presumably) running in the web page, although it could make separate calls
back to code on the server if it wished to.
- Original Message -
There are multiple port numbers emitted during the Jenkins startup (HTTP, JNLP,
maybe others). You need to be sure you are connecting to the HTTP port.
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
Cc: Kevin Fleming (BLOOMBERG/ 731 LEXIN
Pinning is what overrides it; if you unpin it you can upgrade it yourself.
Also, if you are developing a plugin, you don't need to pin the Jenkins version
to 1.480, you can just ensure that your plugin's POM references the 1.480
plugin API so that your plugin will stay compatible with 1.480.
Some of the points you raise were talked about (briefly) in the review of the
spot instance support code; as you have noted, they are primarily driven by the
nature of spot instances.
Using SSH to contact the spot slaves would be quite difficult without
repeatedly polling the spot request to
I would love to do this, but it will be hard to find the time in the next few
months. I'd be happy to be part of the design/review team though. I've been
wanting to rewrite the EC2 plugin anyway, this would be a great way to do that
:-)
- Original Message -
From:
Interesting... the JClouds plugin unfortunately doesn't seem to be as widely
used as the EC2 plugin, and as a result all the new features have been added to
the EC2 plugin only. When I tried the JClouds plugin months ago (trying to
decide which plugin to use for my EC2 slaves), there was really
That's an interesting question. The GitHub plugin doesn't provide a Jenkins
SCM, but it does read the Git SCM URls from projects to interact with GitHub.
Doing what you are talking about would be a bit complex, as logically the
'alternate location' URL (that points to GitHub) would be
reason for mirroring is to improve the performance of git clone, you
could investigate the reference repository idea, and place a copy of the
repository on each slave agent.
Mark Waite
From: Kevin Fleming (BLOOMBERG/ 731 LEXIN) kpflem...@bloomberg.net
To: jenkinsci-dev
Possibly, although I don't know if that would work when the Multiple SCM plugin
is used and multiple Git repositories are used by the project.
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
Cc: Kevin Fleming (BLOOMBERG/ 731 LEXIN)
At: May 9
To: jenkinsci-dev@googlegroups.com
At: Mar 26 2013 14:10:54
On 03/26/2013 11:15 AM, Kevin Fleming (BLOOMBERG/ 731 LEXIN) wrote:
the EC2 plugin deserializes the EC2Slaves and they become Node objects in the
Jenkins node list.
implements EphemeralNode?
--
You received this message because you
Supposedly there is way to make the Git plugin (1.3.x or later) revert to using
command-line git instead of JGit for repository access, but I can't seem to
find any place to configure that.
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
Yeah... the appointed time for the governance meeting came and went without any
discussion.
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
At: Feb 22 2013 15:36:49
On 01/28/2013 04:27 AM, Vojtech Juranek wrote:
1.480.3-candidate (anyone
First, this isn't really a Jenkins development question, so it belongs on the
jenkinsci-users list instead :-)
You didn't mention which version of the EC2 plugin you are using; there have
been a number of attempts to reduce the chances of the plugin launching excess
slaves due to long startup
: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
At: Feb 5 2013 15:14:19
On 02/05/2013 03:03 PM, KEVIN FLEMING (BLOOMBERG/ 731 LEXIN) wrote:
the PlannedNode object is going to be destroyed
Rather, removed from pendingLaunches?
the solution was to know conclusively when
Yeah, that sounds a bit fishy. In addition, even if the plugin archive didn't
include the dependency and required the user to install it in their CLASSPATH
some other way (which would be ugly), it seems like such a license would also
infect derivative works (such as a plugin that uses the
Replying to my own question... Hudson.getInstance().save() does the trick.
- Original Message -
From: jenkinsci-dev@googlegroups.com
To: jenkinsci-dev@googlegroups.com
At: Jan 22 2013 16:22:47
I am working on a new feature in the EC2 plugin, and part of this feature
involves making a
I am working on a new feature in the EC2 plugin, and part of this feature
involves making a change to an existing SlaveTemplate object. SlaveTemplate
objects are owned by the EC2Cloud object, which inherit from the Jenkins Cloud
class, are configured/instantiated in the global configuration
72 matches
Mail list logo