zuul
/
zuul
Code Issues Proposed changes

project templating system 

On setup where Zuul ends up triggering hundreds of projects, you end up
having projects using roughly the same pipeline/jobs.  Whenever one want
to add a job in all the similiar project, he has to edit each project
one by one.

To save some precious time, this patch introduces the concept of project
templates.  It lets you define a set of pipeline and attached jobs
though the job names can be passed parameters defined on a per project
basis.  Thus, updating similiar projects is all about editing a single
template.

A basic example is provided in the documentation.

The voluptuous schema has been updated. It does check whether all
parameters are properly passed to a template but does NOT check whether
the resulting job name exist.

The parameter expansion in templates is borrowed from Jenkins Job
Builder (deep_format function). It has been tweaked to also expand
dictionary keys.

Layout test plan:

  $ nosetests -m layout  --nocapture
  Test layout file validation ...
  <...>
  bad_template1.yaml
     required key not provided @
  data['projects'][0]['template']['project']
  bad_template2.yaml
     extra keys not allowed @
  data['projects'][0]['template']['extraparam']
  good_template1.yaml
  ok
  <...>
  $

A basic test hasbeen added to verify whether a project-template properly
triggers its tests:

  $ nosetests --nocapture \
  tests/test_scheduler.py:testScheduler.test_job_from_templates_launched
  Test whether a job generated via a template can be launched ... ok

  ----------------------------------------------------------------------
  Ran 1 test in 0.863s

  OK
 $

Change-Id: Ib82e4719331c204de87fbb4b20c198842b7e32f4
Reviewed-on: https://review.openstack.org/21881
Reviewed-by: Jeremy Stanley <fungi@yuggoth.org>
Reviewed-by: James E. Blair <corvus@inaugust.com>
Approved: James E. Blair <corvus@inaugust.com>
Tested-by: Jenkins
This commit is contained in:
Antoine Musso 2013-02-13 15:37:53 +01:00 committed by Jenkins
parent 3c5e5b507f
commit 80edd5a8fa
9 changed files with 254 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

34
doc/source/zuul.rst
View File

@ -504,6 +504,40 @@ can help avoid running unnecessary jobs.
.. seealso:: The OpenStack Zuul configuration for a comprehensive example: https://github.com/openstack-infra/config/blob/master/modules/openstack_project/files/zuul/layout.yaml
Project Templates
"""""""""""""""""
Whenever you have lot of similiar projects (such as plugins for a project) you
will most probably want to use the same pipeline configurations. The
project templates let you define pipelines and job name templates to trigger.
One can then just apply the template on its project which make it easier to
update several similiar projects. As an example::
project-templates:
# Name of the template
- name: plugin-triggering
# Definition of pipelines just like for a `project`
check:
- '{jobprefix}-merge':
- '{jobprefix}-pep8'
- '{jobprefix}-pyflakes'
gate:
- '{jobprefix}-merge':
- '{jobprefix}-unittest'
- '{jobprefix}-pep8'
- '{jobprefix}-pyflakes'
In your projects definition, you will then apply the template using the template
key::
projects:
- name: plugin/foobar
template:
- name: plugin-triggering
jobprefix: plugin-foobar
You can pass several parameters to a template. A ``parameter`` value will be
used for expansion of ``{parameter}`` in the template strings.
logging.conf
~~~~~~~~~~~~

11
tests/fixtures/layout.yaml vendored
View File

@ -47,6 +47,12 @@ jobs:
files:
- '.*-requires'
project-templates:
- name: test-one-and-two
check:
- '{projectname}-test1'
- '{projectname}-test2'
projects:
- name: org/project
merge-mode: cherry-pick
@ -124,3 +130,8 @@ projects:
- nonvoting-project-test2
post:
- nonvoting-project-post
- name: org/templated-project
template:
- name: test-one-and-two
projectname: project

19
tests/fixtures/layouts/bad_template1.yaml vendored Normal file
View File

@ -0,0 +1,19 @@
# Template is going to be called but missing a parameter
pipelines:
- name: 'check'
manager: IndependentPipelineManager
trigger:
- event: patchset-created
project-templates:
- name: template-generic
check:
# Template uses the 'project' parameter' which must
- '{project}-merge'
projects:
- name: organization/project
template:
- name: template-generic
# Here we 'forgot' to pass 'project'

22
tests/fixtures/layouts/bad_template2.yaml vendored Normal file
View File

@ -0,0 +1,22 @@
# Template is going to be called with an extra parameter
pipelines:
- name: 'check'
manager: IndependentPipelineManager
trigger:
- event: patchset-created
project-templates:
- name: template-generic
check:
# Template only uses the 'project' parameter'
- '{project}-merge'
projects:
- name: organization/project
template:
- name: template-generic
project: 'MyProjectName'
# Feed an extra parameters which is not going to be used
# by the template. That is an error.
extraparam: 'IShouldNotBeSet'

13
tests/fixtures/layouts/bad_template3.yaml vendored Normal file
View File

@ -0,0 +1,13 @@
# Template refers to an unexisting pipeline
pipelines:
# We have no pipelines at all
project-templates:
- name: template-generic
unexisting-pipeline: # pipeline does not exist
projects:
- name: organization/project
template:
- name: template-generic

16
tests/fixtures/layouts/good_template1.yaml vendored Normal file
View File

@ -0,0 +1,16 @@
pipelines:
- name: 'check'
manager: IndependentPipelineManager
trigger:
- event: patchset-created
project-templates:
- name: template-generic
check:
- '{project}-merge'
projects:
- name: organization/project
template:
- name: template-generic
project: 'myproject'

15
tests/test_scheduler.py
View File

@ -689,6 +689,7 @@ class testScheduler(unittest.TestCase):
init_repo("org/project3")
init_repo("org/one-job-project")
init_repo("org/nonvoting-project")
init_repo("org/templated-project")
self.config = CONFIG
self.statsd = FakeStatsd()
@ -1376,6 +1377,20 @@ class testScheduler(unittest.TestCase):
assert B.reported == 2
self.assertEmptyQueues()
def test_job_from_templates_launched(self):
"Test whether a job generated via a template can be launched"
A = self.fake_gerrit.addFakeChange(
'org/templated-project', 'master', 'A')
self.fake_gerrit.addEvent(A.getPatchsetCreatedEvent(1))
self.waitUntilSettled()
jobs = self.fake_jenkins.job_history
job_names = [x.name for x in jobs]
assert 'project-test1' in job_names
assert 'project-test2' in job_names
assert jobs[0].result == 'SUCCESS'
assert jobs[1].result == 'SUCCESS'
def test_dependent_changes_dequeue(self):
"Test that dependent patches are not needlessly tested"
A = self.fake_gerrit.addFakeChange('org/project', 'master', 'A')

78
zuul/layoutvalidator.py
View File

@ -1,4 +1,6 @@
# Copyright 2013 OpenStack Foundation
# Copyright 2013 Antoine "hashar" Musso
# Copyright 2013 Wikimedia Foundation Inc.
#
# Licensed under the Apache License, Version 2.0 (the "License"); you may
# not use this file except in compliance with the License. You may obtain
@ -13,6 +15,7 @@
# under the License.
import voluptuous as v
import string
# Several forms accept either a single item or a list, this makes
@ -55,6 +58,9 @@ class LayoutSchema(object):
}
pipelines = [pipeline]
project_template = {v.Required('name'): str}
project_templates = [project_template]
job = {v.Required('name'): str,
'failure-message': str,
'success-message': str,
@ -80,21 +86,90 @@ class LayoutSchema(object):
else:
self.job_name.validate(path, self.job_name.schema, value)
def validateTemplateCalls(self, calls):
""" Verify a project pass the parameters required
by a project-template
"""
for call in calls:
schema = self.templates_schemas[call.get('name')]
schema(call)
def collectFormatParam(self, tree):
"""In a nested tree of string, dict and list, find out any named
parameters that might be used by str.format(). This is used to find
out whether projects are passing all the required parameters when
using a project template.
Returns a set() of all the named parameters found.
"""
parameters = set()
if isinstance(tree, str):
# parse() returns a tuple of
# (literal_text, field_name, format_spec, conversion)
# We are just looking for field_name
parameters = set([t[1] for t in string.Formatter().parse(tree)
if t[1] is not None])
elif isinstance(tree, list):
for item in tree:
parameters.update(self.collectFormatParam(item))
elif isinstance(tree, dict):
for item in tree:
parameters.update(self.collectFormatParam(tree[item]))
return parameters
def getSchema(self, data):
pipelines = data.get('pipelines')
if not pipelines:
pipelines = []
pipelines = [p['name'] for p in pipelines if 'name' in p]
# Whenever a project uses a template, it better have to exist
project_templates = data.get('project-templates', [])
template_names = [t['name'] for t in project_templates
if 'name' in t]
# A project using a template must pass all parameters to it.
# We first collect each templates parameters and craft a new
# schema for each of the template. That will later be used
# by validateTemplateCalls().
self.templates_schemas = {}
for t_name in template_names:
# Find out the parameters used inside each templates:
template = [t for t in project_templates
if t['name'] == t_name]
template_parameters = self.collectFormatParam(template)
# Craft the templates schemas
schema = {v.Required('name'): v.Any(*template_names)}
for required_param in template_parameters:
# add this template parameters as requirements:
schema.update({v.Required(required_param): str})
# Register the schema for validateTemplateCalls()
self.templates_schemas[t_name] = v.Schema(schema)
project = {'name': str,
'merge-mode': v.Any('cherry-pick'),
'template': self.validateTemplateCalls,
}
# And project should refers to existing pipelines
for p in pipelines:
project[p] = self.validateJob
projects = [project]
# Sub schema to validate a project template has existing
# pipelines and jobs.
project_template = {'name': str}
for p in pipelines:
project_template[p] = self.validateJob
project_templates = [project_template]
# Gather our sub schemas
schema = v.Schema({'includes': self.includes,
v.Required('pipelines'): self.pipelines,
'jobs': self.jobs,
'project-templates': project_templates,
v.Required('projects'): projects,
})
return schema
@ -116,3 +191,6 @@ class LayoutValidator(object):
if 'jobs' in data:
self.checkDuplicateNames(data['jobs'], ['jobs'])
self.checkDuplicateNames(data['projects'], ['projects'])
if 'project-templates' in data:
self.checkDuplicateNames(
data['project-templates'], ['project-templates'])

46
zuul/scheduler.py
View File

@ -1,5 +1,7 @@
# Copyright 2012 Hewlett-Packard Development Company, L.P.
# Copyright 2013 OpenStack Foundation
# Copyright 2013 Antoine "hashar" Musso
# Copyright 2013 Wikimedia Foundation Inc.
#
# Licensed under the Apache License, Version 2.0 (the "License"); you may
# not use this file except in compliance with the License. You may obtain
@ -32,6 +34,28 @@ import merger
statsd = extras.try_import('statsd.statsd')
def deep_format(obj, paramdict):
"""Apply the paramdict via str.format() to all string objects found within
the supplied obj. Lists and dicts are traversed recursively.
Borrowed from Jenkins Job Builder project"""
if isinstance(obj, str):
ret = obj.format(**paramdict)
elif isinstance(obj, list):
ret = []
for item in obj:
ret.append(deep_format(item, paramdict))
elif isinstance(obj, dict):
ret = {}
for item in obj:
exp_item = item.format(**paramdict)
ret[exp_item] = deep_format(obj[item], paramdict)
else:
ret = obj
return ret
class Scheduler(threading.Thread):
log = logging.getLogger("zuul.Scheduler")
@ -55,6 +79,7 @@ class Scheduler(threading.Thread):
self.pipelines = {}
self.jobs = {}
self.projects = {}
self.project_templates = {}
self.metajobs = {}
def stop(self):
@ -126,6 +151,16 @@ class Scheduler(threading.Thread):
toList(trigger.get('email_filter')))
manager.event_filters.append(f)
for project_template in data.get('project-templates', []):
# Make sure the template only contains valid pipelines
tpl = dict(
(pipe_name, project_template.get(pipe_name))
for pipe_name in self.pipelines.keys()
if pipe_name in project_template
)
self.project_templates[project_template.get('name')] \
= tpl
for config_job in data.get('jobs', []):
job = self.getJob(config_job['name'])
# Be careful to only set attributes explicitly present on
@ -177,6 +212,17 @@ class Scheduler(threading.Thread):
for config_project in data.get('projects', []):
project = Project(config_project['name'])
for requested_template in config_project.get('template', []):
# Fetch the template from 'project-templates'
tpl = self.project_templates.get(
requested_template.get('name'))
# Expand it with the project context
expanded = deep_format(tpl, requested_template)
# Finally merge the expansion with whatever has been already
# defined for this project
config_project.update(expanded)
self.projects[config_project['name']] = project
mode = config_project.get('merge-mode')
if mode and mode == 'cherry-pick':