Skip to content
Commits on Jan 21, 2012
  1. @daniel-pittman

    Merge pull request #151 from joshcooper/ticket/master/12012-env-lang-…

    …windows
    
    (#12012) Fix ENV['LANG'] spec tests on Windows
    daniel-pittman committed Jan 21, 2012
  2. @joshcooper

    (#12012) Fix ENV['LANG'] spec tests on Windows

    Previously, the tests were failing because there isn't an `echo.exe`
    on Windows.
    
    This commit just wraps the command in a `cmd.exe` shell and expands
    the variables using `%LANG%` syntax on Windows.
    joshcooper committed Jan 21, 2012
  3. @daniel-pittman

    Merge pull request #150 from cprice-puppet/bug/master/12012-global-EN…

    …V-LANG-override
    
    (#12012) Fix ruby 1.8.5 incompatibility in new spec test
    daniel-pittman committed Jan 20, 2012
  4. @cprice404
Commits on Jan 20, 2012
  1. @daniel-pittman

    Merge pull request #149 from cprice-puppet/bug/master/12012-global-EN…

    …V-LANG-override
    
    #12012 - don't override lang in the environment globally
    daniel-pittman committed Jan 20, 2012
  2. @cprice404
  3. @cprice404

    (#12012) Remove global override of LANG environment variable

    'LANG' environment variable must be set to 'C' in order to reliably parse output of certain commands when building up core Facter facts (to prevent locale differences from affecting the format of the output).  However, it does not need to happen at a global-ish scope whenever facter is 'require'd;  move this behavior from the facter.rb into the local scope of the Facter::Util::Resolution.exec method; restore original value of environment variable after executing any individual command.
    
    Note that, while this behavior is better than the previous behavior (global override), it is still not possible for a custom facter fact to be written to expect any other locale besides 'C'.  Eventually might prefer to find a solution that allows callers to override this behavior if desired.
    cprice404 committed Jan 19, 2012
  4. @cprice404

    (#12012) Move "with_env" utility method from test code into lib code

    The utility method "with_env", which was defined locally in loader_spec.rb, is actually
    more generally useful (and will be used to complete the resolution of #12012).  Move it
    out of the test code and into Facter::Util::Resolution so that it is accessible to
    production (non-test) code.  Refactor loader_spec test accordingly.
    cprice404 committed Jan 19, 2012
Commits on Jan 17, 2012
  1. @kbarber

    Merge branch '1.6.x'

    * 1.6.x:
      (#11848) Don't hard code ruby install paths in Windows batch files
      Build a Rake task for building Apple Packages
      (#11559) Switch to RbConfig & Provide alias for RbConfig for pre-1.8.5
      (#10271) Identifying 'Amazon' using '/etc/system-release'
      (#11661) EC2 rspec tests were using throw not raise to simulate a timeout
      (#11583) Add basic coverage to the ec2 fact
      (#9599) Generalize zone detection
      (#9599) Add nexenta facts
      (#11583) Switch request method to open-uri monkey patch 'open'
    
    Conflicts:
    	lib/facter/operatingsystem.rb
    kbarber committed Jan 17, 2012
Commits on Jan 13, 2012
  1. @daniel-pittman

    Merge pull request #141 from joshcooper/ticket/1.6.x/11848-relocatabl…

    …e-facter-bat
    
    Ticket/1.6.x/11848 relocatable facter bat
    daniel-pittman committed Jan 12, 2012
  2. @daniel-pittman

    Merge pull request #136 from mv/ticket/1.6.x/10271-identifying_amazon

    (#10271) Identifying 'Amazon' using '/etc/system-release'
    daniel-pittman committed Jan 12, 2012
  3. @daniel-pittman

    Merge pull request #110 from stschulte/feature/master/11082

    add proper operatingsystemrelease fact for solaris
    daniel-pittman committed Jan 12, 2012
Commits on Jan 11, 2012
  1. @joshcooper

    (#11848) Don't hard code ruby install paths in Windows batch files

    Previously, the facter.bat file hard coded the path to the ruby
    installation, making it impossible to move the ruby install directory.
    
    This commit changes the script to use the `%~dp0` batch file modifier,
    which resolves to the drive letter and path of the directory of the
    batch file being executed.
    
    Windows XP and later all support the `%*` modifier, so this commit
    removes the Win 9x code paths that are not supported.
    joshcooper committed Jan 9, 2012
Commits on Jan 10, 2012
  1. @glarizza

    Build a Rake task for building Apple Packages

    The goal is to have our release managers build all Facter packages
    within Rake, and so this commit adds a rake task (apple_package) that
    will build a DMG-encapsulated package for OS X and put it into the
    pkg/apple directory off of the Facter Root.
    
    To accomplish this goal, a folder structure is created in
    /tmp/facter/facter-#{version} that mirrors the structure needed for
    Apple's Packagemaker CLI tool to build a package for OS X. Next, the
    necessary files are copied from the Puppet source into the structure in
    /tmp/facter/facter-#{version} and then packagemaker is run
    (targeting that folder structure) to build an initial package. Hdiutil
    then encapsulates that package into a DMG, and the file is finally
    copied into the pkg/apple directory.
    glarizza committed Jan 10, 2012
  2. @jeffmccune

    Merge pull request #142 from kbarber/ticket/1.6.x/11559-rbconfig_depr…

    …ecation
    
    Ticket/1.6.x/11559 rbconfig deprecation
    jeffmccune committed Jan 10, 2012
  3. @kbarber

    (#11559) Switch to RbConfig & Provide alias for RbConfig for pre-1.8.5

    In Ruby 1.9.3 we started to see warning messages about the deprecated usage of
    the class 'Config' as apposed to 'RbConfig'. This patch switches all our code
    to 'RbConfig' instead.
    
    Now the constant RbConfig was aliased to Config going back to Ruby 1.8.5, but
    for pre-1.8.5 ruby this was not done. To support our Ruby 1.8.1 user base
    (due to RHEL 4 predominantly) this provides a monkey patch util library which
    provides the same alias (conditionally) but for older rubies instead.
    
    This methodology is seen as less intrusive and less maintainance going forward
    then to create conditionals everywhere, or a facter-only backwards compatible
    alternative to RbConfig. And since its the similar methodology adopted by Ruby
    core (more or less) it should be a monkey patch with low surprises.
    
    I've also added testing around the facter/util/config.rb library.
    
    Thanks to James Turnbull <james@lovedthanlost.net> for the original code for
    this patch.
    kbarber committed Jan 3, 2012
Commits on Jan 4, 2012
  1. @daniel-pittman

    Revert "(#11660) Adds feature SSHFP"

    This reverts commit 50f6da6.
    This reverts commit 6c8683a.
    
    These ware submitted oddly, and I didn't notice that I merged the non-working
    tests as well as the correct looking code, since they had two tickets
    associated.  Backing both out now.
    daniel-pittman committed Jan 3, 2012
  2. @daniel-pittman

    Merge pull request #133 from grooverdan/ticket/11660-sshfp_and_ssh_rspec

    (#11660) Adds feature SSHFP
    daniel-pittman committed Jan 3, 2012
Commits on Jan 2, 2012
  1. @mv
Commits on Jan 1, 2012
  1. @adrienthebo

    Merge remote-tracking branch 'kbarber/ticket/1.6.x/11661-ec2_rspec_ra…

    …ises_not_throws' into 1.6.x
    adrienthebo committed Jan 1, 2012
  2. @kbarber

    (#11661) EC2 rspec tests were using throw not raise to simulate a tim…

    …eout
    
    This patch corrects that behaviour. Also, 'throws' wasn't working on our CI
    machines since we were using mocha 0.9.x.
    kbarber committed Jan 1, 2012
  3. @grooverdan

    (#11660) Adds feature SSHFP

    Corrects previous which didn't include the right file.
    
    RFC4255 SSH fingerprints are now facts in the form SSHFP_RSA, SSHFP_DSA.
    No other algorithms are currently supported
    (www.iana.org/assignments/dns-sshfp-rr-parameters).
    grooverdan committed Jan 1, 2012
  4. @grooverdan

    (#11660) Adds feature SSHFP

    (#11659) Adds ssh spec
    
    RFC4255 SSH fingerprints are now facts in the form SSHFP_RSA, SSHFP_DSA.
    No other algorithms are currently supported
    (www.iana.org/assignments/dns-sshfp-rr-parameters).
    
    Adds test cases for the ssh fact module.
    grooverdan committed Jan 1, 2012
Commits on Dec 31, 2011
  1. @jeffmccune

    Merge pull request #128 from kbarber/ticket/1.6.x/11583-ec2_tests

    Ticket/1.6.x/11583 ec2 tests
    jeffmccune committed Dec 31, 2011
  2. @kbarber

    (#11583) Add basic coverage to the ec2 fact

    The only goal here is to add coverage based on the current behaviour so we have
    something to work with when other contributors work on the code.
    
    We've only added very simple bad-case testing (ie. timeout) to match the code
    at the moment, however the good-case testing covers flat & structured meta-data
    and user-data. While we test both ec2 and eucalyptus cases, we only check basic
    user-data for eucalyptus for now (as the code is the same in both cases).
    kbarber committed Dec 30, 2011
Commits on Dec 30, 2011
  1. @grimradical

    Merge pull request #131 from adrienthebo/ticket/1.6.x/9599-nexenta_op…

    …eratingsystem
    
    (#9599) Generalize zone detection
    grimradical committed Dec 30, 2011
  2. @adrienthebo

    (#9599) Generalize zone detection

    All systems using the SunOS kernel can run zones, not just the solaris
    operating system. Generalized this for cases like Nexenta and Open
    Indiana.
    adrienthebo committed Dec 30, 2011
  3. @kbarber

    Merge branch 'ticket/1.6.x/9599-nexenta_operatingsystem' into 1.6.x

    * ticket/1.6.x/9599-nexenta_operatingsystem:
      (#9599) Add nexenta facts
    kbarber committed Dec 30, 2011
  4. @saysjonathan @kbarber

    (#9599) Add nexenta facts

    Adds support for Nexenta to the operatingsystem and osfamily facts.
    Since Nexenta is binary compatible with solaris, it has also been added
    into the Solaris osfamily fact.
    
    Tests and osfamily support added by Adrien Thebo <adrien@puppetlabs.com>
    saysjonathan committed with kbarber Sep 19, 2011
  5. @kbarber

    (#11583) Switch request method to open-uri monkey patch 'open'

    This makes the code consistent with the rest of the EC2 fact code by just using
    open for http requests instead of the full open_uri method. This is to make my
    testing consistent and is also slightly cosmetic.
    
    I also fixed the indentation on those two lines as well.
    kbarber committed Dec 30, 2011
Commits on Dec 22, 2011
  1. @daniel-pittman

    Merge pull request #122 from kbarber/ticket/1.6.x/9708-os_confines

    Ticket/1.6.x/9708 os confines
    daniel-pittman committed Dec 21, 2011
  2. @daniel-pittman

    Merge pull request #121 from kbarber/ticket/1.6.x/9708-os_confines

    (#9708) Confine facts by kernel not operating system and remove confine ...
    daniel-pittman committed Dec 21, 2011
Commits on Dec 21, 2011
  1. @kbarber

    (#9708) Amend requires in specs to use simple requires

    In a previous change to 1.6.x we changed from using a require on a path to just
    a shorter string. This amends the cherry-picked patch from master to use this
    new method and make rspec work again in more modern rspec revisions.
    kbarber committed Dec 21, 2011
  2. @kbarber

    (#9708) Confine facts by kernel not operating system and remove confi…

    …ne for hardwareisa
    
    This will allow the following facts to continue working even when new variants
    of Linux, BSD, etc are supported in the future.
    
    * lsbmajdistrelease
    * macaddress
    * uniqueid
    
    For hardwareisa the confine was removed entirely. On any system that returns
    "uname -p", the value should be this fact. On systems that don't have
    "uname -p", the resolution will not be used. Thus, it shouldn't be necessary to
    confine the resolution by kernel.
    
    Previously we were constantly requiring patches to be added whenever a new
    Linux distribution was added. These changes should reduce this need. The fact
    'operatingsystem', 'osfamily' and 'operatingsystemrelease' will still require
    patches however.
    
    New spec tests were added where they were missing.
    (cherry picked from commit 2eb4ede)
    Michael Kincaid committed with kbarber Sep 27, 2011
  3. @kbarber

    (maint) Fix requires for newer rspec revisions so we don't break build

    In newer revisions of rspec the older methodology for loading rspec:
    
        require File.expand_path(File.dirname(__FILE__) + '/../spec_helper')
    
    Will throw an error when mixed with the new method:
    
        require 'spec_helper'
    
    This patch fixes the 4 spec files that were missing from 1.6.x when the original
    patch changed this (#11436).
    kbarber committed Dec 21, 2011
Something went wrong with that request. Please try again.