- 06 Jun, 2019 1 commit
-
-
Thomas Goirand authored
that is older than 30 days to avoid filling-up a puppet-master hard drive until it's full (Closes: #930033).
-
- 12 Mar, 2019 2 commits
-
-
Apollon Oikonomopoulos authored
-
Apollon Oikonomopoulos authored
Closes: #923976
-
- 16 Jan, 2019 4 commits
-
-
Apollon Oikonomopoulos authored
-
Apollon Oikonomopoulos authored
-
Apollon Oikonomopoulos authored
Update to upstream version '5.5.10' with Debian dir 816f50259b59e36575ce817bc7173fef1371eae4
-
Apollon Oikonomopoulos authored
-
- 14 Jan, 2019 1 commit
-
-
Jenkins CI authored
-
- 05 Jan, 2019 1 commit
-
-
Josh Cooper authored
(PUP-6562) Restrict notice of failed dependency to once.
-
- 04 Jan, 2019 5 commits
-
-
Kris Bosland authored
Failed dependency resources will now only be reported once. Other resources that depend on the failed resource that has already been reported will not cause them to be reported again, but the user will still get the skip message for each skipped resource.
-
Kris Bosland authored
(PUP-9364) Exclude unnecessary module release metadata
-
Jenkins CI authored
-
Josh Cooper authored
The `puppet module install` command will download release metadata[1] for each release of the module and its dependencies so that the PMT can resolve dependencies locally. The v3 'releases' forge API response is verbose as the release metadata includes the module's readme, changelog and license, which are not needed to resolve dependencies. This commit excludes all currently known fields except version, metadata, and file_md5 to reduce the amount of transferred data. Ideally we could specify only the fields we want to be included, but the `include_field` parameter doesn't work that way. For stdlib specifically, this PR drops the amount of data transferred from 2.5MB to 77KB. This will also benefit modules that depend on stdlib. [1] https://forgeapi.puppet.com/#!/release/getReleases
-
Josh Cooper authored
merge 4.10.x -> 5.5.x
-
- 03 Jan, 2019 3 commits
-
-
Sean P. McDonald authored
* upstream/4.10.x: (maint) Update for bundler 2 (packaging) Updating manpage file for 4.10.x (maint) Conditionally update bundler gem for ruby 1.9.3 (packaging) Updating manpage file for 4.10.x (packaging) Bump to version '4.10.14' [no-promote]
-
Jacob Helwig authored
(maint) Update for bundler 2
-
Josh Cooper authored
When running on ruby < 2.3, use bundler 1.x and rubygems 2.7.8. When running on ruby >= 2.3, use bundler 2 and rubygems 3. This is necessary because bundler 2 requires rubygems 3 and ruby 2.3. This also uses the `--no-document` option as `--no-rdoc` and `--no-ri` were removed in bundler 2.
-
- 02 Jan, 2019 3 commits
-
-
Jenkins CI authored
-
Jacob Helwig authored
(maint) Conditionally update bundler gem for ruby 1.9.3
-
Josh Cooper authored
Travis CI with ruby 1.9.3 runs with bundler 1.7.6, which is about 4 years old. It doesn't take supported ruby versions into account during dependency resolution, causing puppet#4.10.x to install rake 12.3.2. However, that version requires ruby 2+, so bundle install fails. It also doesn't understand the :mri_22, etc platform identifiers, as are used in puppet#5.5.x Gemfile. Conditionally updating the bundler gem prior to bundle install resolves both issues.
-
- 19 Dec, 2018 2 commits
-
-
Jenkins CI authored
-
Jenkins CI authored
-
- 14 Dec, 2018 1 commit
-
-
Melissa authored
(maint) Bump minimum Facter gem version & remove obsolete tests
-
- 13 Dec, 2018 12 commits
-
-
Jacob Helwig authored
Facter 2.4.0 has been out since Jan 22, 2015, and is the minimum version to contain all of the Facter.debuggin, Facter.trace, and Facter.on_message methods. By requiring at least version 2.4.0 of the Ruby Facter, we no longer have to carry along tests with fragile setup to test for Puppet working if the version of Facter is too old to have any of these methods.
-
Jacob Helwig authored
(PUP-9330) Add tmpfs to SELinux filesystem whitelist
-
Josh Cooper authored
Merge 5.3.x to 5.5.x
-
Josh Cooper authored
* upstream/5.3.x: (packaging) Updating manpage file for 5.3.x (maint) Update tests requiring a yaml end of document in output (packaging) Updating manpage file for 4.10.x (PUP-9339) Add a status event to resource callback Conflicts: lib/puppet/transaction/event_manager.rb
-
Branan Riley authored
Merge 4.10.x to 5.3.x
-
Josh Cooper authored
* upstream/4.10.x: (packaging) Updating manpage file for 4.10.x (PUP-9339) Add a status event to resource callback Conflicts: lib/puppet/transaction/event_manager.rb man/man8/extlookup2hiera.8 man/man8/puppet-file.8 man/man8/puppet-inspect.8 man/man8/puppet-resource_type.8 spec/integration/transaction_spec.rb Resolved i18n related conflicts, notably the message is pluralized if there's more than one event. Also some applications were deleted in Puppet5.
-
Jenkins CI authored
-
Jacob Helwig authored
(maint) Update tests requiring a yaml end of document in output
-
Henrik Lindberg authored
The behavior of dropping the yaml end of document changed between Ruby 2.4.4p296 and 2.4.5p334. The tests regarding lookup are really not about yaml output compatibility and this commit drops the optional end of document at the end of the output if it is present so that the individual tests can compare against the rest of the produced yaml output.
-
Jenkins CI authored
-
Josh Cooper authored
(PUP-9339) Add a status event to resource callback
-
Josh Cooper authored
(maint) Update Rakefile warnings task to not check generated egrammar.rb
-
- 12 Dec, 2018 5 commits
-
-
Jenkins CI authored
-
Henrik Lindberg authored
Before this, whenever the grammar was regenerated, the warnings rake task would find warnings in it related to indentation. Unfortunately it is not possible to fix those without modifying racc as it produces many lines with inconsistent indentation. This makes the rake warnings task simply skip the generated file puppet/pops/parser/eparser.rb
-
Josh Cooper authored
Revert "Merge pull request #7228 from hlindberg/PUP-9303_fix-heredoc-locator-problems
-
Branan Riley authored
This reverts commit f31fd0f1, reversing changes made to f194cf2c.
-
Kris Bosland authored
With this change, resources that don't sync but are refresh only, will now get a status event recorded in the report.
-