| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* changing the default drive for the config file to be the drive that the executable is launched from. This should allow for an easier launch of chef from a d: drive installation
Signed-off-by: John Kerry <john@kerryhouse.net>
* adding a check and a fallback to systemdrive for the base path when __FILE__ path doesn't have a drive (appears with linux hosts running windows targeted unit tests)
Signed-off-by: John Kerry <john@kerryhouse.net>
* fixing rubocop errors
Signed-off-by: John Kerry <john@kerryhouse.net>
* adding a __FILE__ path override for cache_path tests
Signed-off-by: John Kerry <john@kerryhouse.net>
|
| |
|
|
|
|
|
| |
When we update our changelog it’s a bit confusing since multiple days go by before we ship the new release. Adding a header to the changelog and a readme gem release badge should help clear up what our current release actually is.
Signed-off-by: Tim Smith <tsmith@chef.io>
|
| |
|
|
|
| |
sensitive is the property of the resource, not the provider.
|
| |
|
|
|
| |
Mac: Ensure that launchd limit_load_to_session_type is properly handled
|
| |
|
|\
| |
| | |
Prevent apt_update failures on non-Linux platforms
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
We 1/2 wired up the noop functionality, but specifying that we only provided the resource on Linux caused that code to never run.
This prevents this sort of error on non-Linux platforms:
```
Cannot find a resource for apt_update on windows version 6.3.9600
```
Signed-off-by: Tim Smith <tsmith@chef.io>
|
| | |
|
|/
|
|
|
|
|
| |
Use the latest fedora to avoid having to bump things every X months
We need to make the cache afterwards. I ran into this same issue in the cron cookbook and I’ve updated the dnf_compat recipe in yum to do the same.
Signed-off-by: Tim Smith <tsmith@chef.io>
|
| |
|
|
|
| |
Signed-off-by: Ian C. Ward <ian.ward@disney.com>
|
| |
|
|\
| |
| | |
Cherry Pick omnibus commits
|
| |
| |
| |
| | |
Signed-off-by: Jaymala Sinha <jsinha@chef.io>
|
| | |
|
| |
| |
| |
| |
| | |
scientific linux, oracle, etc, etc...
Signed-off-by: Lamont Granquist <lamont@scriptkiddie.org>
|
| | |
|
|\ \
| | |
| | | |
12.16 Release Notes
|
| | |
| | |
| | |
| | | |
Signed-off-by: Tom Duffield <tom@chef.io>
|
| | | |
|
|\ \ \
| |/ /
|/| | |
Update changelog for upcoming release
|
| |/
| |
| |
| |
| |
| |
| | |
Signed-off-by: Salim Afiune <afiune@chef.io>
Signed-off-by: Tom Duffield <tom@chef.io>
[ci_skip]
|
| | |
|
|\ \
| |/
|/| |
Remove vendored git submodules
|
|/
|
|
|
|
|
| |
Similar to the issue uncovered in chef/chef-server#997 which breaks
Bundler.
Signed-off-by: Seth Chisamore <schisamo@chef.io>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Update changelog generator to better fill our pattern
Configure the changelog generator to better fit our current pattern of
changelogs. By utilizing the HISTORY.md file, we can generate a combined
changelog of the latest version from HISTORY.md to now. Then, when we
cut a release we can simply move our current changelog into
HISTORY.md (using the appropriate rake task) and start the process over
again.
This might even be a first step to having the Chef Versioner generate
the changelog automatically for us each time it bumps the version, and
then we just need to groom it as we go along.
Signed-off-by: Tom Duffield <tom@chef.io>
* Add HISTORY.md for the 12.15.19 release
Signed-off-by: Tom Duffield <tom@chef.io>
|
| |
|
|\
| |
| | |
Fix data collector acceptance
|
| |
| |
| |
| | |
Signed-off-by: Daniel DeLeo <dan@chef.io>
|
| |
| |
| |
| |
| |
| |
| | |
Implementations should be robust to extra fields for forwards
compatibility with future clients.
Signed-off-by: Daniel DeLeo <dan@chef.io>
|
|/
|
|
| |
Signed-off-by: Daniel DeLeo <dan@chef.io>
|
| |
|
|\
| |
| | |
Enable data collector w/o token for solo, but require explicit URL
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Third-party data collector implementations are supported and may not
require a token for authentication. Therefore we must allow the case
where the token is nil in Chef Solo mode. To reduce the chance that the
data collector is accidentally enabled, only set the data collector URL
to a default value in Chef Client mode.
Signed-off-by: Daniel DeLeo <dan@chef.io>
|
| |
| |
| | |
Break up individual changes with headers
|
|/ |
|
|
|
|
|
|
|
| |
Also, Chef Server 12.10 was released, so we will most likely require
12.11+ for the feature to work.
Signed-off-by: Daniel DeLeo <dan@chef.io>
|
|
|
|
| |
Signed-off-by: Daniel DeLeo <dan@chef.io>
|
|
|
|
| |
Signed-off-by: Daniel DeLeo <dan@chef.io>
|
|
|
|
|
|
|
| |
Additionally, add info and debug logging for the auto-detection of
whether to enable the data collector.
Signed-off-by: Daniel DeLeo <dan@chef.io>
|
|
|
|
| |
Signed-off-by: Daniel DeLeo <dan@chef.io>
|
|
|
|
| |
Signed-off-by: Daniel DeLeo <dan@chef.io>
|
|
|
|
| |
Signed-off-by: Daniel DeLeo <dan@chef.io>
|
|
|
|
|
|
|
|
|
|
| |
* When the data collector is enabled but the token is nil, data
collector will use signed header auth.
* Switch data collector to an HTTP client that includes the JSON content
middlewares so both auth methods talk to the `http` object the same
way.
Signed-off-by: Daniel DeLeo <dan@chef.io>
|
| |
|