| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
| |
| |
| |
| |
| |
| |
| |
| | |
Newer versions of lua output lua -v on stdout while older versions use
stderr. This update will handle both and ensure that we continue to
detect the installed lua version.
Signed-off-by: Tim Smith <tsmith@chef.io>
|
|\ \
| | |
| | | |
Fix scala detection when version output contains a warning
|
| |/
| |
| |
| |
| |
| | |
It seems that sometimes scala outputs a bogus line ahead out the version output. Use a match instead of a split to better handle this sort of thing.
Signed-off-by: Tim Smith <tsmith@chef.io>
|
|\ \
| |/
|/| |
rescue github-changelog-generator in Rakefile
|
| |
| |
| |
| | |
Signed-off-by: Lamont Granquist <lamont@scriptkiddie.org>
|
|\ \
| |/
| | |
Fix a chefstyle warning
|
|/
|
|
| |
Signed-off-by: Tim Smith <tsmith@chef.io>
|
|\
| |
| | |
Add Digital Ocean metadata from their metadata API
|
| |
| |
| |
| | |
Signed-off-by: Tim Smith <tsmith@chef.io>
|
| |
| |
| |
| | |
Signed-off-by: Tim Smith <tsmith@chef.io>
|
| |
| |
| |
| | |
Signed-off-by: Tim Smith <tsmith84@gmail.com>
|
| | |
|
| | |
|
| |
| |
| |
| |
| |
| | |
The network plugin is leftover from the old method where we looked up the interfaces based on the IP in the hint. We don't do that anymore so this isn't necessary.
Vendor data can include sensitive bootstrapping information that we shouldn't save to the node. It also doesn't include anything of value so either way the data shouldn't be on the node
|
| |
| |
| |
| | |
Without this it's just a big string and the .each fails
|
| |
| |
| |
| | |
The cloud data on the droplet I spun up didn't have the string listed and it also required disk reads / yaml to be loaded. This should be a lot faster since we're grabbing this data anyways.
|
| | |
|
| | |
|
| |
| |
| |
| | |
: vs. => was killing the spec
|
| | |
|
| | |
|
|\ \
| | |
| | | |
fix logger issues
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
ruby doesn't like optional hashes combined with optional named
parameters -- because hashes and kwargs get magically converted
and ruby gets confused.
plus fixed/added specs
Signed-off-by: Lamont Granquist <lamont@scriptkiddie.org>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Lamont Granquist <lamont@scriptkiddie.org>
|
|/ /
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
i believe this gets the logic correct in that we want to configure the
logger when we're coming from the command line but want to skip it when
we're coming from chef or from other APIs that directly construct an
Ohai::System object.
i suspect this is what thom was trying to do by moving the
Ohai::Log.init call into the Ohai::Application class (which avoids it
being called entirely when Chef creates its Ohai::System object) here:
https://github.com/chef/ohai/pull/942/files
but that change broke the behavior where we were also supposed to skip
the rest of the configure_logging method in Ohai::System when run under
chef client.
I tried going down the route of having the Ohai::Application class
construct the Ohai::System object and then having it be the
responsibility of that caller to do configure_logging work. However,
I suspect that the initializer in Ohai::System does way too much and
that the purpose of configuring the logger where it is, is that it
must be initialized in the middle of object creation before it goes on
and creates the Loader, the Runner, creates Hints and removes constants.
So, I went the route of threading a flag through the initializer so that
Ohai::System can know if its coming from the cli or not and behave
appropriately.
There's also quite a mess with the Ohai::Log class being passed around
to the workstation loader, and it initializes itself at class loading
time and Chef::Application will inject state into Ohai::Log. I think
the eager initialization at class loading time is to not lose early
messages when run from the cli, but it means the Ohai::Log object is always
initialized, so that isn't useful for determining if we have come from the
cli or not.
Signed-off-by: Lamont Granquist <lamont@scriptkiddie.org>
|
|/
|
|
|
|
|
| |
this probably should have been done on master earlier since
breaking changes already went into ohai master
Signed-off-by: Lamont Granquist <lamont@scriptkiddie.org>
|
|\
| |
| | |
Move duplicate http logic into a helper
|
| |
| |
| |
| |
| |
| | |
We probably want to use this in a few other places, but we certainly want to encourage people to use this in their plugins vs. reinventing the wheel.
Signed-off-by: Tim Smith <tsmith@chef.io>
|
|\ \
| | |
| | | |
autofixing new chefstyle issues
|
|/ /
| |
| |
| |
| |
| | |
prep for chefstyle 0.5.0 release to not break master
Signed-off-by: Lamont Granquist <lamont@scriptkiddie.org>
|
|\ \
| |/
|/| |
Config is used in system so require it there.
|
| |
| |
| |
| |
| |
| | |
COOL team found this when using ohai/system directly. It worked in 8.X, but not post deprecation of the old config logic.
Signed-off-by: Tim Smith <tsmith@chef.io>
|
|\ \
| | |
| | | |
Remove a bit of dead code
|
| | |
| | |
| | |
| | | |
Signed-off-by: Tim Smith <tsmith@chef.io>
|
|\ \ \
| |/ /
|/| | |
Remove rescue loading plist in the specs as this is a dep
|
|/ /
| |
| |
| |
| |
| | |
There’s no need for this and we’re not doing this in the dozen other places we load deps.
Signed-off-by: Tim Smith <tsmith@chef.io>
|
|\ \
| |/
|/| |
Minor spec file cleanup
|
|/
|
|
|
|
|
|
|
|
|
|
|
| |
- Use require_relative in all the specs vs. expanding on the current file
path.
- Add missing license headers
- Don't silently swallow ipaddr_extensions not being installed. Make it
a development dep instead
- Remove a rescue on ipaddress as we only do this in one place and it's
a required gem
- Make the Openstack plugin spec setup the same way as every other spec
Signed-off-by: Tim Smith <tsmith@chef.io>
|
|\
| |
| | |
Update travis/appveyer preinstall to resolve rainbow failures
|
|/
|
|
|
|
| |
Update Rubygems, fully uninstall bundler, install the latest bundler.
Signed-off-by: Tim Smith <tsmith@chef.io>
|
|\
| |
| | |
Remove deprecated config logic
|
| |
| |
| |
| | |
Signed-off-by: Tim Smith <tsmith@chef.io>
|
| |
| |
| |
| | |
Signed-off-by: Tim Smith <tsmith@chef.io>
|
|\ \
| | |
| | | |
Remove the deprecated run_command and popen4 methods
|
| | |
| | |
| | |
| | |
| | |
| | | |
This is very much legacy and has long since been replaced by shell_out
Signed-off-by: Tim Smith <tsmith@chef.io>
|
|\ \ \
| | | |
| | | |
| | | |
| | | | |
chef/COOL-626/tduffield/use-cpu-name-for-model-name-windows
Use name for Windows CPU model_name
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Use the "name" value for the model_name, and move description to the
description field.
Signed-off-by: Tom Duffield <tom@chef.io>
|
|\ \ \ \
| | | | |
| | | | | |
Remove usage of the Sigar gem
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Sigar was a really cool idea, but it’s a dead project now. The data is questionable at best and not getting any better as time goes on. We should figure out how to support these platforms vs. hoping that Sigar is doing the right thing. This does remove a lot of support for HP-UX, but I seriously doubt it worked to begin with. We should do that right sometime.
Signed-off-by: Tim Smith <tsmith@chef.io>
|
|\ \ \ \ \
| | | | | |
| | | | | | |
Require Ruby 2.2+
|