Describe the Bug

FreeBSD recently switched from Ruby 3.0 to 3.1 as default FreeBSD version. All packages where rebuild with Ruby 3.1, but after updating them, bolt fails to run:

romain@zappy ~ % bolt 
/usr/local/lib/ruby/site_ruby/3.1/rubygems/specification.rb:2287:in `raise_if_conflicts': Unable to activate puppet_forge-4.0.0, because faraday-1.10.3 conflicts with faraday (~> 2.0) (Gem::ConflictError)
	from /usr/local/lib/ruby/site_ruby/3.1/rubygems/specification.rb:1418:in `activate'
	from /usr/local/lib/ruby/site_ruby/3.1/rubygems/specification.rb:1456:in `block in activate_dependencies'
	from /usr/local/lib/ruby/site_ruby/3.1/rubygems/specification.rb:1438:in `each'
	from /usr/local/lib/ruby/site_ruby/3.1/rubygems/specification.rb:1438:in `activate_dependencies'
	from /usr/local/lib/ruby/site_ruby/3.1/rubygems/specification.rb:1420:in `activate'
	from /usr/local/lib/ruby/site_ruby/3.1/rubygems/specification.rb:1456:in `block in activate_dependencies'
	from /usr/local/lib/ruby/site_ruby/3.1/rubygems/specification.rb:1438:in `each'
	from /usr/local/lib/ruby/site_ruby/3.1/rubygems/specification.rb:1438:in `activate_dependencies'
	from /usr/local/lib/ruby/site_ruby/3.1/rubygems/specification.rb:1420:in `activate'
	from /usr/local/lib/ruby/site_ruby/3.1/rubygems.rb:282:in `block in activate_bin_path'
	from /usr/local/lib/ruby/site_ruby/3.1/rubygems.rb:281:in `synchronize'
	from /usr/local/lib/ruby/site_ruby/3.1/rubygems.rb:281:in `activate_bin_path'
	from /usr/local/bin/bolt:25:in `<main>'

Expected Behavior

Bolt should work as before.

Steps to Reproduce

On FreeBSD:

pkg install bolt

Environment

• Version 0.6.1
• Platform FreeBSD 13.2-RC3

Additional Context

Bolt indirectly depends on two different versions of faraday:

rubygem-bolt-3.26.2
|-> rubygem-r10k-3.15.3_1
|   `-> rubygem-puppet_forge-4.0.0 ([faraday](https://rubygems.org/gems/faraday) ~> 2.0)
|       `-> rubygem-faraday-2.7.4
`-> rubygem-orchestrator_client-0.6.1
    `-> rubygem-faraday1-1.10.3 ([faraday](https://rubygems.org/gems/faraday) ~> 1.4)

Ruby 3.0 seems to cope with this, but 3.1 complains.

Adjusting orchestrator_client-0.6.1.gemspec and changing the version requirement for faraday from ~> 1.4 to ~> 2.0 seems to workaround the issue.

I'll open a PR tomorrow to bump the version requirement.

Describe the Bug

Starting with Ruby 2.7, the URI.escape emits a deprecation warning:

opt/puppetlabs/bolt/lib/ruby/gems/2.7.0/gems/orchestrator_client-0.5.1/lib/orchestrator_client/config.rb:101: warning: URI.escape is obsolete

This method is used in the config class here: https://github.com/puppetlabs/orchestrator_client-ruby/blob/master/lib/orchestrator_client/config.rb#L101

Use Case

CD4PE uses orchestrator_client as part of bolt to run jobs through the PE orchestrator. Under heavy load, these connections can time out. Unfortunately, the connection and read timeouts do not appear to be configurable, so the CD4PE jobs simply fail and users have no way to address the issue.

Describe the Solution You Would Like

Make these timeouts configurable. Otherwise, users are forced to manually update the code to work around issues with busy servers. https://github.com/puppetlabs/orchestrator_client-ruby/blob/0.4.3/lib/orchestrator_client.rb#L35 is the idle timeout. Adding http.read_timeout here will customize the read timeout from the default included in the Faraday gem of 60 seconds. Wiring up overrides to these values that can be populated by bolt.yaml entries would be ideal.

Describe Alternatives You've Considered

Users can change this code manually to get the desired effect. Unfortunately, in the case of CD4PE, this is shipped as part of the container, and thus resets on each container restart. So the user has to re-apply the patch each time. It would be better as just part of the application.

Due to the hardcoded path at

Is this gem supposed to work on Windows?

Looks like puppet tasks are not yet supported, though.

Any plans to add that?