incompatible encoding regexp match for non-printing characters in place of space

[sigv]

It looks like when you declare a resource and provide a key parameter, that is not handled nicely. Swap it out for abc and the error goes away.

An example of where it is in use is the puppetlabs/registry module and the registry::value type it has, which has a key parameter there, at least on latest (v1.1.4) which we are using now.

The linter is not able to deal with non-printing characters in places where it expects whitespace. The parser throws out a syntax error - even though Puppet's validation does not complain about the file's contents.

---

puppet-lint version: 2.2.1
ruby version: 2.1.9-p490
platform: x64-mingw32
file path: ./foo/manifests/bar.pp
file contents:

class foo::bar {
  abc::def { 'tray':
    key => 'A',
  }
}

error:

Encoding::CompatibilityError: incompatible encoding regexp match (UTF-8 regexp with IBM775 string)
C:/Ruby21-x64/lib/ruby/gems/2.1.0/gems/puppet-lint-2.2.1/lib/puppet-lint/plugins/check_whitespace.rb:65:in `scan'
C:/Ruby21-x64/lib/ruby/gems/2.1.0/gems/puppet-lint-2.2.1/lib/puppet-lint/plugins/check_whitespace.rb:65:in `block in check'
C:/Ruby21-x64/lib/ruby/gems/2.1.0/gems/puppet-lint-2.2.1/lib/puppet-lint/plugins/check_whitespace.rb:63:in `each'
C:/Ruby21-x64/lib/ruby/gems/2.1.0/gems/puppet-lint-2.2.1/lib/puppet-lint/plugins/check_whitespace.rb:63:in `each_with_index'
C:/Ruby21-x64/lib/ruby/gems/2.1.0/gems/puppet-lint-2.2.1/lib/puppet-lint/plugins/check_whitespace.rb:63:in `check'
C:/Ruby21-x64/lib/ruby/gems/2.1.0/gems/puppet-lint-2.2.1/lib/puppet-lint/checkplugin.rb:21:in `run'
C:/Ruby21-x64/lib/ruby/gems/2.1.0/gems/puppet-lint-2.2.1/lib/puppet-lint/checks.rb:58:in `block in run'
C:/Ruby21-x64/lib/ruby/gems/2.1.0/gems/puppet-lint-2.2.1/lib/puppet-lint/checks.rb:56:in `each'
C:/Ruby21-x64/lib/ruby/gems/2.1.0/gems/puppet-lint-2.2.1/lib/puppet-lint/checks.rb:56:in `run'
C:/Ruby21-x64/lib/ruby/gems/2.1.0/gems/puppet-lint-2.2.1/lib/puppet-lint.rb:191:in `run'
C:/Ruby21-x64/lib/ruby/gems/2.1.0/gems/puppet-lint-2.2.1/lib/puppet-lint/bin.rb:59:in `block in run'
C:/Ruby21-x64/lib/ruby/gems/2.1.0/gems/puppet-lint-2.2.1/lib/puppet-lint/bin.rb:56:in `each'
C:/Ruby21-x64/lib/ruby/gems/2.1.0/gems/puppet-lint-2.2.1/lib/puppet-lint/bin.rb:56:in `run'
C:/Ruby21-x64/lib/ruby/gems/2.1.0/gems/puppet-lint-2.2.1/bin/puppet-lint:7:in `<top (required)>'
C:/Ruby21-x64/bin/puppet-lint:22:in `load'
C:/Ruby21-x64/bin/puppet-lint:22:in `<main>'

[rnelson0]

I cannot replicate this with your inline sample (btw there is a non-printing character between key and =>, rather than a space) or the file registry/manifests/value.pp:

# With non-printing character intact:
$ be puppet-lint 693.pp
ERROR: Syntax error (try running `puppet parser validate <file>`) on line 3

# Replace non-printing character with a space:
$ be puppet-lint 693.pp
ERROR: foo::bar not in autoload module layout on line 1
WARNING: class not documented on line 1

# Using the contents of registry/manifests/value.pp
$ be puppet-lint 693.pp
ERROR: registry::value not in autoload module layout on line 42

This is consistent with HEAD, 2.2.1 and the 2.0.0 release, with ruby 2.3.1 via rvm and 2.0.0 via EL7 system ruby.

[sigv]

Yes, you are correct - replacing those characters with proper spaces does resolve the linter issue. I had not even noticed they were not actual spaces as the Puppet runs on nodes using the manifests had not complained about them. The error that is produced by the linter suggests running puppet parser validate but the validator does not error out - even setting the verbosity or debugging flags has no effect. So, since Puppet itself does not seem to have any issues parsing the syntax, I believe this should be properly handled by the linter with a proper error or even just a warning, saying that a space character should be put in the place of a non-printing character.

[sigv]

Edited the issue's title and description to represent the actual issue, rather than the one I initially reported.

[rnelson0]

I think this will be addressed successfully in #586, when the actual puppet parser is used.