| Age | Commit message (Collapse) | Author |
|---|
|
Prior to this commit, users coming to the type_of function would not realize that the type function in puppet does the same thing and is preferred over type_of.
After this commit, we have a comment indicating the above.
|
|
(MODULES-3393) Deprecation - Use puppet stacktrace if available
|
|
Without this, some uses of this function do not work in puppet3. e.g.
if $include_src != undef {
deprecation('apt $include_src', "please use \$include => { 'src' => ${include_src} } instead")
}
causes
Function 'deprecation' must be the value of a statement
on puppet 3.8.7.
|
|
A previous PR (#685) was raised on this issue, however once it was merged it was discovered that Puppet 3 with future parser enabled was calling the Puppet 4 version of the deprecation function. The Puppet stacktrace is not available until Puppet 4.6, so this was breaking existing setups. The solution was to check is the stacktrace was defined, and if it was to use it as part of the message output.
|
|
|
|
|
|
Handle array values in join_keys_to_values function
|
|
This function otherwise fails during `puppet preview` on Puppet 3.8.X systems.
|
|
|
|
This is a workaround for PUP-4438 (fixed in https://github.com/puppetlabs/puppet/commit/e01c4dc924cd963ff6630008a5200fc6a2023b08, 4.1.0, 3.8.1).
It works by manually passing through the scope, instead of relying on the
InternalFunction class.
|
|
This now emits one deprecation warning for each function used (but not for each call-site). Prior to this, only a single deprecation warning would have been triggered, potentially misleading users.
Additionally this adds v4 deprecation stubs for the functions that were missed.
|
|
|
|
This reverts commit b63862ff43194194f7428739a32cfe13bad1e7ed, as it would
only show the irrelevant first entry of the ruby stack trace. The
puppetserver log does contain the full trace information, or you can use
--strict=error to cause a hard failure when hitting a deprecation.
# Conflicts:
# lib/puppet/functions/validate_legacy.rb
|
|
`catch` returns value of second argument to `throw`, which until 860a2761f334c964068038b3ef6853f08beb1df5 was `nil`, but now is non-falsey reason for error. short-circuit using return and eval to nil if `throw` was caught.
|
|
|
|
This was not working when the puppet master did not have the newest stdlib
version in its environment.
|
|
Prior to this commit, due to a bug in the exists? method in the
file_line provider, match_for_absence didn't work as described (or
at all really). Update the exists? logic so that match_for_absence
works as described.
Additionally add a unit test to prevent regressions and update the
documentation for the parameter to reflect the fact that it is
ignored when `ensure => present`.
|
|
This is the idiomatic version of #634, and also addresses
the test failures.
Original-Fix-By: Michiel Brandenburg <michiel@silverstreet.com>
|
|
When validate_string is called via the Puppet 4 deprecation wrappers
from deprecation_gen (introduced in 970852d), `undef` is passed as `nil`
where it was previously passed as `''` from the Puppet 3-style function
API.
This change explicitly permits a `nil` value in validate_string, and
adds a test case to `is_string` which also accepts the same.
Fixes test failures in apt, concat etc:
Error while evaluating a Function Call, nil is not a string. It looks to be a NilClass at apt/manifests/source.pp:23:3 [..]
# ./spec/fixtures/modules/stdlib/lib/puppet/parser/functions/validate_string.rb:34:in `block (2 levels) in <module:Functions>'
# ./spec/fixtures/modules/stdlib/lib/puppet/parser/functions/validate_string.rb:32:in `each'
# ./spec/fixtures/modules/stdlib/lib/puppet/parser/functions/validate_string.rb:32:in `block in <module:Functions>'
# puppet-4.7.0/lib/puppet/parser/functions.rb:174:in `block (2 levels) in newfunction'
# puppet-4.7.0/lib/puppet/util/profiler/around_profiler.rb:58:in `profile'
# puppet-4.7.0/lib/puppet/util/profiler.rb:51:in `profile'
# puppet-4.7.0/lib/puppet/parser/functions.rb:167:in `block in newfunction'
# ./spec/fixtures/modules/stdlib/lib/puppet_x/puppetlabs/stdlib/deprecation_gen.rb:13:in `block (2 levels) in deprecation_gen'
|
|
|
|
|
|
Add deprecation warnings to remaining validates
|
|
|
|
|
|
Ensure validate functions use Puppet 4 deprecation
|
|
|
|
|
|
Refactor dig44 function
|
|
* validate_legacy now accepts undef values
* update the TypeMismatch message to include the original validate function name
* only notice, not warn, on newly allowed values
* changed previous_validation to function_name to avoid confusion with the function of the same name
* use deprecation() instead of warn(), when hitting a deprecated value
* prepare the tests and function for MODULES-3735
* rewrite validate_legacy tests to use new rspec-puppet
* move validate_re deprecation to puppet4 only
* adapt validate_re_spec
|
|
|
|
|
|
The current implementation of the dig44 function has the
following problems:
* Doesn't recognise Puppet4's :undef value as an empty value
and doesn't return the default value.
* Doesn't make a different between false and nil value
and returns the default value for a non-empty false value
|
|
(MODULES-3534) Deprecation of ip functions
|
|
|
|
The deprecation function was calling the `Kernel#warn` function which
prints to stderr, rather than the Puppet logger. This causes problems
for Puppet module tests on Travis CI, which has a cap on the amount of
stdout/err permitted in its logs and also prevents users from finding
the deprecation warnings when running under a Puppet master.
|
|
|
|
|
|
(modules-3533) deprecation for 3.x number function
|
|
|
|
|
|
|
|
|
|
|
|
|
|
(MODULES-3529) add deprecation function
|
|
|
|
(MODULES-3568) Move dig to dig44 and deprecate dig
|
|
A new version of dig was introduced in Puppet 4.5.0 that
isn't compatible with the stdlib version of dig. To maintain
backwards compatibility and ensure that tests for stdlib aren't
broken, this patch renames dig to dig44 and adds a deprecation
warning to the stdlib dig function.
|
|
|
|
As described in PUP-6422,
ensure_resources('File[/tmp/a]', { owner => undef })
would not actually create the file. This fixes it, and adds tests to prove
it.
|
