| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| ... | |
| |\| | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | | |
Some specs were in conflict, as they had been recently modified to use
the API functionality that f950061 removes. This merge commit simply
restores the verson of those specs from f950061.
Manually Resolved Conflicts:
spec/unit/dsl/resource_type_api_spec.rb
|
| | | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | | |
Made the following modifications to ResourceTypeAPI:
(1) returned nil from “define”, “hostclass”, and “node”.
(2) renamed “mk_resource_type” and “munge_type_arguments” to
“__mk_resource_type__” and “__munge_type_arguments__” to discourage
customers from calling them.
(3) Made ResourceTypeAPI a class rather than a module, and changed the
parser to evaluate the contents of pure ruby manifests using a
instances of this class.
(4) Changed ResourceTypeAPI to insert newly instantiated types into
Thread.current[:known_resource_types] rather than the default
environment's known_resource_types.
This effectively backports the fix for issue #4657 to 2.6.x.
Also backported the new spec tests from #4657.
|
| |\| | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | | |
There are merge conflicts with commits following this one.
|
| | | | | | | | | | | |
|
| |\| | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | | |
Commit 2b50f30 simplified and fixed bugs in code that had already been
modified extensively by 4da88fb and 6b1dd81. This merge resolution
commit is a manual replay of the changes from 2b50f30 onto next.
Manually Resolved Conflicts:
lib/puppet/parser/type_loader.rb
spec/unit/parser/type_loader_spec.rb
|
| | | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | | |
The function import_if_possible, which was supposed to be responsible
for making sure that no two threads tried to import the same file at
the same time, was not making this decision based on the full pathname
of the file, since it was being invoked before pathnames were
resolved. As a result, if we attempted to import two distinct files
with the same name at the same time (either in two threads or in a
single thread due to recursion), one of the files would not always get
imported.
Fixed this problem by moving the thread-safety logic to happen after
filenames are resolved to absolute paths. This made it possible to
simplify the thread-safety logic significantly.
|
| |\| | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | | |
There are merge conflicts with commits following this one.
|
| | | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | | |
changed the regex so that tagmail allows . in tagname.
|
| | | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | | |
This is a minimal fix for #4631 by reversing one part of the refactor for #1903
commit 0d4fd60c7c143cc1f4e4b0f99f359c09cbfbf21e. It fixes the immediate issue
(implicit classes not being added to resources) but leaves open several other
questions which are defered to #4687.
|
| | | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | | |
Returning
nil in this provider had the effect that on every run, puppet would end
up calling yum erase . Returning the correct data structure
resolves this.
Thanks to Stephen Gran for the patch
|
| | | | | | | | | | | |
|
| | | | | | | | | | | |
|
| | | | | | | | | | | |
|
| | | |_|_|_|_|_|/ /
| |/| | | | | | | |
|
| |\| | | | | | | | |
|
| | | | | | | | | | |
|
| | | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
This fixes spec and unit tests indirectly related to the previous
patch-revert.
One failure was from trying to test the User Type's roles, when,
on many platforms, the roles feature wasn't supported by the default
Provider.
Other tests could fail on some platforms because they assumed that
unsupported attributes would be ignored with a warning, but the code was
crashing instead.
|
| | | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
Restored deleted lines from type.rb and reinstated unit tests
|
| | | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
This restricts the change introduced in #4691 to hostclasses, and leaves defined
resources and nodes alone, thus more closely mimicing the 0.25.x behaviour. It
also includes title, as this was similarly affected.
|
| | | |_|_|_|_|/ /
| |/| | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
A use of namevar apparently slipped through the net or got (re)introduced
in a merge/conflict resolution.
|
| |\| | | | | | | |
|
| | | | | | | | | |
|
| | | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
The class name is now stored internally with an initial capital, which changed
the contents of $name in a way that broke some manifests. This fix stores the
downcased version in $name.
|
| | | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
My fix for #4542 was overly enthusiastic about assuring that all resources had
a stage, resulting in stages designations being serialized for resources in
manifests which did not use resources (everything was in implicit main). This
broke 0.25.x compatibility, as all catalogs now refered to stages.
This patch scales back the change for #4542 slightly, supressing the setting of
main on the puppetmaster and relying on the default behaviour on the client (for
2.6.x and later, treat it as main; for 0.25.x, do nothing).
|
| | | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
The test was expecting a stub object but wasn't providing one.
|
| | | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
If the file streaming is to be thwarted when running standalone the test for
standalone needs to be stubbed when running specs, lest the tests inexplicably
fail.
|
| | | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
Basing pervasive low-level behaviour changes on the application name isn't a
good idea, but if we're going to do it we need to remember to update the test
when we rename the application. Further, if we optimize around such low-level
tests (as we did when implementing file-streaming) we need to carry the tests
over into the new data paths.
This commit addresses both of these isues, restoring 0.25.x stand-alone file
serving by adjusting the process-name test in the indirector to recognize
"apply" as standalone (and thus not use http to fetch the files) and by adding
a branch to the file streaming code to do the same when fetching the content.
I am not certain that this change is globally correct; there may be other
related problems that will not be fixed (and may in fact be exacerbated) by
this patch.
|
| | | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
puppet agent was not catching all species of exceptions, causing it to
crash on net/http timeout errors.
This patch broadens the types of exceptions that are caught.
|
| | | | | | | | | |
|
| | | | | | | | | |
|
| | | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
Autotest was failing on puppet apply tests because it was reading
autotest's args as the manifest file and then erroring because it wasn't
a valid file.
|
| | | | | | | | | |
|
| | | | | | | | | |
|
| | | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
The less stuff being done in the spec_helper the better for reasoning
about what's happening in the tests. puppettest.rb does a lot of things
that aren't necessary for the specs, so this patch gets those things out
of the spec_helper.
Reviewed by: Jesse Wolfe
|
| | | | | | | | | |
|
| | | | | | | | | |
|
| | | | | | | | | |
|
| | | | | | | | | |
|
| | | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
The serialization of much of the memory image was being triggered by a copy of
an unchanging method return value in an instance variable. I had introduced
this as a performance hack. The results were correct, but when the object was
serialized it took a huge amount of data with it (known_resource_types, and
from there...)
As the underlying costly operation is no longer being called (thanks to Paul and
Jesse) the caching is no longer a significant performance boost, and can thus
simply be removed.
|
| | | | | | | | | |
|
| | | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
When a regexp is used in a selector, vim shouldn't choke if more than one slash
is found on the same line.
This patch also adds support for colons in attribute names.
|
| | | | | | | | | |
|
| | | |_|_|_|/ /
| |/| | | | | |
|
| |\ \ \ \ \ \ \
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
Conflicts resolved manually, by Paul Berry:
lib/puppet/parser/ast/astarray.rb
lib/puppet/parser/grammar.ra
lib/puppet/parser/parser.rb (by rebuilding from grammar.ra)
|
| | | |/ / / / /
| |/| | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
or node occurs in a conditional construct so that it contains the
proper line number.
|
| | | |_|_|_|/
| |/| | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Previously, ASTArray#evaluate() was responsible for checking whether
the user had tried to declare a class, define, or node in a prohibited
location (such as a conditional construct). This meant that errors
would only be reported to the user if the conditional code was
actually evaluated.
Moved the checking into the parser, so that errors are always
reported.
|
| |\ \ \ \ \ \ |
|
| | | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Changed the grammar so that the following "plural" constructs always
parse as an ASTArray:
- funcvalues
- rvalues
- resourceinstances
- anyparams
- params
- caseopts
- casevalues
And the following "singluar" construct never parses as an ASTArray:
- statement
The previous behavior was for these constructs to parse as a scalar
when they represented a single item and an ASTArray when they
contained zero or multiple items. ("Statement" could sometimes
represent a single item because a single resource declaration could
represent multiple resources). This complicated other grammar rules
and caused ambiguous handling of nested arrays.
Also made these changes to the AST class hierarchy:
- ResourceInstance no longer derives from ASTArray. This relationship
was not meaningful because a ResourceInstance is a (title,
parameters) pair, not an array, and it produced complications when
we wanted to represent an array of ResourceInstance objects.
- Resource no longer derives from ResourceReference. No significant
functionality was being inherited and the relationship doesn't make
sense in an AST context.
- ResourceOverride no longer derives from Resource. No significant
functionality was being inherited and the relationship doesn't make
sense in an AST context.
- Resource can now represent a compound resource instance such as
"notify { foo: ; bar: }". This saves the parser from having to
use represent a statement as an array of objects.
- ASTArray's evaluate method never flattens out arrays of arrays.
|
| |\ \ \ \ \ \ \ |
|
| | | |/ / / / /
| |/| | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
The test in question (test_parse_line) was nondeterministic because it
was relying on the sort order of a Hash whose keys were symbols. When
the sort order caused a blank line to appear at the end of the file
under test, the blank line was elided by the crontab parser, causing a
failure.
Modified the test to execute in a deterministic order that doesn't
place the blank line at the end.
|
