Add matcher config in configuration file #1176
Conversation
mbj
force-pushed
the
add/matcher-config
branch
from
6ab43f4
to
b8b207f
Compare
mbj
force-pushed
the
add/matcher-config
branch
3 times, most recently
from
af24813
to
c899bf9
Compare
mbj
force-pushed
the
add/matcher-config
branch
4 times, most recently
from
79dc889
to
53b9e40
Compare
| @@ -159,13 +112,14 @@ def self.env | |||
| Transform::Exception.new(YAML::SyntaxError, YAML.method(:safe_load)), | |||
| Transform::Hash.new( | |||
| optional: [ | |||
| Transform::Hash::Key.new('coverage_criteria', CoverageCriteria::TRANSFORM), | |||
| Transform::Hash::Key.new('coverage_criteria', ->(value) { CoverageCriteria::TRANSFORM.call(value) }), | |||
There was a problem hiding this comment.
I "really" like this require sequence:
require 'foo'
require 'foo/bar'
But we have to access the constant from foo/bar here, hence as we are in class scope and the require for `foo/bar' did not happen yet we have to use a lambda indirection.
The alternative would be:
require 'foo/bar'
require 'foo'
which sucks. For the moment I eat it. Would love of Ruby had a module system ;).
There was a problem hiding this comment.
Would love of Ruby had a module system ;)
yeahh... I feel that one!
| @@ -64,7 +64,7 @@ This prints a report like: | |||
|
|
|||
| ```sh | |||
| Mutant environment: | |||
| Matcher: #<Mutant::Matcher::Config match_expressions: [AUOM*]> | |||
| Matcher: #<Mutant::Matcher::Config subjects: [AUOM*]> | |||
There was a problem hiding this comment.
This is a good change I think--it reads more clearly.
| @@ -64,7 +64,7 @@ This prints a report like: | |||
|
|
|||
| ```sh | |||
| Mutant environment: | |||
| Matcher: #<Mutant::Matcher::Config match_expressions: [AUOM*]> | |||
| Matcher: #<Mutant::Matcher::Config subjects: [AUOM*]> | |||
There was a problem hiding this comment.
I think this is much nicer to read. Its clear the domain of this method is matching, else the class would not be named Mutant::Matcher.
So the match_ prefix was redundant always.
Just naming the key expressions would IMO also suck as it than would have a sibling ignore_expressions which does not clearly indicate what the difference between both would be.
Hence replacing this with subjects makes it very clear. And subject_expressions is kind of redundant as we can only match expressions.
| @@ -41,6 +41,7 @@ module Mutant | |||
| SCOPE_OPERATOR = '::' | |||
| end # Mutant | |||
|
|
|||
| require 'mutant/transform' | |||
There was a problem hiding this comment.
The fact this has to move up the require chain to toplevel, indicates it likely belongs to a separate library.
Once I make nicer "input error location rendering" this is worthy of a separate lib.
| Transform::Hash::Key.new('fail_fast', Transform::BOOLEAN), | ||
| Transform::Hash::Key.new('includes', Transform::STRING_ARRAY), | ||
| Transform::Hash::Key.new('integration', Transform::STRING), | ||
| Transform::Hash::Key.new('jobs', Transform::INTEGER), | ||
| Transform::Hash::Key.new('mutation_timeout', Transform::FLOAT), | ||
| Transform::Hash::Key.new('requires', Transform::STRING_ARRAY) | ||
| Transform::Hash::Key.new('requires', Transform::STRING_ARRAY), | ||
| Transform::Hash::Key.new('matcher', Matcher::Config::LOADER) |
There was a problem hiding this comment.
This one, paired with adding the constant Mutant::Config::LOADER is the only real change to make config file support for matchers.
mbj
force-pushed
the
add/matcher-config
branch
from
53b9e40
to
18313e2
Compare
it should, I repushed a lot so have to manage the queue. Merging once done. |
* This class needs to traverse the process tree often, hence it tries to minimize its `marshal` representation. * For that reason we do not even use `adamantium` to avoid the memory. * A custom self freeze `.new` method was added instead. * That clashes with the one `abstract_type` installs before the re-definition of `.new`. Turning `abstract_type` semantics unobservable. * Removing abstract type for this class is the local optimum for the moment.
No description provided.