Add #start_with? and #end_with? mutations
#1201
Conversation
dgollahon
force-pushed
the
start_with-end_with-mutations
branch
from
515e3de to
0a5ade5
Compare
dgollahon
changed the title
/\Astatic/ -> #start_with? mutations#start_with? and #end_with? mutations
| # | ||
| # @return [Parser::AST::Node] | ||
| def self.expand_regexp_ast(node) | ||
| *body, _opts = node.children |
There was a problem hiding this comment.
Note that destructuring instead of using #slice avoids the mutation we chose to ignore before.
There was a problem hiding this comment.
As a followup we could also promote interpolations.
a.match?(/\Afoo#{dynamic}/)
to:
a.end_with?(dynamic)
There was a problem hiding this comment.
Yeah, something like that might be interesting but I am not sure I will go that deep right away. I have a couple of other small improvements for similar cases I am thinking of as well that are pending.
dgollahon
force-pushed
the
start_with-end_with-mutations
branch
from
cfbe4f1 to
6533a8d
Compare
|
|
||
| RSpec.describe Mutant::AST::Regexp, '.expand_regexp_ast' do | ||
| it 'returns the expanded AST' do | ||
| # /foo/ |
There was a problem hiding this comment.
Non blocking: Instead of this comment: what about running the regexp parser / AST transform to get the parser_ast ? This way this test is more meaningful and would have canary properties for future changes?
There was a problem hiding this comment.
I think that makes sense--I have updated this to match. When I was originally writing the spec it felt like it might be clearer to have the raw AST so I started with that but now that I have just swapped it out for Unparser.parse I think it's a lot simpler to read and, as you say, it has a slight "canary" property.
dgollahon
force-pushed
the
start_with-end_with-mutations
branch
from
dd5ab04 to
e0042ad
Compare
- Adds the following mutations:
* `a.match(/\Atext/)` -> `b.start_with?('text')`
* `a.match?(/\Atext/)` -> `b.start_with?('text')`
* `a =~ /\Atext/` -> `b.start_with?('text')`
- NOTE: Adds a `Mutant::AST::Regexp.expand_regexp_ast` to avoid repeating AST expansion logic in the `send` mutator. At that level, the node appears as a simple `parser` `s(:regexp ...)` node so I chose to fully parse the regexp rather than try to do a string pattern.
- Closes #169
- Adds the following mutations:
* `a.match(/text\z/)` -> `b.end_with?('text')`
* `a.match?(/text\z/)` -> `b.end_with?('text')`
* `a =~ /text\z/` -> `b.end_with?('text')`
dgollahon
force-pushed
the
start_with-end_with-mutations
branch
from
c106234 to
635a529
Compare
a.match(/\Atext/)->b.start_with?('text')a.match?(/\Atext/)->b.start_with?('text')a =~ /\Atext/->b.start_with?('text')Mutant::AST::Regexp.expand_regexp_astto avoid repeating AST expansion logic in thesendmutator. At that level, the node appears as a simpleparsers(:regexp ...)node so I chose to fully parse the regexp rather than try to do a string pattern.a.match(/text\z/)->b.end_with?('text')a.match?(/text\z/)->b.end_with?('text')a =~ /text\z/->b.end_with?('text')