Class: RSpec::Core::FilterableItemRepository::QueryOptimized
- Inherits:
-
UpdateOptimized
- Object
- UpdateOptimized
- RSpec::Core::FilterableItemRepository::QueryOptimized
- Defined in:
- lib/rubypitaya/app-template/vendor/bundle/ruby/3.1.0/gems/rspec-core-3.12.0/lib/rspec/core/metadata_filter.rb
Overview
This implementation is much more complex, and is optimized for rare (or hopefully no) updates once the queries start. Updates incur a cost as it has to clear the memoization and keep track of applicable keys. Queries will be O(N) the first time an item is provided with a given set of applicable metadata; subsequent queries with items with the same set of applicable metadata will be O(1) due to internal memoization.
This is ideal for use by config, where filterable items (e.g. hooks) are typically added at the start of the process (e.g. in ‘spec_helper`) and then repeatedly queried as example groups and examples are defined.
Instance Attribute Summary
Attributes inherited from UpdateOptimized
Instance Method Summary collapse
- #append(item, metadata) ⇒ Object
- #delete(item, metadata) ⇒ Object
-
#initialize(applies_predicate) ⇒ QueryOptimized
constructor
A new instance of QueryOptimized.
- #items_for(metadata) ⇒ Object
- #prepend(item, metadata) ⇒ Object
Constructor Details
#initialize(applies_predicate) ⇒ QueryOptimized
Returns a new instance of QueryOptimized.
155 156 157 158 159 160 161 162 |
# File 'lib/rubypitaya/app-template/vendor/bundle/ruby/3.1.0/gems/rspec-core-3.12.0/lib/rspec/core/metadata_filter.rb', line 155 def initialize(applies_predicate) super @applicable_keys = Set.new @proc_keys = Set.new @memoized_lookups = Hash.new do |hash, | hash[] = find_items_for() end end |
Instance Method Details
#append(item, metadata) ⇒ Object
164 165 166 167 |
# File 'lib/rubypitaya/app-template/vendor/bundle/ruby/3.1.0/gems/rspec-core-3.12.0/lib/rspec/core/metadata_filter.rb', line 164 def append(item, ) super handle_mutation() end |
#delete(item, metadata) ⇒ Object
174 175 176 177 |
# File 'lib/rubypitaya/app-template/vendor/bundle/ruby/3.1.0/gems/rspec-core-3.12.0/lib/rspec/core/metadata_filter.rb', line 174 def delete(item, ) super reconstruct_caches end |
#items_for(metadata) ⇒ Object
179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 |
# File 'lib/rubypitaya/app-template/vendor/bundle/ruby/3.1.0/gems/rspec-core-3.12.0/lib/rspec/core/metadata_filter.rb', line 179 def items_for() # The filtering of `metadata` to `applicable_metadata` is the key thing # that makes the memoization actually useful in practice, since each # example and example group have different metadata (e.g. location and # description). By filtering to the metadata keys our items care about, # we can ignore extra metadata keys that differ for each example/group. # For example, given `config.include DBHelpers, :db`, example groups # can be split into these two sets: those that are tagged with `:db` and those # that are not. For each set, this method for the first group in the set is # still an `O(N)` calculation, but all subsequent groups in the set will be # constant time lookups when they call this method. = () if .any? { |k, _| @proc_keys.include?(k) } # It's unsafe to memoize lookups involving procs (since they can # be non-deterministic), so we skip the memoization in this case. find_items_for() else @memoized_lookups[] end end |
#prepend(item, metadata) ⇒ Object
169 170 171 172 |
# File 'lib/rubypitaya/app-template/vendor/bundle/ruby/3.1.0/gems/rspec-core-3.12.0/lib/rspec/core/metadata_filter.rb', line 169 def prepend(item, ) super handle_mutation() end |