From 1bf50badd943e684a56a03392ef0ddafefca0ad7 Mon Sep 17 00:00:00 2001 From: schneems Date: Fri, 24 Jul 2015 22:25:41 -0500 Subject: Decrease string allocations in apply_inflections In `apply_inflections` a string is down cased and some whitespace stripped in the front (which allocate strings). This would normally be fine, however `uncountables` is a fairly small array (10 elements out of the box) and this method gets called a TON. Instead we can keep an array of valid regexes for each uncountable so we don't have to allocate new strings. This change buys us 325,106 bytes of memory and 3,251 fewer objects per request. --- .../lib/active_support/inflector/inflections.rb | 37 ++++++++++++++++++++-- .../lib/active_support/inflector/methods.rb | 4 +-- 2 files changed, 36 insertions(+), 5 deletions(-) (limited to 'activesupport') diff --git a/activesupport/lib/active_support/inflector/inflections.rb b/activesupport/lib/active_support/inflector/inflections.rb index 486838bd15..8ac1820776 100644 --- a/activesupport/lib/active_support/inflector/inflections.rb +++ b/activesupport/lib/active_support/inflector/inflections.rb @@ -27,6 +27,37 @@ module ActiveSupport class Inflections @__instance__ = ThreadSafe::Cache.new + class Uncountables < Array + def initialize + @regex_array = [] + super + end + + def delete(entry) + super entry + @regex_array.delete(to_regex(entry)) + end + + def <<(*word) + add(word) + end + + def add(words) + self.concat(words.flatten.map(&:downcase)) + @regex_array += self.map {|word| to_regex(word) } + self + end + + def uncountable?(str) + @regex_array.detect {|regex| regex.match(str) } + end + + private + def to_regex(string) + /\b#{::Regexp.escape(string)}\Z/i + end + end + def self.instance(locale = :en) @__instance__[locale] ||= new end @@ -34,7 +65,7 @@ module ActiveSupport attr_reader :plurals, :singulars, :uncountables, :humans, :acronyms, :acronym_regex def initialize - @plurals, @singulars, @uncountables, @humans, @acronyms, @acronym_regex = [], [], [], [], {}, /(?=a)b/ + @plurals, @singulars, @uncountables, @humans, @acronyms, @acronym_regex = [], [], Uncountables.new, [], {}, /(?=a)b/ end # Private, for the test suite. @@ -160,7 +191,7 @@ module ActiveSupport # uncountable 'money', 'information' # uncountable %w( money information rice ) def uncountable(*words) - @uncountables += words.flatten.map(&:downcase) + @uncountables.add(words) end # Specifies a humanized form of a string by a regular expression rule or @@ -185,7 +216,7 @@ module ActiveSupport def clear(scope = :all) case scope when :all - @plurals, @singulars, @uncountables, @humans = [], [], [], [] + @plurals, @singulars, @uncountables, @humans = [], [], Uncountables.new, [] else instance_variable_set "@#{scope}", [] end diff --git a/activesupport/lib/active_support/inflector/methods.rb b/activesupport/lib/active_support/inflector/methods.rb index 60ef249e37..899ba70af6 100644 --- a/activesupport/lib/active_support/inflector/methods.rb +++ b/activesupport/lib/active_support/inflector/methods.rb @@ -354,7 +354,7 @@ module ActiveSupport # const_regexp("Foo::Bar::Baz") # => "Foo(::Bar(::Baz)?)?" # const_regexp("::") # => "::" def const_regexp(camel_cased_word) #:nodoc: - parts = camel_cased_word.split("::") + parts = camel_cased_word.split("::".freeze) return Regexp.escape(camel_cased_word) if parts.blank? @@ -372,7 +372,7 @@ module ActiveSupport def apply_inflections(word, rules) result = word.to_s.dup - if word.empty? || inflections.uncountables.include?(result.downcase[/\b\w+\Z/]) + if word.empty? || inflections.uncountables.uncountable?(result) result else rules.each { |(rule, replacement)| break if result.sub!(rule, replacement) } -- cgit v1.2.3 From 57ba9cbc6ccaa20a1ac3ca671d17040946839db3 Mon Sep 17 00:00:00 2001 From: schneems Date: Sun, 26 Jul 2015 13:59:12 -0500 Subject: Decrease allocations in transliterate We can save a few objects by freezing the `replacement` string. We save a few more by down-casing the string in memory instead of allocating a new one. We save far more objects by checking for the default separator `"-"`, and using pre-generated regular expressions. We will save 209,231 bytes and 1,322 objects. --- .../lib/active_support/inflector/transliterate.rb | 18 +++++++++++++----- 1 file changed, 13 insertions(+), 5 deletions(-) (limited to 'activesupport') diff --git a/activesupport/lib/active_support/inflector/transliterate.rb b/activesupport/lib/active_support/inflector/transliterate.rb index 2c03956672..7b28eeb6e2 100644 --- a/activesupport/lib/active_support/inflector/transliterate.rb +++ b/activesupport/lib/active_support/inflector/transliterate.rb @@ -58,7 +58,7 @@ module ActiveSupport # I18n.locale = :de # transliterate('Jürgen') # # => "Juergen" - def transliterate(string, replacement = "?") + def transliterate(string, replacement = "?".freeze) I18n.transliterate(ActiveSupport::Multibyte::Unicode.normalize( ActiveSupport::Multibyte::Unicode.tidy_bytes(string), :c), :replacement => replacement) @@ -75,13 +75,21 @@ module ActiveSupport # Turn unwanted chars into the separator parameterized_string.gsub!(/[^a-z0-9\-_]+/i, sep) unless sep.nil? || sep.empty? - re_sep = Regexp.escape(sep) + if sep == "-".freeze + re_duplicate_seperator = /-{2,}/ + re_leading_trailing_separator = /^-|-$/i + else + re_sep = Regexp.escape(sep) + re_duplicate_seperator = /#{re_sep}{2,}/ + re_leading_trailing_separator = /^#{re_sep}|#{re_sep}$/i + end # No more than one of the separator in a row. - parameterized_string.gsub!(/#{re_sep}{2,}/, sep) + parameterized_string.gsub!(re_duplicate_seperator, sep) # Remove leading/trailing separator. - parameterized_string.gsub!(/^#{re_sep}|#{re_sep}$/i, ''.freeze) + parameterized_string.gsub!(re_leading_trailing_separator, ''.freeze) end - parameterized_string.downcase + parameterized_string.downcase! + parameterized_string end end end -- cgit v1.2.3 From 0d7a714dcae69668429c4b79e36f1b47d6c898e4 Mon Sep 17 00:00:00 2001 From: schneems Date: Sun, 26 Jul 2015 14:31:18 -0500 Subject: String#freeze optimizations --- activesupport/lib/active_support/core_ext/string/inflections.rb | 2 +- activesupport/lib/active_support/multibyte/unicode.rb | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'activesupport') diff --git a/activesupport/lib/active_support/core_ext/string/inflections.rb b/activesupport/lib/active_support/core_ext/string/inflections.rb index 97f9720b2b..0d5e02cd77 100644 --- a/activesupport/lib/active_support/core_ext/string/inflections.rb +++ b/activesupport/lib/active_support/core_ext/string/inflections.rb @@ -164,7 +164,7 @@ class String # # <%= link_to(@person.name, person_path) %> # # => Donald E. Knuth - def parameterize(sep = '-') + def parameterize(sep = '-'.freeze) ActiveSupport::Inflector.parameterize(self, sep) end diff --git a/activesupport/lib/active_support/multibyte/unicode.rb b/activesupport/lib/active_support/multibyte/unicode.rb index f4de4a80d9..4de7fa4e4b 100644 --- a/activesupport/lib/active_support/multibyte/unicode.rb +++ b/activesupport/lib/active_support/multibyte/unicode.rb @@ -273,7 +273,7 @@ module ActiveSupport compose(reorder_characters(decompose(:compatibility, codepoints))) else raise ArgumentError, "#{form} is not a valid normalization variant", caller - end.pack('U*') + end.pack('U*'.freeze) end def downcase(string) -- cgit v1.2.3