logfmt 0.0.3 → 0.0.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 8ea9ce7afd32bfdaa73e9399460c34f6c5cf2a60
-  data.tar.gz: c54d8967a623133fbabfbf50d160fb284e62a207
+  metadata.gz: ebbeb9a0b99c4d279a682953bd40f6585acfc8ff
+  data.tar.gz: 0c24353022ea70cf672b07cc9126de375b0400d2
 SHA512:
-  metadata.gz: 451bd0bc7ab6db03156c1bcd16277af82f36a597eef4d2a8e5ec7a989cbeb1224662a497918523478dc64850f868afb242a954453a882165027e27387ae186ae
-  data.tar.gz: 38780bd602894e7a57800ec52bbbe0325d476f155f5e09de3964232268d1906932576f576ce5d0d797dc34840d221e616f046a48330e19d40df4f0c94574a91f
+  metadata.gz: a14c02145816f5ec84487ef96b4fff690c6d1692bd3209c59eb58db7bf1af5b20148e19c6d56d4b99eb6ba8175e540b61852ddac2748441552fcc9d4faf7f4b8
+  data.tar.gz: 5db326288526225bd1084acaf9ef2fb57a785de7b0f6f7086c92b0e0a4eae58252f052632897f49d60dca488b03f5bd68c4cebe8d263e086cb5a94ada3774c6d

data/bench.rb

@@ -0,0 +1,15 @@
+$:.unshift File.expand_path('../lib', __FILE__)
+
+require 'benchmark'
+require 'logfmt'
+
+N = 1_000
+line = 'foo=bar a=14 baz="hello kitty" ƒ=2h3s cool%story=bro f %^asdf'
+
+Benchmark.bm(20) do |x|
+  x.report('char-by-char') do
+    N.times do
+      Logfmt.parse(line)
+    end
+  end
+end

data/lib/logfmt/parser.rb

@@ -1,49 +1,88 @@
-require "parslet"
-
 module Logfmt
-  class Parser < Parslet::Parser
-    rule(:ident_byte) { match['^\s"='] }
-    rule(:string_byte) { match['^\\\\"'] }
-    rule(:garbage) { match['\s"='].repeat }
-    rule(:ident) { ident_byte >> ident_byte.repeat }
-    rule(:key) { ident }
-    rule(:value) {
-      ident.as(:value) |
-      str('"') >> (string_byte | str('\"')).repeat.as(:value) >> str('"')
-    }
-    rule(:pair) {
-      (key.as(:key) >> str('=') >> value) |
-      (key.as(:key) >> str('=')) |
-      key.as(:key)
-    }
-    rule(:message) {
-      (garbage >> pair.as(:pair)).repeat.as(:message) >> garbage
-    }
-    root(:message)
-  end
-
-  class Transformer < Parslet::Transform
-    class Pair < Struct.new(:key, :val); end
-
-    rule(:message => subtree(:ob)) {
-      (ob.is_a?(Array) ? ob : [ ob ]).inject({}) { |h, p| h[p.key] = p.val; h }
-    }
-    rule(:pair => { :key => simple(:key), :value => simple(:val) }) {
-      Pair.new(key.to_s, val.to_s)
-    }
-    rule(:pair => { :key => simple(:key), :value => sequence(:val) }) {
-      Pair.new(key.to_s, "")
-    }
-    rule(:pair => { :key => simple(:key) }) {
-      Pair.new(key.to_s, true)
-    }
-  end
-
-  def self.parse(logs)
-    parser = Parser.new
-    transformer = Transformer.new
+  GARBAGE = 0
+  KEY = 1
+  EQUAL = 2
+  IVALUE = 3
+  QVALUE = 4
 
-    tree = parser.parse(logs)
-    transformer.apply(tree)
+  def self.parse(line)
+    output = {}
+    key, value = "", ""
+    escaped = false
+    state = GARBAGE
+    i = 0
+    line.each_char do |c|
+      i += 1
+      if state == GARBAGE
+        if c > ' ' && c != '"' && c != '='
+          key = c
+          state = KEY
+        end
+        next
+      end
+      if state == KEY
+        if c > ' ' && c != '"' && c != '='
+          state = KEY
+          key << c
+        elsif c == '='
+          output[key.strip()] = true
+          state = EQUAL
+        else
+          output[key.strip()] = true
+          state = GARBAGE
+        end
+        if i >= line.length
+          output[key.strip()] = true
+        end
+        next
+      end
+      if state == EQUAL
+        if c > ' ' && c != '"' && c != '='
+          value = c
+          state = IVALUE
+        elsif c == '"'
+          value = ""
+          escaped = false
+          state = QVALUE
+        else
+          state = GARBAGE
+        end
+        if i >= line.length
+          output[key.strip()] = true
+        end
+        next
+      end
+      if state == IVALUE
+        if not (c > ' ' && c != '"' && c != '=')
+          output[key.strip()] = value
+          state = GARBAGE
+        else
+          value << c
+        end
+        if i >= line.length
+          output[key.strip()] = value
+        end
+        next
+      end
+      if state == QVALUE
+        if c == '\\'
+          escaped = true
+          value << "\\"
+        elsif c == '"'
+          if escaped
+            escaped = false
+            value << c
+            next
+          end
+          output[key.strip()] = value
+          state = GARBAGE
+        else
+          escaped = false
+          value << c
+        end
+        next
+      end
+    end
+    output
   end
 end

data/lib/logfmt/version.rb

@@ -1,3 +1,3 @@
 module Logfmt
-  VERSION = "0.0.3"
+  VERSION = "0.0.4"
 end

data/spec/logfmt/parser_spec.rb

@@ -2,7 +2,7 @@
 require 'spec_helper'
 require 'logfmt/parser'
 
-describe Logfmt::Parser do
+describe Logfmt do
   it 'parse empty log line' do
     data = Logfmt.parse("")
     expect(data).to eq({})
@@ -55,12 +55,12 @@ describe Logfmt::Parser do
 
   it 'parse escaped quote value ' do
     data = Logfmt.parse('key="quoted \" value" r="esc\t"')
-    expect(data).to eq({"key" => 'quoted " value', "r" => "esc\t"})
+    expect(data).to eq({"key" => 'quoted \" value', "r" => "esc\\t"})
   end
 
   it 'parse mixed pairs' do
     data = Logfmt.parse('key1="quoted \" value" key2 key3=value3')
-    expect(data).to eq({"key1" => 'quoted " value', "key2" => true, "key3" => "value3"})
+    expect(data).to eq({"key1" => 'quoted \" value', "key2" => true, "key3" => "value3"})
   end
 
   it 'parse mixed characters pairs' do

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: logfmt
 version: !ruby/object:Gem::Version
-  version: 0.0.3
+  version: 0.0.4
 platform: ruby
 authors:
 - Timothée Peignier
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-03-15 00:00:00.000000000 Z
+date: 2014-04-24 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: parslet
@@ -65,6 +65,7 @@ files:
 - Gemfile
 - README.md
 - Rakefile
+- bench.rb
 - lib/logfmt.rb
 - lib/logfmt/parser.rb
 - lib/logfmt/version.rb