lz4-ruby 0.1.2 → 0.1.3

data/README.rdoc

@@ -22,7 +22,7 @@ Ruby bindings for {LZ4}[http://code.google.com/p/lz4/].
 
 == Benchmark
 
-Tested on VirtualBox VM : 2-core, 4GB RAM (Host : Core i5-2520M / 8GB RAM).
+Tested on VirtualBox VM : 2-core / 4GB RAM (Host : Core i5-2520M / 8GB RAM).
 
 === {enwik8}[http://mattmahoney.net/dc/enwik8.zip]
     method  ratio(bpc)      comp.time(ms)   uncomp.time(ms)
@@ -34,7 +34,6 @@ Tested on VirtualBox VM : 2-core, 4GB RAM (Host : Core i5-2520M / 8GB RAM).
 == TODO
 
 * Write API documents.
-* Write test codes.
 * Support mswin32 platform.
 
 == Copyright

data/VERSION

@@ -1 +1 @@
-0.1.2
+0.1.3

data/ext/extconf.rb

@@ -4,8 +4,9 @@ LZ4_REV_NO = 66
 
 URL = "http://lz4.googlecode.com/svn-history/r#{LZ4_REV_NO}/trunk/"
 
-def download_from_web(url)
-  `wget #{url}`
+def download_from_web(file)
+  File.delete(file) if File.file?(file)
+  `wget #{URL}/#{file}`
   
   if $?.exitstatus != 0
     return false
@@ -14,10 +15,12 @@ def download_from_web(url)
   return true
 end
 
-return if !download_from_web("#{URL}/lz4.c")
-return if !download_from_web("#{URL}/lz4.h")
-return if !download_from_web("#{URL}/lz4hc.c")
-return if !download_from_web("#{URL}/lz4hc.h")
+[ "lz4.c",
+  "lz4.h",
+  "lz4hc.c",
+  "lz4hc.h" ].each do |filename|
+  exit if !download_from_web(filename)
+end
 
 $CFLAGS += " -Wall "
 

data/ext/lz4ruby.c

@@ -7,11 +7,58 @@ typedef int (*CompressFunc)(const char *source, char *dest, int isize);
 static VALUE lz4;
 static VALUE lz4_error;
 
+static int encode_varbyte(int value, char *buf) {
+  buf[0] = value & 0x7f; value >>= 7;
+  if (value == 0) { return 1; }
+  buf[0] |= 0x80;
+  
+  buf[1] = value & 0x7f; value >>= 7;
+  if (value == 0) { return 2; }
+  buf[1] |= 0x80;
+  
+  buf[2] = value & 0x7f; value >>= 7;
+  if (value == 0) { return 3; }
+  buf[2] |= 0x80;
+  
+  buf[3] = value & 0x7f; value >>= 7;
+  if (value == 0) { return 4; }
+  buf[3] |= 0x80;
+
+  buf[4] = value & 0x7f;
+  return 5;
+}
+
+static int decode_varbyte(const char *src, int len, int *value) {
+  if (len < 1) { return -1; }
+
+  *value = src[0] & 0x7f;
+  if ((src[0] & 0x80) == 0) { return 1; }
+  if (len < 2) { return -1; }
+
+  *value |= (src[1] & 0x7f) << 7;
+  if ((src[1] & 0x80) == 0) { return 2; }
+  if (len < 3) { return -1; }
+
+  *value |= (src[2] & 0x7f) << 14;
+  if ((src[2] & 0x80) == 0) { return 3; }
+  if (len < 4) { return -1; }
+
+  *value |= (src[3] & 0x7f) << 21;
+  if ((src[3] & 0x80) == 0) { return 4; }
+  if (len < 5) { return -1; }
+
+  *value |= (src[4] & 0x7f) << 28;
+
+  return 5;
+}
+
 static VALUE compress(CompressFunc compressor, VALUE self, VALUE source) {
   const char *src_p = NULL;
+  char varbyte[5];
   char *buf = NULL;
   VALUE result;
   int src_size;
+  int varbyte_len;
   int buf_size;
   int comp_size;
 
@@ -20,16 +67,15 @@ static VALUE compress(CompressFunc compressor, VALUE self, VALUE source) {
   src_size = RSTRING_LEN(source);
   buf_size = LZ4_compressBound(src_size);
 
-  result = rb_str_new(NULL, buf_size + 4);
+  varbyte_len = encode_varbyte(src_size, varbyte);
+
+  result = rb_str_new(NULL, buf_size + varbyte_len);
   buf = RSTRING_PTR(result);
 
-  buf[0] = (char)((src_size >> 24) & 0xff);
-  buf[1] = (char)((src_size >> 16) & 0xff);
-  buf[2] = (char)((src_size >> 8) & 0xff);
-  buf[3] = (char)(src_size & 0xff);
+  memcpy(buf, varbyte, varbyte_len);
 
-  comp_size = compressor(src_p, buf + 4, src_size);
-  rb_str_resize(result, comp_size + 4);
+  comp_size = compressor(src_p, buf + varbyte_len, src_size);
+  rb_str_resize(result, comp_size + varbyte_len);
 
   return result;
 }
@@ -47,22 +93,20 @@ static VALUE lz4_ruby_uncompress(VALUE self, VALUE source) {
   char *buf = NULL;
   VALUE result;
   int src_size;
-  int buf_size;
+  int varbyte_len;
+  int buf_size = 0;
   int read_bytes;
 
   Check_Type(source, T_STRING);
   src_p = RSTRING_PTR(source);
   src_size = RSTRING_LEN(source);
 
-  buf_size = ((src_p[0] & 0xffU) << 24)
-    | ((src_p[1] & 0xffU) << 16)
-    | ((src_p[2] & 0xffU) << 8)
-    | (src_p[3] & 0xffU);
+  varbyte_len = decode_varbyte(src_p, src_size, &buf_size);
 
   result = rb_str_new(NULL, buf_size);
   buf = RSTRING_PTR(result);
 
-  read_bytes = LZ4_uncompress(src_p + 4, buf, buf_size);
+  read_bytes = LZ4_uncompress(src_p + varbyte_len, buf, buf_size);
   if (read_bytes < 0) {
     rb_raise(lz4_error, "Compressed data is maybe corrupted.");
   }

data/lz4-ruby.gemspec

@@ -5,11 +5,11 @@
 
 Gem::Specification.new do |s|
   s.name = "lz4-ruby"
-  s.version = "0.1.2"
+  s.version = "0.1.3"
 
   s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
   s.authors = ["KOMIYA Atsushi"]
-  s.date = "2012-06-04"
+  s.date = "2012-06-05"
   s.description = "Ruby bindings for LZ4."
   s.email = "komiya.atsushi@gmail.com"
   s.extensions = ["ext/extconf.rb"]

data/test/test_lz4-ruby.rb

@@ -1,8 +1,18 @@
 require 'helper'
 
 class TestLz4Ruby < Test::Unit::TestCase
+  LOOP_COUNT = 257
+  
+  @@random = Random.new(123)
+
   context "LZ4::compress" do
-    257.times do |t|
+    should "empty text" do
+      compressed = LZ4::compress("")
+      uncompressed = LZ4::uncompress(compressed)
+      assert_empty("", uncompressed)
+    end
+    
+    LOOP_COUNT.times do |t|
       len = t + 1
       text = "a" * len
 
@@ -12,5 +22,46 @@ class TestLz4Ruby < Test::Unit::TestCase
         assert_equal(text, uncompressed)
       end
     end
+
+    LOOP_COUNT.times do |t|
+      len = t + 1
+      text = @@random.bytes(len)
+      
+      should "random text of #{len} bytes" do
+        compressed = LZ4::compress(text)
+        uncompressed = LZ4::uncompress(compressed)
+        assert_equal(text, uncompressed)
+      end
+    end
+  end
+
+  context "LZ4::compressHC" do
+    should "empty text" do
+      compressed = LZ4::compressHC("")
+      uncompressed = LZ4::uncompress(compressed)
+      assert_empty("", uncompressed)
+    end
+    
+    LOOP_COUNT.times do |t|
+      len = t + 1
+      text = "a" * len
+
+      should "text of #{len} \"a\"'s" do
+        compressed = LZ4::compressHC(text)
+        uncompressed = LZ4::uncompress(compressed)
+        assert_equal(text, uncompressed)
+      end
+    end
+
+    LOOP_COUNT.times do |t|
+      len = t + 1
+      text = @@random.bytes(len)
+      
+      should "random text of #{len} bytes" do
+        compressed = LZ4::compressHC(text)
+        uncompressed = LZ4::uncompress(compressed)
+        assert_equal(text, uncompressed)
+      end
+    end
   end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: lz4-ruby
 version: !ruby/object:Gem::Version
-  version: 0.1.2
+  version: 0.1.3
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2012-06-04 00:00:00.000000000 Z
+date: 2012-06-05 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: shoulda
@@ -116,7 +116,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: 492549213
+      hash: -379957487
 required_rubygems_version: !ruby/object:Gem::Requirement
   none: false
   requirements: