zscan 1.0 → 1.0.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: c343abbf9b043bcb7643ac193bb5f9a0fd76245e
-  data.tar.gz: fe57d1e6e352cad9d9ac2f8095afdbbd4affbebe
+  metadata.gz: 8baa1419230fe03d1ec41cfa65202dce10d0512c
+  data.tar.gz: 0cf1335b2d7221ca4c6329dd6e8e2b5b0f3ac59d
 SHA512:
-  metadata.gz: 687a22c21f5da837039b557365807a89b5afd33cea5301baa1d5321c3a5b324f63f7b96524166f388da1c296e2e775ce8007e97281ad084c1583aa20c247ff6f
-  data.tar.gz: d6fde3485b2bb8e4db9bb80513c370dec17e331c1e20c5eb087742d86fdb06c403136b98be3e14dc65747b897182f4eaa0132ebc454b427e1e5b1cc622a1c8e0
+  metadata.gz: a9ccf8455843673336365dd446f808cb52f1538a53c86cc69d102640e4a6afb40332e3c65f6f7a5a4178895a1095571e4328a501e64797a01a6af1cdc4256549
+  data.tar.gz: 8c3d109803e0e9d8e91b8787f98190fd7e400613772f57347ff2d0a99725202c8eb3444ba11efc7a606897d78cec5d9932638292c5ba2583fb697f9e7c8c2540

data/bench.rb

@@ -0,0 +1,23 @@
+require "strscan"
+require_relative "lib/zscan"
+require "benchmark"
+
+s = "word\n"
+s *= 3_000_000
+
+puts "StringScanner:"
+sc = StringScanner.new s
+puts Benchmark.measure{
+  until sc.eos?
+    sc.scan(/\w+/) and (sc.pos += 1)
+  end
+}
+
+puts
+puts "ZScan should be nearly as fast as StringScanner"
+zc = ZScan.new s
+puts Benchmark.measure{
+  until zc.eos?
+    zc.scan(/\w+/) and zc.advance(1)
+  end
+}

data/ext/zscan.c

@@ -166,7 +166,7 @@ static VALUE zscan_match_bytesize(VALUE self, VALUE pattern) {
       return ULONG2NUM(RSTRING_LEN(pattern));
     }
   } else if (TYPE(pattern) == T_REGEXP) {
-    regex_t *re = rb_reg_prepare_re(pattern, p->s);
+    regex_t *re = rb_reg_prepare_re(pattern, p->s); // prepare with compatible encoding
     int tmpreg = re != RREGEXP(pattern)->ptr;
     if (!tmpreg) {
       RREGEXP(pattern)->usecnt++;
@@ -177,9 +177,6 @@ static VALUE zscan_match_bytesize(VALUE self, VALUE pattern) {
     UChar* ptr_match_from = (UChar*)(ptr + p->bytepos);
     long ret = onig_match(re, (UChar*)ptr, ptr_end, ptr_match_from, NULL, ONIG_OPTION_NONE);
 
-    if (!tmpreg) {
-      RREGEXP(pattern)->usecnt--;
-    }
     if (tmpreg) {
       if (RREGEXP(pattern)->usecnt) {
         onig_free(re);
@@ -187,6 +184,8 @@ static VALUE zscan_match_bytesize(VALUE self, VALUE pattern) {
         onig_free(RREGEXP(pattern)->ptr);
         RREGEXP(pattern)->ptr = re;
       }
+    } else {
+      RREGEXP(pattern)->usecnt--;
     }
 
     if (ret == -2) {
@@ -201,6 +200,21 @@ static VALUE zscan_match_bytesize(VALUE self, VALUE pattern) {
   return Qnil;
 }
 
+static VALUE zscan_scan(VALUE self, VALUE pattern) {
+  VALUE v_bytelen = zscan_match_bytesize(self, pattern);
+  if (v_bytelen == Qnil) {
+    return Qnil;
+  } else {
+    P;
+    long bytelen = NUM2LONG(v_bytelen);
+    volatile VALUE ret = rb_funcall(p->s, rb_intern("byteslice"), 2, LONG2NUM(p->bytepos), v_bytelen);
+    VALUE v_len = rb_str_length(ret);
+    p->bytepos += bytelen;
+    p->pos += NUM2LONG(v_len);
+    return ret;
+  }
+}
+
 static VALUE zscan_push(VALUE self) {
   P;
   if (p->stack_i + 1 == p->stack_cap) {
@@ -242,6 +256,27 @@ static VALUE zscan_restore(VALUE self) {
   return self;
 }
 
+static VALUE zscan_clear_pos_stack(VALUE self) {
+  P;
+  p->stack_i = 0;
+  return self;
+}
+
+static VALUE zscan_try(VALUE self) {
+  if (!rb_block_given_p()) {
+    rb_raise(rb_eRuntimeError, "need a block");
+  }
+  VALUE r;
+  zscan_push(self);
+  r = rb_yield(Qnil);
+  if (RTEST(r)) {
+    zscan_drop(self);
+  } else {
+    zscan_pop(self);
+  }
+  return r;
+}
+
 void Init_zscan() {
   VALUE zscan = rb_define_class("ZScan", rb_cObject);
   rb_define_alloc_func(zscan, zscan_alloc);
@@ -254,8 +289,11 @@ void Init_zscan() {
   rb_define_method(zscan, "advance", zscan_advance, 1);
   rb_define_method(zscan, "eos?", zscan_eos_p, 0);
   rb_define_method(zscan, "match_bytesize", zscan_match_bytesize, 1);
+  rb_define_method(zscan, "scan", zscan_scan, 1);
   rb_define_method(zscan, "push", zscan_push, 0);
   rb_define_method(zscan, "pop", zscan_pop, 0);
   rb_define_method(zscan, "drop", zscan_drop, 0);
   rb_define_method(zscan, "restore", zscan_restore, 0);
+  rb_define_method(zscan, "clear_pos_stack", zscan_clear_pos_stack, 0);
+  rb_define_method(zscan, "try", zscan_try, 0);
 }

data/lib/zscan.rb

@@ -1,7 +1,7 @@
 require_relative "../ext/zscan"
 
 class ZScan
-  VERSION = '1.0'
+  VERSION = '1.0.1'
 
   def initialize s, dup=false
     _internal_init dup ? s.dup : s
@@ -11,14 +11,6 @@ class ZScan
     _internal_string.dup
   end
 
-  def scan re_or_str
-    if sz = match_bytesize(re_or_str)
-      r = _internal_string.byteslice bytepos, sz
-      self.bytepos += sz
-      r
-    end
-  end
-
   def skip re_or_str
     if sz = match_bytesize(re_or_str)
       self.bytepos += sz
@@ -42,5 +34,31 @@ class ZScan
     self.pos = _internal_string.size
   end
 
+  def << substring
+    _internal_string << substring
+  end
+
+  def []= range, substring
+    start = range.start
+    if start < 0
+      start = _internal_string.size + start
+    end
+    if start < pos
+      self.pos = start
+    end
+    _internal_string[range] = substring
+  end
+
+  def size
+    _internal_string.size
+  end
+
+  def bytesize
+    _internal_string.bytesize
+  end
+
   private :_internal_init, :_internal_string
 end
+
+# rooooobust!
+Zscan = ZScan

data/rakefile

@@ -9,9 +9,7 @@ task :default => [:test, gem_package]
 
 desc "build and run test"
 task :test do
-  Dir.chdir "ext"
-  sh "make"
-  Dir.chdir ".."
+  sh "make -C ext"
   sh "rspec"
 end
 

data/readme.md

@@ -46,14 +46,14 @@ z.scan /^/     #=> nil
 
 See also https://bugs.ruby-lang.org/issues/7092
 
-## Methods
+## Essential methods
 
 - `ZScan.new string, dup=false`
 - `#scan regexp_or_string`
 - `#skip regexp_or_string`
-- `#match_bytesize regexp_or_string` returns length of matched bytes or nil
+- `#match_bytesize regexp_or_string` return length of matched bytes or `nil`.
 - `#eos?`
-- `#string` note: returns a COW dup
+- `#string` note: return a dup. Don't worry the performance because it is a copy-on-write string.
 - `#rest`
 
 ## Pos management
@@ -68,13 +68,23 @@ See also https://bugs.ruby-lang.org/issues/7092
 
 ## Efficient pos stack manipulation
 
-- `#push` pushes current pos into the stack.
-- `#pop` sets current pos to top of the stack, and pops it.
-- `#drop` drops top of pos stack without changing current pos.
-- `#restore` sets current pos to top of the stack.
+- `#push` push current pos into the stack.
+- `#pop` set current pos to top of the stack, and pop it.
+- `#drop` drop top of pos stack without changing current pos.
+- `#restore` set current pos to top of the stack.
+- `#clear_pos_stack` clear pos stack.
+- `#try` try to do several scans in the given block, fall back to init pos if block returns `nil` or `false`. Returns block's return, can be nested.
+
+## String delegates
+
+- `#<< append_string`
+- `#[]= range, replace_string` note: if `range` starts before pos, moves pos left, also clears the stack.
+- `#size`
+- `#bytesize`
 
 ## License
 
+```
 Copyright (C) 2013 by Zete Lui (BSD)
 
 Permission is hereby granted, free of charge, to any person obtaining a copy of
@@ -93,3 +103,4 @@ FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR
 COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
 IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
 CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+```

data/spec/zscan_spec.rb

@@ -64,4 +64,33 @@ describe ZScan do
     assert_equal '', z.scan(/(?<=a)/)
     assert_equal nil, z.scan(/^/)
   end
+
+  it "stack doesn't underflow" do
+    @z.push
+    @z.pop
+    @z.pop
+    @z.pos = 3
+    @z.push
+    @z.pos = 4
+    @z.pop
+    assert_equal 3, @z.pos
+  end
+
+  it "#try restores pos" do
+    z = ZScan.new "hello"
+    return1 = z.try do
+      z.scan 'h'
+      z.scan 'e'
+    end
+    assert_equal 'e', return1
+    assert_equal 2, z.pos
+
+    return2 = z.try do
+      z.scan 'l'
+      z.scan 'l'
+      z.scan 'p' # fails
+    end
+    assert_equal nil, return2
+    assert_equal 2, z.pos
+  end
 end

data/zscan.gemspec

@@ -1,6 +1,6 @@
 Gem::Specification.new do |s|
   s.name = "zscan"
-  s.version = "1.0" # version mapped from zscan.rb, don't change here
+  s.version = "1.0.1" # version mapped from zscan.rb, don't change here
   s.author = "Zete Lui"
   s.homepage = "https://github.com/luikore/zscan"
   s.platform = Gem::Platform::RUBY

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: zscan
 version: !ruby/object:Gem::Version
-  version: '1.0'
+  version: 1.0.1
 platform: ruby
 authors:
 - Zete Lui
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-05-06 00:00:00.000000000 Z
+date: 2013-05-07 00:00:00.000000000 Z
 dependencies: []
 description: improved string scanner, respects anchors and lookbehinds, supports codepoint
   positioning
@@ -21,6 +21,7 @@ files:
 - rakefile
 - zscan.gemspec
 - readme.md
+- bench.rb
 - ext/extconf.rb
 - lib/zscan.rb
 - spec/zscan_spec.rb