[ruby-dev:34052] IO should provide #each_char and #chars

Yuguiです。

IOにはバイトストリームだけではなく文字ストリームとしての機能 
も期待されていると思うのですが、
IO#each_charやIO#charsがありません。

バッファに届いていないバイトのせいで不完全な文字が出現する可 
能性を考えたのかとも思いましたが、getcがあるなら
each_charもあるべきではないでしょうか。

とりあえず、getcをベースに書いてみました。

-- 
Yugui
[e-mail]
http://idm.s9.xrea.com




Index: io.c
===================================================================
--- io.c	(revision 15781)
+++ io.c	(working copy)
@@ -2226,8 +2226,79 @@
      return io;
  }

+static VALUE
+io_getc(rb_io_t *fptr, rb_encoding *enc)
+{
+    int r, n;
+    VALUE str;
+
+    if (io_fillbuf(fptr) < 0) {
+	return Qnil;
+    }
+    r = rb_enc_precise_mbclen(fptr->rbuf+fptr->rbuf_off, fptr->rbuf 
+fptr->rbuf_off+fptr->rbuf_len, enc);
+    if (MBCLEN_CHARFOUND_P(r) &&
+        (n = MBCLEN_CHARFOUND_LEN(r)) <= fptr->rbuf_len) {
+	str = rb_str_new(fptr->rbuf+fptr->rbuf_off, n);
+	fptr->rbuf_off += n;
+	fptr->rbuf_len -= n;
+    }
+    else if (MBCLEN_NEEDMORE_P(r)) {
+	str = rb_str_new(fptr->rbuf+fptr->rbuf_off, fptr->rbuf_len);
+        fptr->rbuf_len = 0;
+      getc_needmore:
+        if (io_fillbuf(fptr) != -1) {
+            rb_str_cat(str, fptr->rbuf+fptr->rbuf_off, 1);
+            fptr->rbuf_off++;
+            fptr->rbuf_len--;
+            r = rb_enc_precise_mbclen(RSTRING_PTR(str),  
RSTRING_PTR(str)+RSTRING_LEN(str), enc);
+            if (MBCLEN_NEEDMORE_P(r)) {
+                goto getc_needmore;
+            }
+        }
+    }
+    else {
+	str = rb_str_new(fptr->rbuf+fptr->rbuf_off, 1);
+	fptr->rbuf_off++;
+	fptr->rbuf_len--;
+    }
+    return io_enc_str(str, fptr);
+}
+
  /*
   *  call-seq:
+ *     ios.each_char {|c| block }  => ios
+ *
+ *  Calls the given block once for each character in <em>ios</em>,
+ *  passing the character as an argument. The stream must be opened for
+ *  reading or an <code>IOError</code> will be raised.
+ *
+ *     f = File.new("testfile")
+ *     f.each_char {|c| print c, ' ' }   #=> #<File:testfile>
+ */
+
+static VALUE
+rb_io_each_char(VALUE io)
+{
+    rb_io_t *fptr;
+    rb_encoding *enc;
+    VALUE c;
+
+    RETURN_ENUMERATOR(io, 0, 0);
+    GetOpenFile(io, fptr);
+    rb_io_check_readable(fptr);
+
+    enc = io_input_encoding(fptr);
+    READ_CHECK(fptr);
+    while (!NIL_P(c = io_getc(fptr, enc))) {
+        rb_yield(c);
+    }
+    return io;
+}
+
+
+
+/*
+ *  call-seq:
   *     str.lines(sep=$/)     => anEnumerator
   *     str.lines(limit)      => anEnumerator
   *     str.lines(sep, limit) => anEnumerator
@@ -2239,9 +2310,9 @@
   */

  static VALUE
-rb_io_lines(int argc, VALUE *argv, VALUE str)
+rb_io_lines(int argc, VALUE *argv, VALUE io)
  {
-    return rb_enumeratorize(str, ID2SYM(rb_intern("each_line")),  
argc, argv);
+    return rb_enumeratorize(io, ID2SYM(rb_intern("each_line")), argc,  
argv);
  }

  /*
@@ -2254,13 +2325,31 @@
   */

  static VALUE
-rb_io_bytes(VALUE str)
+rb_io_bytes(VALUE io)
  {
-    return rb_enumeratorize(str, ID2SYM(rb_intern("each_byte")), 0, 0);
+    return rb_enumeratorize(io, ID2SYM(rb_intern("each_byte")), 0, 0);
  }

  /*
   *  call-seq:
+ *     ios.chars   => anEnumerator
+ *
+ *  Returns an enumerator that gives each character in <em>ios</em>.
+ *  The stream must be opened for reading or an <code>IOError</code>
+ *  will be raised.
+ *
+ *     f = File.new("testfile)
+ *     f.chars.each {|c| print c, ' ' }
+ */
+
+static VALUE
+rb_io_chars(VALUE io)
+{
+    return rb_enumeratorize(io, ID2SYM(rb_intern("each_char")), 0, 0);
+}
+
+/*
+ *  call-seq:
   *     ios.getc   => fixnum or nil
   *
   *  Reads a one-character string from <em>ios</em>. Returns
@@ -2284,38 +2373,8 @@

      enc = io_input_encoding(fptr);
      READ_CHECK(fptr);
-    if (io_fillbuf(fptr) < 0) {
-	return Qnil;
-    }
-    r = rb_enc_precise_mbclen(fptr->rbuf+fptr->rbuf_off, fptr->rbuf 
+fptr->rbuf_off+fptr->rbuf_len, enc);
-    if (MBCLEN_CHARFOUND_P(r) &&
-        (n = MBCLEN_CHARFOUND_LEN(r)) <= fptr->rbuf_len) {
-	str = rb_str_new(fptr->rbuf+fptr->rbuf_off, n);
-	fptr->rbuf_off += n;
-	fptr->rbuf_len -= n;
-    }
-    else if (MBCLEN_NEEDMORE_P(r)) {
-	str = rb_str_new(fptr->rbuf+fptr->rbuf_off, fptr->rbuf_len);
-        fptr->rbuf_len = 0;
-      getc_needmore:
-        if (io_fillbuf(fptr) != -1) {
-            rb_str_cat(str, fptr->rbuf+fptr->rbuf_off, 1);
-            fptr->rbuf_off++;
-            fptr->rbuf_len--;
-            r = rb_enc_precise_mbclen(RSTRING_PTR(str),  
RSTRING_PTR(str)+RSTRING_LEN(str), enc);
-            if (MBCLEN_NEEDMORE_P(r)) {
-                goto getc_needmore;
-            }
-        }
-    }
-    else {
-	str = rb_str_new(fptr->rbuf+fptr->rbuf_off, 1);
-	fptr->rbuf_off++;
-	fptr->rbuf_len--;
-    }
-    return io_enc_str(str, fptr);
+    return io_getc(fptr, enc);
  }
-
  int
  rb_getc(FILE *f)
  {
@@ -6477,6 +6536,17 @@
  }

  static VALUE
+argf_each_char(VALUE argf)
+{
+    RETURN_ENUMERATOR(argf, 0, 0);
+    for (;;) {
+	if (!next_argv()) return Qnil;
+	rb_block_call(current_file, rb_intern("each_char"), 0, 0, rb_yield,  
0);
+	next_p = 1;
+    }
+}
+
+static VALUE
  argf_filename(VALUE argf)
  {
      next_argv();
@@ -6773,8 +6843,10 @@
      rb_define_method(rb_cIO, "each",  rb_io_each_line, -1);
      rb_define_method(rb_cIO, "each_line",  rb_io_each_line, -1);
      rb_define_method(rb_cIO, "each_byte",  rb_io_each_byte, 0);
+    rb_define_method(rb_cIO, "each_char",  rb_io_each_char, 0);
      rb_define_method(rb_cIO, "lines",  rb_io_lines, -1);
      rb_define_method(rb_cIO, "bytes",  rb_io_bytes, 0);
+    rb_define_method(rb_cIO, "chars",  rb_io_chars, 0);

      rb_define_method(rb_cIO, "syswrite", rb_io_syswrite, 1);
      rb_define_method(rb_cIO, "sysread",  rb_io_sysread, -1);
@@ -6870,6 +6942,7 @@
      rb_define_method(rb_cARGF, "each",  argf_each_line, -1);
      rb_define_method(rb_cARGF, "each_line",  argf_each_line, -1);
      rb_define_method(rb_cARGF, "each_byte",  argf_each_byte, 0);
+    rb_define_method(rb_cARGF, "each_char",  argf_each_char, 0);

      rb_define_method(rb_cARGF, "read",  argf_read, -1);
      rb_define_method(rb_cARGF, "readpartial",  argf_readpartial, -1);