ox 2.1.6 → 2.1.7

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 46587c384333cacf569a8e839535a5558571bcf4
-  data.tar.gz: c5ef06082e351544aad217a18e485dac9e48e598
+  metadata.gz: 065c6a76ec5d44344dd607e266b9208f32aead7d
+  data.tar.gz: 242089d4a9ebe913eb825721085d1c6c8c7187f8
 SHA512:
-  metadata.gz: 86781ade4fe9945b6d05d1c3ca5462e94f7f9947ddba0d79d29765091aeddc0f839630258a624afa899b5486235054cd6dbe6945c70af080cdedf1c7aaa06523
-  data.tar.gz: a16bfd3f9808401f4b76bd0740ec9c5942c65698e0b3c668a36e1095209fdabbbae479d665096789948322c28e533929ce75ec1e9fe09eafb489e014343f3cd0
+  metadata.gz: 37a3c5ff01e5e238ae83bde8cb3f722624e17b927a8005586ff5b6bba425c1980ba41a120ea8b0889c3da37fff802f42ed5980beb83290ed3c81ace4184b9050
+  data.tar.gz: 5b830c8b6c8e8cd8f41349c2962a7fd510c6540111e41dcb46619983e1829b47e9ae4e8323c6836e03b04530551c6dd0f5374a4d08e9bc89a6d7c7238b1f8469

data/README.md

@@ -34,7 +34,14 @@ A fast XML parser and Object marshaller as a Ruby gem.
 
 ## Release Notes
 
-### Current Release 2.1.6
+### Current Release 2.1.7
+
+ - Empty elements such as <foo></foo> are now called back with empty text.
+
+ - Fixed GC problem that occurs with the new GC in Ruby 2.2 that garbage
+   collects Symbols.
+
+### Release 2.1.6
 
  - Update licenses. No other changes.
 
@@ -44,11 +51,6 @@ A fast XML parser and Object marshaller as a Ruby gem.
    rb_intern(). There does not seem to be a way to force symbols created with
    encoding to be pinned.
 
-### Release 2.1.4
-
- - Fixed bug where the parser always started at the first position in a stringio
-   instead of the current position.
-
 ## Description
 
 Optimized XML (Ox), as the name implies was written to provide speed optimized
@@ -89,7 +91,7 @@ callbacks. Unlike Nokogiri and LibXML, Ox can be tuned to use only the SAX
 callbacks that are of interest to the caller. (See the perf_sax.rb file for an
 example.)
 
-Ox is compatible with Ruby 1.8.7, 1.9.3, 2.1.2, and RBX.
+Ox is compatible with Ruby 1.8.7, 1.9.3, 2.1.2, 2.2.0 and RBX.
 
 ### Object Dump Sample:
 

data/ext/ox/gen_load.c

@@ -300,13 +300,14 @@ add_element(PInfo pi, const char *ename, Attr attrs, int hasChildren) {
     e = rb_obj_alloc(ox_element_clas);
     rb_ivar_set(e, ox_at_value_id, s);
     if (0 != attrs->name) {
-        VALUE   ah = rb_hash_new();
+        volatile VALUE	ah = rb_hash_new();
         
         for (; 0 != attrs->name; attrs++) {
-            VALUE   sym;
-            VALUE   *slot;
+            volatile VALUE	sym;
 
 	    if (Yes == pi->options->sym_keys) {
+		VALUE	*slot;
+
 		if (Qundef == (sym = ox_cache_get(ox_symbol_cache, attrs->name, &slot, 0))) {
 #if HAS_ENCODING_SUPPORT
 		    if (0 != pi->options->rb_enc) {
@@ -329,6 +330,9 @@ add_element(PInfo pi, const char *ename, Attr attrs, int hasChildren) {
 #else
 		    sym = ID2SYM(rb_intern(attrs->name));
 #endif
+		    // Needed for Ruby 2.2 to get around the GC of symbols
+		    // created with to_sym which is needed for encoded symbols.
+		    rb_ary_push(ox_sym_bank, sym);
 		    *slot = sym;
 		}
 	    } else {
@@ -437,6 +441,9 @@ add_instruct(PInfo pi, const char *name, Attr attrs, const char *content) {
 #else
 		    sym = ID2SYM(rb_intern(attrs->name));
 #endif
+		    // Needed for Ruby 2.2 to get around the GC of symbols
+		    // created with to_sym which is needed for encoded symbols.
+		    rb_ary_push(ox_sym_bank, sym);
 		    *slot = sym;
 		}
 	    } else {

data/ext/ox/obj_load.c

@@ -82,11 +82,14 @@ name2var(const char *name, void *encoding) {
     } else if (Qundef == (var_id = ox_cache_get(ox_attr_cache, name, &slot, 0))) {
 #ifdef HAVE_RUBY_ENCODING_H
 	if (0 != encoding) {
-	    VALUE	rstr = rb_str_new2(name);
-	    VALUE	sym;
+	    volatile VALUE	rstr = rb_str_new2(name);
+	    volatile VALUE	sym;
 	    
 	    rb_enc_associate(rstr, (rb_encoding*)encoding);
 	    sym = rb_funcall(rstr, ox_to_sym_id, 0);
+	    // Needed for Ruby 2.2 to get around the GC of symbols
+	    // created with to_sym which is needed for encoded symbols.
+	    rb_ary_push(ox_sym_bank, sym);
 	    var_id = SYM2ID(sym);
 	} else {
 	    var_id = rb_intern(name);
@@ -462,6 +465,9 @@ add_text(PInfo pi, char *text, int closed) {
 
 	if (Qundef == (sym = ox_cache_get(ox_symbol_cache, text, &slot, 0))) {
 	    sym = str2sym(text, (void*)pi->options->rb_enc);
+	    // Needed for Ruby 2.2 to get around the GC of symbols created with
+	    // to_sym which is needed for encoded symbols.
+	    rb_ary_push(ox_sym_bank, sym);
 	    *slot = sym;
 	}
 	h->obj = sym;
@@ -513,6 +519,9 @@ add_text(PInfo pi, char *text, int closed) {
 	from_base64(text, (uchar*)str);
 	if (Qundef == (sym = ox_cache_get(ox_symbol_cache, str, &slot, 0))) {
 	    sym = str2sym(str, (void*)pi->options->rb_enc);
+	    // Needed for Ruby 2.2 to get around the GC of symbols created with
+	    // to_sym which is needed for encoded symbols.
+	    rb_ary_push(ox_sym_bank, sym);
 	    *slot = sym;
 	}
 	h->obj = sym;

data/ext/ox/ox.c

@@ -76,6 +76,7 @@ VALUE	ox_encoding_sym;
 
 VALUE	ox_empty_string;
 VALUE	ox_zero_fixnum;
+VALUE	ox_sym_bank; // Array
 
 VALUE	ox_arg_error_class;
 VALUE	ox_bag_clas;
@@ -995,6 +996,7 @@ void Init_ox() {
 
     ox_empty_string = rb_str_new2("");				rb_gc_register_address(&ox_empty_string);
     ox_zero_fixnum = INT2NUM(0);				rb_gc_register_address(&ox_zero_fixnum);
+    ox_sym_bank = rb_ary_new();					rb_gc_register_address(&ox_sym_bank);
 
     ox_document_clas = rb_const_get_at(Ox, rb_intern("Document"));
     ox_element_clas = rb_const_get_at(Ox, rb_intern("Element"));

data/ext/ox/ox.h

@@ -229,6 +229,7 @@ extern VALUE	ox_stringio_class;
 extern VALUE	ox_struct_class;
 extern VALUE	ox_time_class;
 extern VALUE	ox_zero_fixnum;
+extern VALUE	ox_sym_bank; // Array
 
 extern VALUE	ox_document_clas;
 extern VALUE	ox_element_clas;

data/ext/ox/sax.c

@@ -103,6 +103,9 @@ str2sym(SaxDrive dr, const char *str, const char **strp) {
 
 		rb_funcall(rstr, ox_force_encoding_id, 1, dr->encoding);
 		sym = rb_funcall(rstr, ox_to_sym_id, 0);
+		// Needed for Ruby 2.2 to get around the GC of symbols created
+		// with to_sym which is needed for encoded symbols.
+		rb_ary_push(ox_sym_bank, sym);
 		*slot = Qundef;
 	    } else {
 		sym = ID2SYM(rb_intern(str));
@@ -262,12 +265,10 @@ static void
 parse(SaxDrive dr) {
     char        c = skipBOM(dr);
     int		state = START_STATE;
+    Nv		parent;
 
     while ('\0' != c) {
 	buf_protect(&dr->buf);
-        if (is_white(c) && '\0' == (c = buf_next_non_white(&dr->buf))) {
-            break;
-        }
 	if ('<' == c) {
 	    c = buf_get(&dr->buf);
 	    switch (c) {
@@ -325,6 +326,11 @@ parse(SaxDrive dr) {
 			ox_sax_drive_error(dr, CASE_ERROR "expected CDATA all in caps");
 			c = read_cdata(dr);
 		    } else {
+			Nv	parent = stack_peek(&dr->stack);
+
+			if (0 != parent) {
+			    parent->childCnt++;
+			}
 			ox_sax_drive_error_at(dr, WRONG_CHAR "DOCTYPE, CDATA, or comment expected", line, col);
 			c = read_name_token(dr);
 			if ('>' == c) {
@@ -334,6 +340,30 @@ parse(SaxDrive dr) {
 		}
 		break;
 	    case '/': /* element end */
+		parent = stack_peek(&dr->stack);
+		if (0 != parent && 0 == parent->childCnt) {
+		    VALUE	args[1];
+		    int		line = dr->buf.line;
+		    int		col = dr->buf.col - 1;
+
+		    args[0] = rb_str_new2("");
+#if HAS_ENCODING_SUPPORT
+		    if (0 != dr->encoding) {
+			rb_enc_associate(args[0], dr->encoding);
+		    }
+#elif HAS_PRIVATE_ENCODING
+		    if (Qnil != dr->encoding) {
+			rb_funcall(args[0], ox_force_encoding_id, 1, dr->encoding);
+		    }
+#endif
+		    if (dr->has.line) {
+			rb_ivar_set(dr->handler, ox_at_line_id, LONG2NUM(line));
+		    }
+		    if (dr->has.column) {
+			rb_ivar_set(dr->handler, ox_at_column_id, LONG2NUM(col));
+		    }
+		    rb_funcall2(dr->handler, ox_text_id, 1, args);
+		}
 		c = read_element_end(dr);
 		if (0 == stack_peek(&dr->stack)) {
 		    state = AFTER_STATE;
@@ -450,7 +480,7 @@ read_instruction(SaxDrive dr) {
     dr->err = 0;
     c = read_attrs(dr, c, '?', '?', is_xml, 1);
     if (dr->has.attrs_done) {
-	    rb_funcall(dr->handler, ox_attrs_done_id, 0);
+	rb_funcall(dr->handler, ox_attrs_done_id, 0);
     }
     if (dr->err) {
 	if (dr->has.text) {
@@ -557,6 +587,7 @@ read_doctype(SaxDrive dr) {
     int		line = dr->buf.line;
     int		col = dr->buf.col - 10;
     char	*s;
+    Nv		parent = stack_peek(&dr->stack);
 
     buf_backup(&dr->buf); /* back up to the start in case the cdata is empty */
     buf_protect(&dr->buf);
@@ -568,6 +599,9 @@ read_doctype(SaxDrive dr) {
 	}
     }
     *(dr->buf.tail - 1) = '\0';
+    if (0 != parent) {
+	parent->childCnt++;
+    }
     if (dr->has.doctype) {
         VALUE       args[1];
 
@@ -590,11 +624,16 @@ read_doctype(SaxDrive dr) {
 static char
 read_cdata(SaxDrive dr) {
     char        	c;
+    char        	zero = '\0';
     int         	end = 0;
     int			line = dr->buf.line;
     int			col = dr->buf.col - 10;
     struct _CheckPt	cp = CHECK_PT_INIT;
+    Nv			parent = stack_peek(&dr->stack);
 
+    if (0 != parent) {
+	parent->childCnt++;
+    }
     buf_backup(&dr->buf); /* back up to the start in case the cdata is empty */
     buf_protect(&dr->buf);
     while (1) {
@@ -624,6 +663,7 @@ read_cdata(SaxDrive dr) {
 	    if (buf_checkset(&cp)) {
 		c = buf_checkback(&dr->buf, &cp);
 		ox_sax_drive_error(dr, NO_TERM "CDATA not terminated");
+		zero = c;
 		*(dr->buf.tail - 1) = '\0';
 		goto CB;
 	    }
@@ -659,6 +699,9 @@ read_cdata(SaxDrive dr) {
 	}
         rb_funcall2(dr->handler, ox_cdata_id, 1, args);
     }
+    if ('\0' != zero) {
+	*(dr->buf.tail - 1) = zero;
+    }
     dr->buf.str = 0;
 
     return c;
@@ -669,6 +712,7 @@ read_cdata(SaxDrive dr) {
 static char
 read_comment(SaxDrive dr) {
     char        	c;
+    char        	zero = '\0';
     int         	end = 0;
     int			line = dr->buf.line;
     int			col = dr->buf.col - 4;
@@ -703,6 +747,7 @@ read_comment(SaxDrive dr) {
 	    if (buf_checkset(&cp)) {
 		c = buf_checkback(&dr->buf, &cp);
 		ox_sax_drive_error(dr, NO_TERM "comment not terminated");
+		zero = c;
 		*(dr->buf.tail - 1) = '\0';
 		goto CB;
 	    }
@@ -738,6 +783,9 @@ read_comment(SaxDrive dr) {
 	}
         rb_funcall2(dr->handler, ox_comment_id, 1, args);
     }
+    if ('\0' != zero) {
+	*(dr->buf.tail - 1) = zero;
+    }
     dr->buf.str = 0;
 
     return c;
@@ -756,10 +804,15 @@ read_element_start(SaxDrive dr) {
     int			col = dr->buf.col - 1;
     Hint		h = 0;
     int			stackless = 0;
+    Nv			parent = stack_peek(&dr->stack);
+
 
     if ('\0' == (c = read_name_token(dr))) {
         return '\0';
     }
+    if (0 != parent) {
+	parent->childCnt++;
+    }
     if (dr->options.smart && 0 == dr->hints && stack_empty(&dr->stack) && 0 == strcasecmp("html", dr->buf.str)) {
 	dr->hints = ox_hints_html();
     }
@@ -833,7 +886,7 @@ read_element_start(SaxDrive dr) {
 	closed = ('/' == c);
     }
     if (dr->has.attrs_done) {
-	    rb_funcall(dr->handler, ox_attrs_done_id, 0);
+	rb_funcall(dr->handler, ox_attrs_done_id, 0);
     }
     if (closed) {
 	c = buf_next_non_white(&dr->buf);
@@ -950,18 +1003,46 @@ read_text(SaxDrive dr) {
     char        c;
     int		line = dr->buf.line;
     int		col = dr->buf.col - 1;
+    Nv		parent = stack_peek(&dr->stack);
+    int		allWhite = 1;
 
     buf_backup(&dr->buf);
     buf_protect(&dr->buf);
     while ('<' != (c = buf_get(&dr->buf))) {
-        if ('\0' == c) {
+	switch(c) {
+	case ' ':
+	case '\t':
+	case '\f':
+	case '\n':
+	case '\r':
+	    break;
+	case '\0':
+	    if (allWhite) {
+		return c;
+	    }
             ox_sax_drive_error(dr, NO_TERM "text not terminated");
+	    goto END_OF_BUF;
 	    break;
-        }
+	default:
+	    allWhite = 0;
+	    break;
+	}
     }
+ END_OF_BUF:
     if ('\0' != c) {
 	*(dr->buf.tail - 1) = '\0';
     }
+    if (allWhite) {
+	int	isEnd = ('/' == buf_get(&dr->buf));
+
+	buf_backup(&dr->buf);
+	if (!isEnd || 0 == parent || 0 < parent->childCnt) {
+	    return c;
+	}
+    }
+    if (0 != parent) {
+	parent->childCnt++;
+    }
     if (dr->has.value) {
 	if (dr->has.line) {
 	    rb_ivar_set(dr->handler, ox_at_line_id, LONG2NUM(line));

data/ext/ox/sax_stack.h

@@ -13,6 +13,7 @@
 typedef struct _Nv {
     const char	*name;
     VALUE	val;
+    int		childCnt;
     Hint	hint;
 } *Nv;
 
@@ -60,6 +61,7 @@ stack_push(NStack stack, const char *name, VALUE val, Hint hint) {
     stack->tail->name = name;
     stack->tail->val = val;
     stack->tail->hint = hint;
+    stack->tail->childCnt = 0;
     stack->tail++;
 }
 

data/lib/ox/version.rb

@@ -1,5 +1,5 @@
 
 module Ox
   # Current version of the module. 
-  VERSION = '2.1.6'
+  VERSION = '2.1.7'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: ox
 version: !ruby/object:Gem::Version
-  version: 2.1.6
+  version: 2.1.7
 platform: ruby
 authors:
 - Peter Ohler
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-12-31 00:00:00.000000000 Z
+date: 2015-01-31 00:00:00.000000000 Z
 dependencies: []
 description: "A fast XML parser and object serializer that uses only standard C lib.\n
   \           \nOptimized XML (Ox), as the name implies was written to provide speed
@@ -92,7 +92,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: ox
-rubygems_version: 2.2.2
+rubygems_version: 2.4.5
 signing_key: 
 specification_version: 4
 summary: A fast XML parser and object serializer.