@@ -113,6 +113,120 @@ static void end_element(void * ctx, const xmlChar *name) ); } +/** + * start_element_ns was borrowed heavily from libxml-ruby. + */ +static void +start_element_ns ( + void * ctx, + const xmlChar * localname, + const xmlChar * prefix, + const xmlChar * URI, + int nb_namespaces, + const xmlChar ** namespaces, + int nb_attributes, + int nb_defaulted, + const xmlChar ** attributes) +{ + VALUE self = (VALUE)ctx; + VALUE doc = rb_funcall(self, rb_intern("document"), 0); + VALUE MAYBE_UNUSED(enc) = rb_iv_get(self, "@encoding"); + + VALUE attrHash = rb_hash_new(); + VALUE nsHash = rb_hash_new(); + + if (attributes) + { + /* Each attribute is an array of [localname, prefix, URI, value, end] */ + int i; + for (i = 0; i < nb_attributes * 5; i += 5) + { + rb_hash_aset( attrHash, + NOKOGIRI_STR_NEW2((const char*)attributes[i+0], RTEST(enc) ? StringValuePtr(enc) : NULL), + NOKOGIRI_STR_NEW((const char*)attributes[i+3], (attributes[i+4] - attributes[i+3]), RTEST(enc) ? StringValuePtr(enc) : NULL)); + } + } + + if (namespaces) + { + int i; + for (i = 0; i < nb_namespaces * 2; i += 2) + { + rb_hash_aset( nsHash, + namespaces[i+0] ? NOKOGIRI_STR_NEW2((const char*)namespaces[i+0], RTEST(enc) ? StringValuePtr(enc) : NULL) : Qnil, + namespaces[i+1] ? NOKOGIRI_STR_NEW2((const char*)namespaces[i+1], RTEST(enc) ? StringValuePtr(enc) : NULL) : Qnil); + } + } + + rb_funcall( doc, + rb_intern("start_element_ns"), + 5, + NOKOGIRI_STR_NEW2(localname, RTEST(enc) ? StringValuePtr(enc) : NULL), + attrHash, + prefix ? NOKOGIRI_STR_NEW2(prefix, RTEST(enc) ? StringValuePtr(enc) : NULL) : Qnil, + URI ? NOKOGIRI_STR_NEW2(URI, RTEST(enc) ? StringValuePtr(enc) : NULL) : Qnil, + nsHash + ); + + /* Call start element if it's there' */ + if (rb_respond_to(doc, rb_intern("start_element"))) + { + VALUE name; + if (prefix) + { + name = NOKOGIRI_STR_NEW2(prefix, RTEST(enc) ? StringValuePtr(enc) : NULL); + rb_funcall(name, rb_intern("<<"), 1, NOKOGIRI_STR_NEW2(":", RTEST(enc) ? StringValuePtr(enc) : NULL)); + rb_funcall(name, rb_intern("<<"), 1, NOKOGIRI_STR_NEW2(localname, RTEST(enc) ? StringValuePtr(enc) : NULL)); + } + else + { + name = NOKOGIRI_STR_NEW2(localname, RTEST(enc) ? StringValuePtr(enc) : NULL); + } + VALUE attrArray = rb_funcall(attrHash, rb_intern("to_a"), 0); + attrArray = rb_funcall(attrArray, rb_intern("flatten"), 0); + rb_funcall(doc, rb_intern("start_element"), 2, name, attrArray); + } + +} + +/** + * end_element_ns was borrowed heavily from libxml-ruby. + */ +static void +end_element_ns ( + void * ctx, + const xmlChar * localname, + const xmlChar * prefix, + const xmlChar * URI) +{ + VALUE self = (VALUE)ctx; + VALUE doc = rb_funcall(self, rb_intern("document"), 0); + + rb_funcall(doc, rb_intern("end_element_ns"), 3, + NOKOGIRI_STR_NEW2(localname, RTEST(enc) ? StringValuePtr(enc) : NULL), + prefix ? NOKOGIRI_STR_NEW2(prefix, RTEST(enc) ? StringValuePtr(enc) : NULL) : Qnil, + URI ? NOKOGIRI_STR_NEW2(URI, RTEST(enc) ? StringValuePtr(enc) : NULL) : Qnil + ); + + /* Call end element for old-times sake */ + if (rb_respond_to(doc, rb_intern("end_element"))) + { + VALUE name; + if (prefix) + { + name = NOKOGIRI_STR_NEW2(prefix, RTEST(enc) ? StringValuePtr(enc) : NULL); + rb_funcall(name, rb_intern("<<"), 1, NOKOGIRI_STR_NEW2(":", RTEST(enc) ? StringValuePtr(enc) : NULL)); + rb_funcall(name, rb_intern("<<"), 1, NOKOGIRI_STR_NEW2(localname, RTEST(enc) ? StringValuePtr(enc) : NULL)); + } + else + { + name = NOKOGIRI_STR_NEW2(localname, RTEST(enc) ? StringValuePtr(enc) : NULL); + } + rb_funcall(doc, rb_intern("end_element"), 1, name); + } + +} + static void characters_func(void * ctx, const xmlChar * ch, int len) { VALUE self = (VALUE)ctx; @@ -192,11 +306,14 @@ static VALUE allocate(VALUE klass) handler->endDocument = end_document; handler->startElement = start_element; handler->endElement = end_element; + handler->startElementNs = start_element_ns; + handler->endElementNs = end_element_ns; handler->characters = characters_func; handler->comment = comment_func; handler->warning = warning_func; handler->error = error_func; handler->cdataBlock = cdata_block; + handler->initialized = XML_SAX2_MAGIC; return Data_Wrap_Struct(klass, NULL, deallocate, handler); } ext/nokogiri/xml_sax_parser.c @@ -65,6 +65,7 @@ static VALUE initialize_native(VALUE self, VALUE _xml_sax, VALUE _filename) if(ctx == NULL) rb_raise(rb_eRuntimeError, "Could not create a parser context"); + ctx->sax2 = 1; DATA_PTR(self) = ctx; return self; } ext/nokogiri/xml_sax_push_parser.c @@ -91,6 +91,24 @@ module Nokogiri end ### + # Called at the beginning of an element + # +name+ is the element name + # +attrs+ is a hash of attributes + # +prefix+ is the namespace prefix for the element + # +uri+ is the associated namespace URI + # +namespaces+ is a hash of namespace prefix:urls associated with the element + def start_element_ns(name, attrs = {}, prefix = nil, uri = nil, namespaces = {}) + end + + ### + # Called at the end of an element + # +name+ is the element's name + # +prefix+ is the namespace prefix associated with the element + # +uri+ is the associated namespace URI + def end_element_ns(name, prefix = nil, uri = nil) + end + + ### # Characters read between a tag # +string+ contains the character data def characters string lib/nokogiri/xml/sax/document.rb @@ -54,8 +54,8 @@ module Nokogiri module SAX class TestCase < Nokogiri::TestCase class Doc < XML::SAX::Document - attr_reader :start_elements, :start_document_called - attr_reader :end_elements, :end_document_called + attr_reader :start_elements, :start_elements_ns, :start_document_called + attr_reader :end_elements, :end_elements_ns, :end_document_called attr_reader :data, :comments, :cdata_blocks attr_reader :errors, :warnings @@ -84,11 +84,21 @@ module Nokogiri super end + def start_element_ns *args + (@start_elements_ns ||= []) << args + super + end + def end_element *args (@end_elements ||= []) << args super end + def end_element_ns *args + (@end_elements_ns ||= []) << args + super + end + def characters string @data ||= [] @data += [string] test/helper.rb @@ -38,6 +38,30 @@ module Nokogiri @parser.finish end + def test_start_element_ns + @parser.<<(<<-eoxml) + <stream:stream xmlns='jabber:client' xmlns:stream='http://etherx.jabber.org/streams' version='1.0'></stream:stream> + eoxml + + assert_equal [[ 'stream', + {'version' => '1.0'}, + 'stream', + 'http://etherx.jabber.org/streams', + {nil => 'jabber:client', 'stream' => 'http://etherx.jabber.org/streams'}]], + @parser.document.start_elements_ns + @parser.finish + end + + def test_end_element_ns + @parser.<<(<<-eoxml) + <stream:stream xmlns='jabber:client' xmlns:stream='http://etherx.jabber.org/streams' version='1.0'></stream:stream> + eoxml + + assert_equal [['stream', 'stream', 'http://etherx.jabber.org/streams']], + @parser.document.end_elements_ns + @parser.finish + end + def test_chevron_partial_xml @parser.<<(<<-eoxml) <p id="asdfasdf"> test/xml/sax/test_push_parser.rb e5a79545f4ee4d40ac073a30e266c9eace30b313 Jeff Smick sprsquish@gmail.com http://github.com/sprsquish/nokogiri/commit/660362805ff06cd36465d08bf6dda67af6867c4f 660362805ff06cd36465d08bf6dda67af6867c4f 2009-05-19T09:58:38-07:00 2009-05-18T17:16:45-07:00 add support for SAX2 start_element_ns and end_element_ns (code was heavily borrowed from libxml-ruby) f1cd130d5d19d18cf6ae0061ab34a0844f21d905 Jeff Smick sprsquish@gmail.com