Skip to content
GitLab
Commit ddbff8ff authored by Braden McDaniel's avatar Braden McDaniel Committed by Braden McDaniel
Browse files

Add support for character data parsing (and reporting)

parent d19d1cd3
Hide whitespace changes
Inline Side-by-side
src/endoframe/xml.h
View file @ ddbff8ff
......@@ -51,6 +51,13 @@ namespace endoframe {
/// \param[in] name Element name.
///
void end_element(std::string name) {}
///
/// Called when character data is encountered.
///
/// \param[in] chars Character data.
///
void characters(std::string chars) {}
};
template <typename ContentHandler>
......@@ -81,12 +88,10 @@ bool endoframe::xml::parse(std::istream & in, ContentHandler & content_handler)
try {
content_handler.start_document();
result =
phrase_parse(
pos, end,
x3::with<grammar::content_handler_tag>(
std::ref(content_handler))
[grammar::document],
x3::space);
parse(pos, end,
x3::with<grammar::content_handler_tag>(
std::ref(content_handler))
[grammar::document]);
content_handler.end_document();
} catch (const x3::expectation_failure<decltype(pos)> & ex)
{
......
src/endoframe/xml/grammar.h
View file @ ddbff8ff
......@@ -41,6 +41,19 @@ struct on_element_end {
struct e_tag_class : on_element_end {};
struct on_char_data {
template <typename T, typename Iterator, typename Context>
void on_success(const Iterator & /* first */, const Iterator & /* last */,
T & chars, const Context & context)
{
namespace x3 = boost::spirit::x3;
auto & doc_handler = x3::get<content_handler_tag>(context).get();
doc_handler.characters(std::move(chars));
}
};
struct char_data_class : on_char_data {};
const boost::spirit::x3::rule<class document> document = "document";
const boost::spirit::x3::rule<class prolog> prolog = "prolog";
const boost::spirit::x3::rule<class xml_decl> xml_decl = "XMLDecl";
......@@ -60,22 +73,28 @@ const boost::spirit::x3::rule<class attribute_, xml::attribute> attribute =
const boost::spirit::x3::rule<class name, std::string> name = "Name";
const boost::spirit::x3::rule<class att_value, std::string> att_value =
"AttValue";
const boost::spirit::x3::rule<char_data_class, std::string> char_data = "CharData";
#ifdef __GNUG__
#pragma GCC diagnostic push
#pragma GCC diagnostic ignored "-Wparentheses"
#endif
const auto version_info_def =
boost::spirit::x3::lit("version") > '='
> (boost::spirit::x3::lit(R"("1.0")") | "'1.0'")
const auto s
= boost::spirit::x3::omit[+boost::spirit::x3::space]
;
const auto eq
= -s > '=' > -s
;
const auto version_info_def
= s > "version" > eq > (boost::spirit::x3::lit(R"("1.0")") | "'1.0'")
;
const auto encoding_decl_def
= boost::spirit::x3::lit("encoding") > '='
> ( boost::spirit::x3::lexeme['\'' >> enc_name >> '\'']
| boost::spirit::x3::lexeme['"' >> enc_name >> '"']
)
= s > "encoding" > eq > ( boost::spirit::x3::lexeme['\'' >> enc_name >> '\'']
| boost::spirit::x3::lexeme['"' >> enc_name >> '"'] )
;
const auto enc_name_def
......@@ -84,11 +103,15 @@ const auto enc_name_def
;
const auto xml_decl_def
= boost::spirit::x3::lit("<?xml") > version_info > -encoding_decl > "?>"
= boost::spirit::x3::lit("<?xml") > version_info > -encoding_decl > -s > "?>"
;
const auto misc
= s
;
const auto prolog_def
= -xml_decl
= -xml_decl >> *misc
;
const auto name_start_char
......@@ -168,23 +191,28 @@ const auto att_value_def
;
const auto attribute_def
= name > '=' > att_value
= name > eq > att_value
;
const auto elem_start_def
= boost::spirit::x3::lexeme['<' >> name] >> *attribute
= boost::spirit::x3::lexeme['<' >> name] >> *(s >> attribute)
;
const auto empty_elem_tag_def
= elem_start >> "/>"
= elem_start >> -s >> "/>"
;
const auto s_tag_def
= elem_start >> '>'
= elem_start >> -s >> '>'
;
const auto char_data_def
= *(boost::spirit::x3::char_ - boost::spirit::x3::char_("<&"))
>> -boost::spirit::x3::lit("]]>")
;
const auto content
= *(element | reference)
= -char_data > *((element | reference) >> -char_data)
;
const auto e_tag_def
......@@ -216,6 +244,7 @@ BOOST_SPIRIT_DEFINE(
s_tag,
elem_start,
e_tag,
char_data,
name,
attribute,
att_value
......
tests/parse.cpp
View file @ ddbff8ff
......@@ -5,6 +5,18 @@
#include <boost/variant.hpp>
#include <endoframe/xml.h>
BOOST_AUTO_TEST_CASE(xml_decl)
{
namespace x3 = boost::spirit::x3;
const std::string decl = R"(<?xml version="1.0" encoding="UTF-8"?>)";
auto pos = decl.begin();
auto result = x3::parse(pos, decl.end(),
endoframe::xml::grammar::xml_decl);
BOOST_TEST_REQUIRE(result == true);
BOOST_TEST((pos == decl.end()) == true);
}
const std::string good_names[] = {
"my_name",
"my-name",
......@@ -22,8 +34,7 @@ BOOST_DATA_TEST_CASE(name_good, good_names)
auto pos = sample.begin();
std::string parse_result;
auto result =
phrase_parse(pos, sample.end(),
endoframe::xml::grammar::name, x3::space, parse_result);
parse(pos, sample.end(), endoframe::xml::grammar::name, parse_result);
BOOST_TEST_REQUIRE(result == true);
BOOST_TEST_REQUIRE((pos == sample.end()) == true);
BOOST_TEST(parse_result == sample);
......@@ -42,7 +53,7 @@ BOOST_DATA_TEST_CASE(name_bad, bad_names)
auto pos = sample.begin();
std::string parse_result;
phrase_parse(pos, sample.end(), endoframe::xml::grammar::name, x3::space, parse_result);
parse(pos, sample.end(), endoframe::xml::grammar::name, parse_result);
BOOST_TEST_REQUIRE((pos == sample.end()) == false);
}
......@@ -70,8 +81,8 @@ BOOST_DATA_TEST_CASE(char_ref_good, good_char_refs)
auto pos = sample.test_val.begin();
std::uint32_t parse_result;
auto result =
phrase_parse(pos, sample.test_val.end(),
endoframe::xml::grammar::char_ref, x3::space, parse_result);
parse(pos, sample.test_val.end(),
endoframe::xml::grammar::char_ref, parse_result);
BOOST_TEST_REQUIRE(result == true);
BOOST_TEST_REQUIRE((pos == sample.test_val.end()) == true);
BOOST_TEST(parse_result == sample.expect);
......@@ -101,8 +112,8 @@ BOOST_DATA_TEST_CASE(entity_ref_good, good_entity_refs)
auto pos = sample.test_val.begin();
std::string parse_result;
auto result =
phrase_parse(pos, sample.test_val.end(),
endoframe::xml::grammar::entity_ref, x3::space, parse_result);
parse(pos, sample.test_val.end(),
endoframe::xml::grammar::entity_ref, parse_result);
BOOST_TEST_REQUIRE(result == true);
BOOST_TEST_REQUIRE((pos == sample.test_val.end()) == true);
BOOST_TEST(parse_result == sample.expect);
......@@ -146,8 +157,8 @@ BOOST_DATA_TEST_CASE(reference_good, good_references)
auto pos = sample.test_val.begin();
boost::variant<std::string, std::uint32_t> parse_result;
auto result =
phrase_parse(pos, sample.test_val.end(),
endoframe::xml::grammar::reference, x3::space, parse_result);
parse(pos, sample.test_val.end(),
endoframe::xml::grammar::reference, parse_result);
BOOST_TEST_REQUIRE(result == true);
BOOST_TEST_REQUIRE((pos == sample.test_val.end()) == true);
BOOST_TEST(parse_result == sample.expect);
......@@ -179,9 +190,8 @@ BOOST_DATA_TEST_CASE(att_value_good, good_att_values)
auto pos = sample.test_val.begin();
std::string parse_result;
auto result =
phrase_parse(pos, sample.test_val.end(),
endoframe::xml::grammar::att_value, x3::space,
parse_result);
parse(pos, sample.test_val.end(),
endoframe::xml::grammar::att_value, parse_result);
BOOST_TEST_REQUIRE(result == true);
BOOST_TEST_REQUIRE((pos == sample.test_val.end()) == true);
BOOST_TEST(parse_result == sample.expect);
......@@ -201,7 +211,10 @@ static std::ostream & operator<<(std::ostream & out,
}
const good_attribute_sample good_attributes[] = {
{ "foo=\"abc&#123;abc\"", { "foo", "abc&#123;abc" } }
{ "foo=\"abc&#123;abc\"", { "foo", "abc&#123;abc" } },
{ "foo =\"abc&#123;abc\"", { "foo", "abc&#123;abc" } },
{ "foo= \"abc&#123;abc\"", { "foo", "abc&#123;abc" } },
{ "foo = \"abc&#123;abc\"", { "foo", "abc&#123;abc" } }
};
BOOST_DATA_TEST_CASE(attribute_good, good_attributes)
......@@ -211,9 +224,9 @@ BOOST_DATA_TEST_CASE(attribute_good, good_attributes)
auto pos = sample.test_val.begin();
endoframe::xml::attribute parse_result;
auto result =
phrase_parse(pos, sample.test_val.end(),
endoframe::xml::grammar::attribute, x3::space,
parse_result);
parse(pos, sample.test_val.end(),
endoframe::xml::grammar::attribute,
parse_result);
BOOST_TEST_REQUIRE(result == true);
BOOST_TEST_REQUIRE((pos == sample.test_val.end()) == true);
BOOST_TEST(parse_result.name == sample.expect.name);
......@@ -222,7 +235,6 @@ BOOST_DATA_TEST_CASE(attribute_good, good_attributes)
BOOST_AUTO_TEST_CASE(start_and_end_document)
{
namespace x3 = boost::spirit::x3;
namespace xml = endoframe::xml;
struct content_handler : xml::content_handler {
......@@ -241,8 +253,9 @@ BOOST_AUTO_TEST_CASE(start_and_end_document)
}
};
const std::string doc = R"(
<?xml version="1.0" encoding="UTF-8"?> <element> </element>
const std::string doc =
R"(<?xml version="1.0" encoding="UTF-8"?>
<element> </element>
)";
std::istringstream in{doc};
......@@ -254,7 +267,6 @@ BOOST_AUTO_TEST_CASE(start_and_end_document)
BOOST_AUTO_TEST_CASE(start_and_end_element)
{
namespace x3 = boost::spirit::x3;
namespace xml = endoframe::xml;
struct content_handler : xml::content_handler {
......@@ -281,8 +293,9 @@ BOOST_AUTO_TEST_CASE(start_and_end_element)
}
};
const std::string doc = R"(
<?xml version="1.0" encoding="UTF-8"?> <element name1="val1" name2="val2"> </element>
const std::string doc =
R"(<?xml version="1.0" encoding="UTF-8"?>
<element name1="val1" name2="val2"> </element>
)";
std::istringstream in{doc};
......@@ -291,3 +304,28 @@ BOOST_AUTO_TEST_CASE(start_and_end_element)
BOOST_TEST_REQUIRE(result == true);
BOOST_TEST(doc_handler.end_element_called == true);
}
BOOST_AUTO_TEST_CASE(char_data)
{
namespace xml = endoframe::xml;
struct content_handler : xml::content_handler {
bool characters_called = false;
void characters(std::string chars) {
characters_called = true;
BOOST_TEST(chars == " foo ");
}
};
const std::string doc =
R"(<?xml version="1.0" encoding="UTF-8"?>
<element> foo </element>
)";
std::istringstream in{doc};
content_handler doc_handler;
auto result = xml::parse(in, doc_handler);
BOOST_TEST_REQUIRE(result == true);
BOOST_TEST(doc_handler.characters_called == true);
}
Supports Markdown
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment