From 1e5502b4a97ed9d73d19889a2829f2900dc4a8ce Mon Sep 17 00:00:00 2001
From: Adam Radziszewski <adam.radziszewski@pwr.wroc.pl>
Date: Tue, 16 Apr 2013 13:35:13 +0200
Subject: [PATCH] fx premorph writer: escape XML entities
---
CMakeLists.txt | 2 +-
libcorpus2/io/premorphwriter.cpp | 8 ++++++--
2 files changed, 7 insertions(+), 3 deletions(-)
diff --git a/CMakeLists.txt b/CMakeLists.txt
index 75783dd..01a8890 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -2,7 +2,7 @@ PROJECT(Corpus2Library)
set(corpus2_ver_major "1")
set(corpus2_ver_minor "3")
-set(corpus2_ver_patch "1")
+set(corpus2_ver_patch "2")
cmake_minimum_required(VERSION 2.8.0)
diff --git a/libcorpus2/io/premorphwriter.cpp b/libcorpus2/io/premorphwriter.cpp
index c47e84d..e330794 100644
--- a/libcorpus2/io/premorphwriter.cpp
+++ b/libcorpus2/io/premorphwriter.cpp
@@ -17,6 +17,9 @@ or FITNESS FOR A PARTICULAR PURPOSE.
#include <libcorpus2/io/premorphwriter.h>
#include <boost/foreach.hpp>
+// for entity encoding
+#include <libcorpus2/io/xcescommon.h>
+
namespace Corpus2 {
bool PremorphWriter::registered = TokenWriter::register_writer<PremorphWriter>(
@@ -42,14 +45,15 @@ PremorphWriter::~PremorphWriter()
void PremorphWriter::write_token(const Token &t)
{
- os() << PwrNlp::Whitespace::to_whitespace(t.wa()) << t.orth_utf8();
+ os() << PwrNlp::Whitespace::to_whitespace(t.wa());
+ encode_xml_entities_into(os(), t.orth_utf8());
}
void PremorphWriter::write_sentence(const Sentence &s)
{
os() << "<chunk type=\"s\">";
if (!s.tokens().empty()) {
- os() << s[0]->orth_utf8();
+ encode_xml_entities_into(os(), s[0]->orth_utf8());
}
for (size_t i = 1; i < s.tokens().size(); ++i) {
write_token(*s[i]);
--
GitLab