From 7d489d583980c7c4b6afb4f0bd91853e896f0fd3 Mon Sep 17 00:00:00 2001
From: John Benediktsson <mrjbq7@gmail.com>
Date: Thu, 6 Feb 2014 21:12:57 -0800
Subject: [PATCH] xml.tokenize: fix cdata parsing problem.
---
basis/xml/tests/cdata.factor | 22 ++++++++++++++++++++++
basis/xml/tokenize/tokenize.factor | 14 +++++++++++---
2 files changed, 33 insertions(+), 3 deletions(-)
create mode 100644 basis/xml/tests/cdata.factor
diff --git a/basis/xml/tests/cdata.factor b/basis/xml/tests/cdata.factor
new file mode 100644
index 0000000000..ab79679602
--- /dev/null
+++ b/basis/xml/tests/cdata.factor
@@ -0,0 +1,22 @@
+USING: xml xml.writer tools.test ;
+IN: xml.tests
+
+{
+"""<?xml version="1.0" encoding="UTF-8"?>
+<rss version="2.0">
+ <channel>
+ <item>
+ <description>Python has a n class property in [&#8230;]</description>
+ </item>
+ </channel>
+</rss>"""
+} [
+"""<?xml version="1.0" encoding="UTF-8"?>
+<rss version="2.0">
+ <channel>
+ <item>
+ <description><![CDATA[Python has a n class property in […]]]></description>
+ </item>
+ </channel>
+</rss>""" string>xml xml>string
+] unit-test
diff --git a/basis/xml/tokenize/tokenize.factor b/basis/xml/tokenize/tokenize.factor
index a85c48a022..660cbeaf36 100644
--- a/basis/xml/tokenize/tokenize.factor
+++ b/basis/xml/tokenize/tokenize.factor
@@ -92,10 +92,18 @@ HINTS: next* { spot } ;
: string-matcher ( str -- quot: ( pos char -- pos ? ) )
dup length 1 - '[ _ next-matching dup _ > ] ; inline
+:: (take-string) ( match spot -- sbuf matched? )
+ 10 <sbuf> f [
+ spot char>> [
+ nip over push
+ spot next*
+ dup match tail? dup not
+ ] [ f ] if*
+ ] loop ; inline
+
: take-string ( match -- string )
- [ 0 swap string-matcher take-until nip ] keep
- dupd [ length ] bi@ 1 - - head
- get-char [ missing-close ] unless next ;
+ [ spot get (take-string) [ missing-close ] unless ]
+ [ dupd [ length ] bi@ - over shorten "" like ] bi ;
: expect ( string -- )
dup length spot get '[ _ [ char>> ] keep next* ] "" replicate-as