Import htmlcxx-0.85 as www/htmlcxx.

htmlcxx is a simple non-validating CSS1 and HTML parser for C++.
Although there are several other HTML parsers available, htmlcxx
has some characteristics that make it unique:

* STL like navigation of DOM tree, using the excellent tree.hh library
  from Kasper Peeters
* It is possible to reproduce exactly, character by character, the
  original document from the parse tree
* Bundled css parser
* Optional parsing of attributes
* C++ code that looks like C++ (not so true anymore)
* Offsets of tags/elements in the original document are stored in
  the nodes of the DOM tree

The parsing politics of htmlcxx were created trying to mimic Mozilla
Firefox behavior. So you should expect parse trees similar to those
create by Firefox. However, differently from Firefox, htmlcxx does
not insert non-existent stuff in your html. Therefore, serializing
the DOM tree gives exactly the same bytes contained in the original
HTML document.

1 files changed, 15 insertions, 0 deletions

diff --git a/www/htmlcxx/patches/patch-html_CharsetConverter.cc b/www/htmlcxx/patches/patch-html_CharsetConverter.cc
new file mode 100644
index 00000000000..72dc5fa8a12
--- /dev/null
+++ b/www/htmlcxx/patches/patch-html_CharsetConverter.cc
@@ -0,0 +1,15 @@
+$NetBSD: patch-html_CharsetConverter.cc,v 1.1 2014/02/16 22:58:51 wiz Exp $
+
+Fix build on NetBSD.
+
+--- html/CharsetConverter.cc.orig	2011-05-15 17:32:10.000000000 +0000
++++ html/CharsetConverter.cc
+@@ -37,7 +37,7 @@ string CharsetConverter::convert(const s
+ 
+ 	size_t ret;
+ 	while (1) {
+-		ret = iconv(mIconvDescriptor, const_cast<char**>(&inbuf), &inbytesleft, &outbuf, &outbytesleft);
++		ret = iconv(mIconvDescriptor, (const char**)(&inbuf), &inbytesleft, &outbuf, &outbytesleft);
+ 		if (ret == 0) break;
+ 		if (ret == (size_t)-1 && errno == E2BIG) return string();
+