From 3c3f615d7c4b77392c565d05712322a85eea3cc5 Mon Sep 17 00:00:00 2001
From: SheetJS <dev@sheetjs.com>
Date: Thu, 9 Oct 2014 22:12:25 -0400
Subject: [PATCH] updating to 0.7.11

---
 index.html     |   4 +-
 ods.js         | 352 +++++++++++++++++++++++++++++++++++++++++++++++++
 xlsx.js        |  38 ++++--
 xlsxworker.js  |   2 +
 xlsxworker1.js |   2 +
 xlsxworker2.js |   2 +
 6 files changed, 389 insertions(+), 11 deletions(-)
 create mode 100644 ods.js

diff --git a/index.html b/index.html
index b8a9ff5..2b67e9e 100644
--- a/index.html
+++ b/index.html
@@ -34,7 +34,7 @@ Output Format:
 <input type="radio" name="format" value="json"> JSON<br>
 <input type="radio" name="format" value="form"> FORMULAE<br> -->
 
-<div id="drop">Drop an XLSX or XLSM or XLSB file here to see sheet data</div>
+<div id="drop">Drop an XLSX / XLSM / XLSB / ODS file here to see sheet data</div>
 <p><input type="file" name="xlfile" id="xlf" /> ... or click here to select a file</p>
 <textarea id="b64data">... or paste a base64-encoding here</textarea>
 <input type="button" id="dotext" value="Click here to process the base64 text" onclick="b64it();"/><br />
@@ -49,6 +49,8 @@ Use readAsBinaryString: (when available) <input type="checkbox" name="userabs" c
 <script src="shim.js"></script>
 <script src="jszip.js"></script>
 <script src="xlsx.js"></script>
+<!-- uncomment the next line here and in xlsxworker.js for ODS support -->
+<script src="ods.js"></script>
 <script>
 var rABS = typeof FileReader !== "undefined" && typeof FileReader.prototype !== "undefined" && typeof FileReader.prototype.readAsBinaryString !== "undefined";
 if(!rABS) {
diff --git a/ods.js b/ods.js
new file mode 100644
index 0000000..d5a23ca
--- /dev/null
+++ b/ods.js
@@ -0,0 +1,352 @@
+/* ods.js (C) 2014 SheetJS -- http://sheetjs.com */
+/* vim: set ts=2: */
+/*jshint -W041 */
+var ODS = {};
+(function make_ods(ODS) {
+/* Open Document Format for Office Applications (OpenDocument) Version 1.2 */
+var get_utils = function() {
+	if(typeof XLSX !== 'undefined') return XLSX.utils;
+	if(typeof module !== "undefined" && typeof require !== 'undefined') return require('xl' + 'sx').utils;
+	throw new Error("Cannot find XLSX utils");
+};
+var has_buf = (typeof Buffer !== 'undefined');
+function getdata(data) {
+	if(!data) return null;
+	if(data.data) return data.data;
+	if(data.asNodeBuffer && has_buf) return data.asNodeBuffer().toString('binary');
+	if(data.asBinary) return data.asBinary();
+	if(data._data && data._data.getContent) return cc2str(Array.prototype.slice.call(data._data.getContent(),0));
+	return null;
+}
+
+function safegetzipfile(zip, file) {
+	var f = file; if(zip.files[f]) return zip.files[f];
+	f = file.toLowerCase(); if(zip.files[f]) return zip.files[f];
+	f = f.replace(/\//g,'\\'); if(zip.files[f]) return zip.files[f];
+	return null;
+}
+
+function getzipfile(zip, file) {
+	var o = safegetzipfile(zip, file);
+	if(o == null) throw new Error("Cannot find file " + file + " in zip");
+	return o;
+}
+
+function getzipdata(zip, file, safe) {
+	if(!safe) return getdata(getzipfile(zip, file));
+	if(!file) return null;
+	try { return getzipdata(zip, file); } catch(e) { return null; }
+}
+
+var _fs, jszip;
+if(typeof JSZip !== 'undefined') jszip = JSZip;
+if (typeof exports !== 'undefined') {
+	if (typeof module !== 'undefined' && module.exports) {
+		if(has_buf && typeof jszip === 'undefined') jszip = require('js'+'zip');
+		if(typeof jszip === 'undefined') jszip = require('./js'+'zip').JSZip;
+		_fs = require('f'+'s');
+	}
+}
+var attregexg=/\b[\w:-]+=["'][^"]*['"]/g;
+var tagregex=/<[^>]*>/g;
+var nsregex=/<\w*:/, nsregex2 = /<(\/?)\w+:/;
+function parsexmltag(tag, skip_root) {
+	var z = [];
+	var eq = 0, c = 0;
+	for(; eq !== tag.length; ++eq) if((c = tag.charCodeAt(eq)) === 32 || c === 10 || c === 13) break;
+	if(!skip_root) z[0] = tag.substr(0, eq);
+	if(eq === tag.length) return z;
+	var m = tag.match(attregexg), j=0, w="", v="", i=0, q="", cc="";
+	if(m) for(i = 0; i != m.length; ++i) {
+		cc = m[i];
+		for(c=0; c != cc.length; ++c) if(cc.charCodeAt(c) === 61) break;
+		q = cc.substr(0,c); v = cc.substring(c+2, cc.length-1);
+		for(j=0;j!=q.length;++j) if(q.charCodeAt(j) === 58) break;
+		if(j===q.length) z[q] = v;
+		else z[(j===5 && q.substr(0,5)==="xmlns"?"xmlns":"")+q.substr(j+1)] = v;
+	}
+	return z;
+}
+function strip_ns(x) { return x.replace(nsregex2, "<$1"); }
+
+var encodings = {
+	'&quot;': '"',
+	'&apos;': "'",
+	'&gt;': '>',
+	'&lt;': '<',
+	'&amp;': '&'
+};
+var rencoding = {
+	'"': '&quot;',
+	"'": '&apos;',
+	'>': '&gt;',
+	'<': '&lt;',
+	'&': '&amp;'
+};
+var rencstr = "&<>'\"".split("");
+
+// TODO: CP remap (need to read file version to determine OS)
+var encregex = /&[a-z]*;/g, coderegex = /_x([\da-fA-F]+)_/g;
+function unescapexml(text){
+	var s = text + '';
+	return s.replace(encregex, function($$) { return encodings[$$]; }).replace(coderegex,function(m,c) {return String.fromCharCode(parseInt(c,16));});
+}
+var decregex=/[&<>'"]/g, charegex = /[\u0000-\u0008\u000b-\u001f]/g;
+function escapexml(text){
+	var s = text + '';
+	return s.replace(decregex, function(y) { return rencoding[y]; }).replace(charegex,function(s) { return "_x" + ("000"+s.charCodeAt(0).toString(16)).substr(-4) + "_";});
+}
+
+function parsexmlbool(value, tag) {
+	switch(value) {
+		case '1': case 'true': case 'TRUE': return true;
+		/* case '0': case 'false': case 'FALSE':*/
+		default: return false;
+	}
+}
+
+function datenum(v) {
+	var epoch = Date.parse(v);
+	return (epoch + 2209161600000) / (24 * 60 * 60 * 1000);
+}
+
+/* ISO 8601 Duration */
+function parse_isodur(s) {
+	var sec = 0, mt = 0, time = false;
+	var m = s.match(/P([0-9\.]+Y)?([0-9\.]+M)?([0-9\.]+D)?T([0-9\.]+H)?([0-9\.]+M)?([0-9\.]+S)?/);
+	if(!m) throw new Error("|" + s + "| is not an ISO8601 Duration");
+	for(var i = 1; i != m.length; ++i) {
+		if(!m[i]) continue;
+		mt = 1;
+		if(i > 3) time = true;
+		switch(m[i].substr(m[i].length-1)) {
+			case 'Y':
+				throw new Error("Unsupported ISO Duration Field: " + m[i].substr(m[i].length-1));
+			case 'D': mt *= 24;
+				/* falls through */
+			case 'H': mt *= 60;
+				/* falls through */
+			case 'M':
+				if(!time) throw new Error("Unsupported ISO Duration Field: M");
+				else mt *= 60;
+				/* falls through */
+			case 'S': break;
+		}
+		sec += mt * parseInt(m[i], 10);
+	}
+	return sec;
+}
+/* copied from js-xls (C) SheetJS Apache2 license */
+function xlml_normalize(d) {
+	if(has_buf && Buffer.isBuffer(d)) return d.toString('utf8');
+	if(typeof d === 'string') return d;
+	throw "badf";
+}
+
+var xlmlregex = /<(\/?)([a-z0-9]*:|)([\w-]+)[^>]*>/mg;
+/* Part 3 Section 4 Manifest File */
+var CT_ODS = "application/vnd.oasis.opendocument.spreadsheet";
+var parse_manifest = function(d, opts) {
+	var str = xlml_normalize(d);
+	var Rn;
+	var FEtag;
+	while((Rn = xlmlregex.exec(str))) switch(Rn[3]) {
+		case 'manifest': break; // 4.2 <manifest:manifest>
+		case 'file-entry': // 4.3 <manifest:file-entry>
+			FEtag = parsexmltag(Rn[0]);
+			if(FEtag.path == '/' && FEtag.type !== CT_ODS) throw new Error("This OpenDocument is not a spreadsheet");
+			break;
+		case 'encryption-data': // 4.4 <manifest:encryption-data>
+		case 'algorithm': // 4.5 <manifest:algorithm>
+		case 'start-key-generation': // 4.6 <manifest:start-key-generation>
+		case 'key-derivation': // 4.7 <manifest:key-derivation>
+			throw new Error("Unsupported ODS Encryption");
+		default: throw Rn;
+	}
+};
+var parse_text_p = function(text, tag) {
+	return text;
+};
+var parse_content_xml = (function() {
+
+	var number_formats = {
+		/* ods name: [short ssf fmt, long ssf fmt] */
+		day: ["d", "dd"],
+		month: ["m", "mm"],
+		year: ["y", "yy"],
+		hours: ["h", "hh"],
+		minutes: ["m", "mm"],
+		seconds: ["s", "ss"],
+		"am-pm": ["A/P", "AM/PM"],
+		"day-of-week": ["ddd", "dddd"]
+	};
+
+	return function pcx(d, opts) {
+		var str = xlml_normalize(d);
+		var state = [], tmp;
+		var tag;
+		var NFtag, NF, pidx;
+		var sheetag;
+		var Sheets = {}, SheetNames = [], ws = {};
+		var Rn, q;
+		var ctag;
+		var textp, textpidx, textptag;
+		var R, C, range = {s: {r:1000000,c:10000000}, e: {r:0, c:0}};
+		var number_format_map = {};
+
+		while((Rn = xlmlregex.exec(str))) switch(Rn[3]) {
+
+			case 'table': // 9.1.2 <table:table>
+				if(Rn[1]==='/') {
+					if(range.e.c >= range.s.c && range.e.r >= range.s.r) ws['!ref'] = get_utils().encode_range(range);
+					SheetNames.push(sheetag.name);
+					Sheets[sheetag.name] = ws;
+				}
+				else if(Rn[0].charAt(Rn[0].length-2) !== '/') {
+					sheetag = parsexmltag(Rn[0]);
+					R = C = -1;
+					range.s.r = range.s.c = 10000000; range.e.r = range.e.c = 0;
+					ws = {};
+				}
+				break;
+
+			case 'table-row':
+				if(Rn[1] === '/') break;
+				++R; C = -1; break; // 9.1.3 <table:table-row>
+			case 'table-cell':
+				if(Rn[0].charAt(Rn[0].length-2) === '/') { ++C; break; } /* stub */
+				if(Rn[1]!=='/') {
+					++C;
+					if(C > range.e.c) range.e.c = C;
+					if(R > range.e.r) range.e.r = R;
+					if(C < range.s.c) range.s.c = C;
+					if(R < range.s.r) range.s.r = R;
+					ctag = parsexmltag(Rn[0]);
+					q = {t:ctag['value-type'], v:null};
+					/* 19.385 office:value-type */
+					switch(q.t) {
+						case 'boolean': q.t = 'b'; q.v = parsexmlbool(ctag['boolean-value']); break;
+						case 'float': q.t = 'n'; q.v = parseFloat(ctag.value); break;
+						case 'percentage': q.t = 'n'; q.v = parseFloat(ctag.value); break;
+						case 'date': q.t = 'n'; q.v = datenum(ctag['date-value']); q.z = 'm/d/yy'; break;
+						case 'time': q.t = 'n'; q.v = parse_isodur(ctag['time-value'])/86400; break;
+						case 'string': q.t = 'str'; break;
+						default: throw new Error('Unsupported value type ' + q.t);
+					}
+				} else {
+					if(q.t === 'str') q.v = textp;
+					if(textp) q.w = textp;
+					ws[get_utils().encode_cell({r:R,c:C})] = q;
+					q = null;
+				}
+				break; // 9.1.4 <table:table-cell>
+
+			case 'document-content': // 3.1.3.2 <office:document-content>
+			case 'spreadsheet': // 3.7 <office:spreadsheet>
+			case 'scripts': // 3.12 <office:scripts>
+			case 'font-face-decls': // 3.14 <office:font-face-decls>
+				if(Rn[1]==='/'){if((tmp=state.pop())[0]!==Rn[3]) throw "Bad state: "+tmp;}
+				else if(Rn[0].charAt(Rn[0].length-2) !== '/') state.push([Rn[3], true]);
+				break;
+
+			case 'number-style': // 16.27.2 <number:number-style>
+			case 'percentage-style': // 16.27.9 <number:percentage-style>
+			case 'date-style': // 16.27.10 <number:date-style>
+			case 'time-style': // 16.27.18 <number:time-style>
+				if(Rn[1]==='/'){
+					number_format_map[NFtag.name] = NF;
+					if((tmp=state.pop())[0]!==Rn[3]) throw "Bad state: "+tmp;
+				} else if(Rn[0].charAt(Rn[0].length-2) !== '/') {
+					NF = "";
+					NFtag = parsexmltag(Rn[0]);
+					state.push([Rn[3], true]);
+				} break;
+
+			case 'script': break; // 3.13 <office:script>
+			case 'automatic-styles': break; // 3.15.3 <office:automatic-styles>
+
+			case 'style': break; // 16.2 <style:style>
+			case 'font-face': break; // 16.21 <style:font-face>
+
+			case 'paragraph-properties': break; // 17.6 <style:paragraph-properties>
+			case 'table-properties': break; // 17.15 <style:table-properties>
+			case 'table-column-properties': break; // 17.16 <style:table-column-properties>
+			case 'table-row-properties': break; // 17.17 <style:table-row-properties>
+			case 'table-cell-properties': break; // 17.18 <style:table-cell-properties>
+
+			case 'number': // 16.27.3 <number:number>
+				switch(state[state.length-1][0]) {
+					case 'time-style':
+					case 'date-style':
+						tag = parsexmltag(Rn[0]);
+						NF += number_formats[Rn[3]][tag.style==='long'?1:0]; break;
+				} break;
+
+			case 'day': // 16.27.11 <number:day>
+			case 'month': // 16.27.12 <number:month>
+			case 'year': // 16.27.13 <number:year>
+			case 'era': // 16.27.14 <number:era>
+			case 'day-of-week': // 16.27.15 <number:day-of-week>
+			case 'week-of-year': // 16.27.16 <number:week-of-year>
+			case 'quarter': // 16.27.17 <number:quarter>
+			case 'hours': // 16.27.19 <number:hours>
+			case 'minutes': // 16.27.20 <number:minutes>
+			case 'seconds': // 16.27.21 <number:seconds>
+			case 'am-pm': // 16.27.22 <number:am-pm>
+				switch(state[state.length-1][0]) {
+					case 'time-style':
+					case 'date-style':
+						tag = parsexmltag(Rn[0]);
+						NF += number_formats[Rn[3]][tag.style==='long'?1:0]; break;
+				} break;
+
+			case 'boolean-style': break; // 16.27.23 <number:boolean-style>
+			case 'boolean': break; // 16.27.24 <number:boolean>
+			case 'text-style': break; // 16.27.25 <number:text-style>
+			case 'text': // 16.27.26 <number:text>
+				if(Rn[0].substr(-2) === "/>") break;
+				else if(Rn[1]==="/") switch(state[state.length-1][0]) {
+					case 'number-style':
+					case 'date-style':
+					case 'time-style':
+						NF += str.slice(pidx, Rn.index);
+						break;
+				}
+				else pidx = Rn.index + Rn[0].length;
+				break;
+			case 'text-content': break; // 16.27.27 <number:text-content>
+			case 'text-properties': break; // 16.27.27 <style:text-properties>
+
+			case 'body': break; // 3.3 16.9.6 19.726.3
+
+			case 'forms': break; // 12.25.2 13.2
+			case 'table-column': break; // 9.1.6 <table:table-column>
+
+			case 'graphic-properties': break;
+			case 'calculation-settings': break; // 9.4.1 <table:calculation-settings>
+			case 'named-expressions': break; // 9.4.11 <table:named-expressions>
+			case 'named-range': break; // 9.4.11 <table:named-range>
+			case 'span': break; // <text:span>
+			case 'p':
+				if(Rn[1]==='/') textp = parse_text_p(str.slice(textpidx,Rn.index), textptag);
+				else { textptag = parsexmltag(Rn[0]); textpidx = Rn.index + Rn[0].length; }
+				break; // <text:p>
+			case 's': break; // <text:s>
+			case 'date': break; // <*:date>
+			case 'annotation': break;
+			default: throw Rn;
+		}
+		var out = {
+			Sheets: Sheets,
+			SheetNames: SheetNames
+		};
+		return out;
+	};
+})();
+/* Part 3: Packages */
+var parse_ods = function(zip, opts) {
+	//var manifest = parse_manifest(getzipdata(zip, 'META-INF/manifest.xml'));
+	return parse_content_xml(getzipdata(zip, 'content.xml'));
+};
+ODS.parse_ods = parse_ods;
+})(typeof exports !== 'undefined' ? exports : ODS);
diff --git a/xlsx.js b/xlsx.js
index d100a15..7a6a376 100644
--- a/xlsx.js
+++ b/xlsx.js
@@ -3,7 +3,7 @@
 /*jshint -W041 */
 var XLSX = {};
 (function(XLSX){
-XLSX.version = '0.7.10';
+XLSX.version = '0.7.11';
 var current_codepage = 1252, current_cptable;
 if(typeof module !== "undefined" && typeof require !== 'undefined') {
 	if(typeof cptable === 'undefined') cptable = require('./dist/cpexcel');
@@ -847,11 +847,17 @@ function getdata(data) {
 	return null;
 }
 
-function getzipfile(zip, file) {
+function safegetzipfile(zip, file) {
 	var f = file; if(zip.files[f]) return zip.files[f];
 	f = file.toLowerCase(); if(zip.files[f]) return zip.files[f];
 	f = f.replace(/\//g,'\\'); if(zip.files[f]) return zip.files[f];
-	throw new Error("Cannot find file " + file + " in zip");
+	return null;
+}
+
+function getzipfile(zip, file) {
+	var o = safegetzipfile(zip, file);
+	if(o == null) throw new Error("Cannot find file " + file + " in zip");
+	return o;
 }
 
 function getzipdata(zip, file, safe) {
@@ -2934,13 +2940,13 @@ function write_ws_xml_cell(cell, ref, ws, opts, idx, wb) {
 }
 
 var parse_ws_xml_data = (function parse_ws_xml_data_factory() {
-	var cellregex = /<(?:\w+:)?c /, rowregex = /<\/(?:\w+:)?row>/;
+	var cellregex = /<(?:\w+:)?c[ >]/, rowregex = /<\/(?:\w+:)?row>/;
 	var rregex = /r=["']([^"']*)["']/, isregex = /<is>([\S\s]*?)<\/is>/;
 	var match_v = matchtag("v"), match_f = matchtag("f");
 
 return function parse_ws_xml_data(sdata, s, opts, guess) {
 	var ri = 0, x = "", cells = [], cref = [], idx = 0, i=0, cc=0, d="", p;
-	var tag;
+	var tag, tagr = 0, tagc = 0;
 	var sstr;
 	var fmtid = 0, fillid = 0, do_format = Array.isArray(styles.CellXf), cf;
 	for(var marr = sdata.split(rowregex), mt = 0, marrlen = marr.length; mt != marrlen; ++mt) {
@@ -2951,18 +2957,19 @@ return function parse_ws_xml_data(sdata, s, opts, guess) {
 		/* 18.3.1.73 row CT_Row */
 		for(ri = 0; ri < xlen; ++ri) if(x.charCodeAt(ri) === 62) break; ++ri;
 		tag = parsexmltag(x.substr(0,ri), true);
-		var tagr = parseInt(tag.r, 10);
+		/* SpreadSheetGear uses implicit r/c */
+		tagr = typeof tag.r !== 'undefined' ? parseInt(tag.r, 10) : tagr+1; tagc = -1;
 		if(opts.sheetRows && opts.sheetRows < tagr) continue;
 		if(guess.s.r > tagr - 1) guess.s.r = tagr - 1;
 		if(guess.e.r < tagr - 1) guess.e.r = tagr - 1;
 
 		/* 18.3.1.4 c CT_Cell */
 		cells = x.substr(ri).split(cellregex);
-		for(ri = 1, cellen = cells.length; ri != cellen; ++ri) {
+		for(ri = typeof tag.r === 'undefined' ? 0 : 1; ri != cells.length; ++ri) {
 			x = cells[ri].trim();
 			if(x.length === 0) continue;
 			cref = x.match(rregex); idx = ri; i=0; cc=0;
-			x = "<c " + x;
+			x = "<c " + (x.substr(0,1)=="<"?">":"") + x;
 			if(cref !== null && cref.length === 2) {
 				idx = 0; d=cref[1];
 				for(i=0; i != d.length; ++i) {
@@ -2970,10 +2977,11 @@ return function parse_ws_xml_data(sdata, s, opts, guess) {
 					idx = 26*idx + cc;
 				}
 				--idx;
-			}
-
+				tagc = idx;
+			} else ++tagc;
 			for(i = 0; i != x.length; ++i) if(x.charCodeAt(i) === 62) break; ++i;
 			tag = parsexmltag(x.substr(0,i), true);
+			if(!tag.r) tag.r = utils.encode_cell({r:tagr-1, c:tagc});
 			d = x.substr(i);
 			p = {t:""};
 
@@ -4804,6 +4812,12 @@ var RecordEnum = {
 };
 
 var evert_RE = evert_key(RecordEnum, 'n');
+/* Helper function to call out to ODS parser */
+function parse_ods(zip, opts) {
+	if(typeof module !== "undefined" && typeof require !== 'undefined' && typeof ODS === 'undefined') ODS = require('./dist/od' + 's');
+	if(typeof ODS === 'undefined' || !ODS.parse_ods) throw new Error("Unsupported ODS");
+	return ODS.parse_ods(zip, opts);
+}
 function fix_opts_func(defaults) {
 	return function fix_opts(opts) {
 		for(var i = 0; i != defaults.length; ++i) {
@@ -4861,6 +4875,10 @@ function parse_zip(zip, opts) {
 	opts = opts || {};
 	fix_read_opts(opts);
 	reset_cp();
+
+	/* OpenDocument Part 3 Section 2.2.1 OpenDocument Package */
+	if(safegetzipfile(zip, 'META-INF/manifest.xml')) return parse_ods(zip, opts);
+
 	var entries = keys(zip.files).filter(nodirs).sort();
 	var dir = parse_ct(getzipdata(zip, '[Content_Types].xml'), opts);
 	var xlsb = false;
diff --git a/xlsxworker.js b/xlsxworker.js
index 9bdc953..a0a4edd 100644
--- a/xlsxworker.js
+++ b/xlsxworker.js
@@ -3,6 +3,8 @@
 //importScripts('dist/cpexcel.js');
 importScripts('jszip.js');
 importScripts('xlsx.js');
+/* uncomment the next line for ODS support */
+importScripts('ods.js');
 postMessage({t:"ready"});
 
 onmessage = function (oEvent) {
diff --git a/xlsxworker1.js b/xlsxworker1.js
index 9567b05..5570bdc 100644
--- a/xlsxworker1.js
+++ b/xlsxworker1.js
@@ -3,6 +3,8 @@
 //importScripts('dist/cpexcel.js');
 importScripts('jszip.js');
 importScripts('xlsx.js');
+/* uncomment the next line for ODS support */
+importScripts('ods.js');
 postMessage({t:"ready"});
 
 function ab2str(data) {
diff --git a/xlsxworker2.js b/xlsxworker2.js
index 6b51645..170a114 100644
--- a/xlsxworker2.js
+++ b/xlsxworker2.js
@@ -3,6 +3,8 @@
 //importScripts('dist/cpexcel.js');
 importScripts('jszip.js');
 importScripts('xlsx.js');
+/* uncomment the next line for ODS support */
+importScripts('ods.js');
 postMessage({t:"ready"});
 
 function ab2str(data) {