version bump 0.7.3: performance

- require cpexcel instead of full codepage library
- hardcode the unescapexml regexp
- utf8read short circuit for ASCII strings
- cellFormulae also acts on XLSX/XLSM
- bin/xlsx.njs do not process formulae unless requested

README.md

@@ -125,7 +125,7 @@ The exported `read` and `readFile` functions accept an options argument:
 
 | Option Name | Default | Description |
 | :---------- | ------: | :---------- |
-| cellFormula | true    | Save formulae to the .f field ** |
+| cellFormula | true    | Save formulae to the .f field |
 | cellHTML    | true    | Parse rich text and save HTML to the .h field |
 | cellNF      | false   | Save number format string to the .z field |
 | sheetStubs  | false   | Create cell objects for stub cells |
@@ -136,8 +136,6 @@ The exported `read` and `readFile` functions accept an options argument:
 | bookSheets  | false   | If true, only parse enough to get the sheet names |
 | bookVBA     | false   | If true, expose vbaProject.bin to `vbaraw` field ** |
 
-- `cellFormula` only applies to constructing XLSB formulae.  XLSX/XLSM formulae
-  are stored in plaintext, but XLSB formulae are stored in a binary format.
 - Even if `cellNF` is false, formatted text (.w) will be generated
 - In some cases, sheets may be parsed even if `bookSheets` is false.
 - `bookSheets` and `bookProps` combine to give both sets of information

bin/xlsx.njs

@@ -20,6 +20,8 @@ program
 	.option('-F, --field-sep <sep>', 'CSV field separator', ",")
 	.option('-R, --row-sep <sep>', 'CSV row separator', "\n")
 	.option('-n, --sheet-rows <num>', 'Number of rows to process (0=all rows)')
+	.option('--no-sst', 'do not generate sst')
+	.option('--perf', 'do not generate output')
 	.option('--dev', 'development mode')
 	.option('--read', 'read but do not print out contents')
 	.option('-q, --quiet', 'quiet mode');
@@ -57,6 +59,9 @@ if(program.xlsx || program.xlsm || program.xlsb) {
 	opts.cellNF = true;
 	if(program.output) sheetname = program.output;
 }
+else if(program.formulae);
+else opts.cellFormula = false;
+
 if(program.dev) {
 	X.verbose = 2;
 	opts.WTF = true;
@@ -77,7 +82,7 @@ if(program.listSheets) {
 	process.exit(0);
 }
 
-var wopts = {WTF:opts.WTF};
+var wopts = {WTF:opts.WTF, bookSST:program.sst};
 
 if(program.xlsx) return X.writeFile(wb, sheetname || (filename + ".xlsx"), wopts);
 if(program.xlsm) return X.writeFile(wb, sheetname || (filename + ".xlsm"), wopts);
@@ -95,6 +100,8 @@ try {
 	process.exit(4);
 }
 
+if(program.perf) return;
+
 var oo = ""; 
 if(!program.quiet) console.error(target_sheet);
 if(program.formulae) oo = X.utils.get_formulae(ws).join("\n");

bits/01_version.js

@@ -1 +1 @@
-XLSX.version = '0.7.2';
+XLSX.version = '0.7.3';

bits/02_codepage.js

@@ -1,6 +1,6 @@
 var current_codepage = 1252, current_cptable;
 if(typeof module !== "undefined" && typeof require !== 'undefined') {
-	if(typeof cptable === 'undefined') cptable = require('codepage');
+	if(typeof cptable === 'undefined') cptable = require('./dist/cpexcel');
 	current_cptable = cptable[current_codepage];
 }
 function reset_cp() { set_cp(1252); }

bits/36_xmlutils.js

@@ -27,7 +27,8 @@ var rencstr = "&<>'\"".split("");
 // TODO: CP remap (need to read file version to determine OS)
 function unescapexml(text){
 	var s = text + '';
-	for(var y in encodings) s = s.replace(new RegExp(y,'g'), encodings[y]);
+	s = s.replace(/&quot;/g, '"').replace(/&apos;/g, "'").replace(/&gt;/g, ">").replace(/&lt;/g, "<").replace(/&amp;/g, "&");
+	//for(var y in encodings) s = s.replace(new RegExp(y,'g'), encodings[y]);
 	return s.replace(/_x([0-9a-fA-F]*)_/g,function(m,c) {return _chr(parseInt(c,16));});
 }
 function escapexml(text){
@@ -48,6 +49,7 @@ function parsexmlbool(value, tag) {
 
 var utf8read = function(orig) {
 	var out = [], i = 0, c = 0, c1 = 0, c2 = 0, c3 = 0;
+	if(!orig.match(/[\u0080-\uffff]/)) return orig;
 	while (i < orig.length) {
 		c = orig.charCodeAt(i++);
 		if (c < 128) out.push(_chr(c));

bits/72_wsxml.js

@@ -18,7 +18,7 @@ function parse_ws_xml(data, opts, rels) {
 	}
 
 	var refguess = {s: {r:1000000, c:1000000}, e: {r:0, c:0} };
-	var q = ["v","f"];
+	var q = (opts.cellFormula ? ["v","f"] : ["v"]);
 	var sidx = 0;
 
 	/* 18.3.1.80 sheetData CT_SheetData ? */

dist/cpexcel.js

@@ -959,10 +959,12 @@ if (typeof module !== 'undefined' && module.exports) module.exports = cptable;
   var encache = function(cp) {
     if(typeof Buffer !== 'undefined') {
       sbcs_cache.forEach(function(s) {
+        if(!cpt[s]) return;
         cpdcache[s] = sbcs_decode(s);
         cpecache[s] = sbcs_encode(s);
       });
       dbcs_cache.forEach(function(s) {
+        if(!cpt[s]) return;
         cpdcache[s] = dbcs_decode(s);
         cpecache[s] = dbcs_encode(s);
       });

dist/xlsx.js

@@ -2,10 +2,10 @@
 /* vim: set ts=2: */
 var XLSX = {};
 (function(XLSX){
-XLSX.version = '0.7.2';
+XLSX.version = '0.7.3';
 var current_codepage = 1252, current_cptable;
 if(typeof module !== "undefined" && typeof require !== 'undefined') {
-	if(typeof cptable === 'undefined') cptable = require('codepage');
+	if(typeof cptable === 'undefined') cptable = require('./dist/cpexcel');
 	current_cptable = cptable[current_codepage];
 }
 function reset_cp() { set_cp(1252); }
@@ -675,7 +675,8 @@ var rencstr = "&<>'\"".split("");
 // TODO: CP remap (need to read file version to determine OS)
 function unescapexml(text){
 	var s = text + '';
-	for(var y in encodings) s = s.replace(new RegExp(y,'g'), encodings[y]);
+	s = s.replace(/&quot;/g, '"').replace(/&apos;/g, "'").replace(/&gt;/g, ">").replace(/&lt;/g, "<").replace(/&amp;/g, "&");
+	//for(var y in encodings) s = s.replace(new RegExp(y,'g'), encodings[y]);
 	return s.replace(/_x([0-9a-fA-F]*)_/g,function(m,c) {return _chr(parseInt(c,16));});
 }
 function escapexml(text){
@@ -696,6 +697,7 @@ function parsexmlbool(value, tag) {
 
 var utf8read = function(orig) {
 	var out = [], i = 0, c = 0, c1 = 0, c2 = 0, c3 = 0;
+	if(!orig.match(/[\u0080-\uffff]/)) return orig;
 	while (i < orig.length) {
 		c = orig.charCodeAt(i++);
 		if (c < 128) out.push(_chr(c));
@@ -2165,7 +2167,7 @@ function parse_ws_xml(data, opts, rels) {
 	}
 
 	var refguess = {s: {r:1000000, c:1000000}, e: {r:0, c:0} };
-	var q = ["v","f"];
+	var q = (opts.cellFormula ? ["v","f"] : ["v"]);
 	var sidx = 0;
 
 	/* 18.3.1.80 sheetData CT_SheetData ? */

package.json

@@ -1,6 +1,6 @@
 {
 	"name": "xlsx",
-	"version": "0.7.2-a",
+	"version": "0.7.3",
 	"author": "sheetjs",
 	"description": "XLSB/XLSX/XLSM (Excel 2007+ Spreadsheet) parser and writer",
 	"keywords": [ "xlsx", "xlsb", "xlsm", "office", "excel", "spreadsheet" ],

tests.lst

@@ -82,6 +82,7 @@ apachepoi_53568.xlsx
 apachepoi_53734.xlsx
 apachepoi_53798.xlsx
 apachepoi_53798_shiftNegative_TMPL.xlsx
+apachepoi_54034.xlsx
 apachepoi_54071.xlsx
 apachepoi_54084 - Greek - beyond BMP.xlsx
 apachepoi_54206.xlsx

xlsx.js

@@ -2,10 +2,10 @@
 /* vim: set ts=2: */
 var XLSX = {};
 (function(XLSX){
-XLSX.version = '0.7.2';
+XLSX.version = '0.7.3';
 var current_codepage = 1252, current_cptable;
 if(typeof module !== "undefined" && typeof require !== 'undefined') {
-	if(typeof cptable === 'undefined') cptable = require('codepage');
+	if(typeof cptable === 'undefined') cptable = require('./dist/cpexcel');
 	current_cptable = cptable[current_codepage];
 }
 function reset_cp() { set_cp(1252); }
@@ -675,7 +675,8 @@ var rencstr = "&<>'\"".split("");
 // TODO: CP remap (need to read file version to determine OS)
 function unescapexml(text){
 	var s = text + '';
-	for(var y in encodings) s = s.replace(new RegExp(y,'g'), encodings[y]);
+	s = s.replace(/&quot;/g, '"').replace(/&apos;/g, "'").replace(/&gt;/g, ">").replace(/&lt;/g, "<").replace(/&amp;/g, "&");
+	//for(var y in encodings) s = s.replace(new RegExp(y,'g'), encodings[y]);
 	return s.replace(/_x([0-9a-fA-F]*)_/g,function(m,c) {return _chr(parseInt(c,16));});
 }
 function escapexml(text){
@@ -696,6 +697,7 @@ function parsexmlbool(value, tag) {
 
 var utf8read = function(orig) {
 	var out = [], i = 0, c = 0, c1 = 0, c2 = 0, c3 = 0;
+	if(!orig.match(/[\u0080-\uffff]/)) return orig;
 	while (i < orig.length) {
 		c = orig.charCodeAt(i++);
 		if (c < 128) out.push(_chr(c));
@@ -2165,7 +2167,7 @@ function parse_ws_xml(data, opts, rels) {
 	}
 
 	var refguess = {s: {r:1000000, c:1000000}, e: {r:0, c:0} };
-	var q = ["v","f"];
+	var q = (opts.cellFormula ? ["v","f"] : ["v"]);
 	var sidx = 0;
 
 	/* 18.3.1.80 sheetData CT_SheetData ? */