From 6bea47aaefef2059ea4a5dfc01cd93e3aed11bdc Mon Sep 17 00:00:00 2001
From: Hafez <a.hafez852@gmail.com>
Date: Mon, 8 Aug 2022 22:41:08 +0200
Subject: [PATCH] allow reading base64 files from a dataURI scheme (#2763)

---
 bits/04_base64.js    |  4 +++-
 modules/04_base64.js |  4 +++-
 modules/04_base64.ts |  3 ++-
 test.js              | 33 +++++++++++++++++++++++++++++----
 4 files changed, 37 insertions(+), 7 deletions(-)

diff --git a/bits/04_base64.js b/bits/04_base64.js
index 53f40f5..0164e3f 100644
--- a/bits/04_base64.js
+++ b/bits/04_base64.js
@@ -48,7 +48,9 @@ function Base64_encode_pass(input) {
 function Base64_decode(input) {
   var o = "";
   var c1 = 0, c2 = 0, c3 = 0, e1 = 0, e2 = 0, e3 = 0, e4 = 0;
-  input = input.replace(/[^\w\+\/\=]/g, "");
+  input = input.replace(/^data:([^\/]+\/[^\/]+)?;base64,/, '')
+               .replace(/[^\w\+\/\=]/g, "");
+
   for (var i = 0; i < input.length; ) {
     e1 = Base64_map.indexOf(input.charAt(i++));
     e2 = Base64_map.indexOf(input.charAt(i++));
diff --git a/modules/04_base64.js b/modules/04_base64.js
index 53f40f5..fdd658a 100644
--- a/modules/04_base64.js
+++ b/modules/04_base64.js
@@ -48,7 +48,9 @@ function Base64_encode_pass(input) {
 function Base64_decode(input) {
   var o = "";
   var c1 = 0, c2 = 0, c3 = 0, e1 = 0, e2 = 0, e3 = 0, e4 = 0;
-  input = input.replace(/[^\w\+\/\=]/g, "");
+  input = input.replace(/^data:.+\/.+;base64\,/,'')
+               .replace(/[^\w\+\/\=]/g, "")
+  
   for (var i = 0; i < input.length; ) {
     e1 = Base64_map.indexOf(input.charAt(i++));
     e2 = Base64_map.indexOf(input.charAt(i++));
diff --git a/modules/04_base64.ts b/modules/04_base64.ts
index 154a446..adbd40a 100644
--- a/modules/04_base64.ts
+++ b/modules/04_base64.ts
@@ -40,7 +40,8 @@ function Base64_encode_pass(input: string): string {
 function Base64_decode(input: string): string {
 	var o = "";
 	var c1=0, c2=0, c3=0, e1=0, e2=0, e3=0, e4=0;
-	input = input.replace(/[^\w\+\/\=]/g, "");
+	input = input.replace(/^data:.+\/.+;base64\,/,'')
+	             .replace(/[^\w\+\/\=]/g, "")
 	for(var i = 0; i < input.length;) {
 		e1 = Base64_map.indexOf(input.charAt(i++));
 		e2 = Base64_map.indexOf(input.charAt(i++));
diff --git a/test.js b/test.js
index 1c25e3e..f7d9cd5 100644
--- a/test.js
+++ b/test.js
@@ -688,6 +688,31 @@ describe('input formats', function() {
 	it('should read base64 strings', function() { artifax.forEach(function(p) {
 		X.read(fs.readFileSync(p, 'base64'), {type: 'base64'});
 	}); });
+
+	it('handles base64 within data URI scheme (gh-2762)', function() {
+		// Arrange
+		var fileInBase64 = 'TmFtZXMNCkhhZmV6DQpTYW0NCg==';
+		var fileInBase64WithDataURIScheme = 'data:text/csv;base64,TmFtZXMNCkhhZmV6DQpTYW0NCg==';
+
+		// Act
+		var workBookFromRawBase64 = X.read(fileInBase64, { type: 'base64' });
+		var workBookFromBase64WithinDataURI = X.read(fileInBase64WithDataURIScheme, { type: 'base64' });
+
+		// Assert
+		assert.deepStrictEqual(workBookFromRawBase64, workBookFromBase64WithinDataURI);
+	});
+	it('handles base64 where data URI has no media type (gh-2762)', function() {
+		// Arrange
+		var fileInBase64 = 'TmFtZXMNCkhhZmV6DQpTYW0NCg==';
+		var fileInBase64WithDataURIScheme = 'data:;base64,TmFtZXMNCkhhZmV6DQpTYW0NCg==';
+
+		// Act
+		var workBookFromRawBase64 = X.read(fileInBase64, { type: 'base64' });
+		var workBookFromBase64WithinDataURI = X.read(fileInBase64WithDataURIScheme, { type: 'base64' });
+
+		// Assert
+		assert.deepStrictEqual(workBookFromRawBase64, workBookFromBase64WithinDataURI);
+	});
 	if(typeof Uint8Array !== 'undefined') it('should read array', function() { artifax.forEach(function(p) {
 		X.read(fs.readFileSync(p, 'binary').split("").map(function(x) { return x.charCodeAt(0); }), {type:'array'});
 	}); });
@@ -1393,7 +1418,7 @@ describe('parse features', function() {
 	});
 
 	describe('data types formats', function() {[
-		['xlsx', paths.dtfxlsx],
+		['xlsx', paths.dtfxlsx]
 	].forEach(function(m) { it(m[0], function() {
 		var wb = X.read(fs.readFileSync(m[1]), {type: TYPE, cellDates: true});
 		var ws = wb.Sheets[wb.SheetNames[0]];
@@ -2109,7 +2134,7 @@ function plaintext_test(wb, raw) {
 	var sheet = wb.Sheets[wb.SheetNames[0]];
 	plaintext_val.forEach(function(x) {
 		var cell = get_cell(sheet, x[0]);
-		var tcval = x[2+(!!raw ? 1 : 0)];
+		var tcval = x[2+(raw ? 1 : 0)];
 		var type = raw ? 's' : x[1];
 		if(x.length == 1) { if(cell) { assert.equal(cell.t, 'z'); assert.ok(!cell.v); } return; }
 		assert.equal(cell.v, tcval); assert.equal(cell.t, type);
@@ -2201,8 +2226,8 @@ describe('CSV', function() {
 			var aoa = [
 				["3a", "3 a", "3 a-1"],
 				["3b", "3 b", "3 b-1"],
-				["3p", "3 P", "3 p-1"],
-			]
+				["3p", "3 P", "3 p-1"]
+			];
 			var ws = X.read(aoa.map(function(row) { return row.join(","); }).join("\n"), {type: "string", cellDates: true}).Sheets.Sheet1;
 			for(var R = 0; R < 3; ++R) {
 				assert.equal(get_cell(ws, "A" + (R+1)).v, aoa[R][0]);