Parse numbers according to YAML 1.2 instead of YAML 1.1

dennisleung · Dec 4, 2020 · 1918519 · 1918519
1 parent d1692b9
commit 1918519
Show file tree

Hide file tree

Showing 9 changed files with 73 additions and 66 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -16,6 +16,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - `yaml.Schema.create(schema, tags)` is removed, use `schema.extend(tags)` instead.
 - `!!binary` now always mapped to `Uint8Array` on load.
 - Reduced nesting of `/lib` folder.
+- Parse numbers according to YAML 1.2 instead of YAML 1.1 (`01234` is now decimal,
+  `0o1234` is octal, `1:23` is parsed as string instead of base60).
 
 ### Added
 - Added `.mjs` (es modules) support.

diff --git a/lib/type/float.js b/lib/type/float.js
@@ -5,12 +5,10 @@ var Type   = require('../type');
 
 var YAML_FLOAT_PATTERN = new RegExp(
   // 2.5e4, 2.5 and integers
-  '^(?:[-+]?(?:0|[1-9][0-9_]*)(?:\\.[0-9_]*)?(?:[eE][-+]?[0-9]+)?' +
+  '^(?:[-+]?(?:[0-9][0-9_]*)(?:\\.[0-9_]*)?(?:[eE][-+]?[0-9]+)?' +
   // .2e4, .2
   // special case, seems not from spec
   '|\\.[0-9_]+(?:[eE][-+]?[0-9]+)?' +
-  // 20:59
-  '|[-+]?[0-9][0-9_]*(?::[0-5]?[0-9])+\\.[0-9_]*' +
   // .inf
   '|[-+]?\\.(?:inf|Inf|INF)' +
   // .nan
@@ -30,11 +28,10 @@ function resolveYamlFloat(data) {
 }
 
 function constructYamlFloat(data) {
-  var value, sign, base, digits;
+  var value, sign;
 
   value  = data.replace(/_/g, '').toLowerCase();
   sign   = value[0] === '-' ? -1 : 1;
-  digits = [];
 
   if ('+-'.indexOf(value[0]) >= 0) {
     value = value.slice(1);
@@ -45,22 +42,6 @@ function constructYamlFloat(data) {
 
   } else if (value === '.nan') {
     return NaN;
-
-  } else if (value.indexOf(':') >= 0) {
-    value.split(':').forEach(function (v) {
-      digits.unshift(parseFloat(v, 10));
-    });
-
-    value = 0.0;
-    base = 1;
-
-    digits.forEach(function (d) {
-      value += d * base;
-      base *= 60;
-    });
-
-    return sign * value;
-
   }
   return sign * parseFloat(value, 10);
 }

diff --git a/lib/type/int.js b/lib/type/int.js
@@ -68,25 +68,29 @@ function resolveYamlInteger(data) {
       return hasDigits && ch !== '_';
     }
 
-    // base 8
-    for (; index < max; index++) {
-      ch = data[index];
-      if (ch === '_') continue;
-      if (!isOctCode(data.charCodeAt(index))) return false;
-      hasDigits = true;
+
+    if (ch === 'o') {
+      // base 8
+      index++;
+
+      for (; index < max; index++) {
+        ch = data[index];
+        if (ch === '_') continue;
+        if (!isOctCode(data.charCodeAt(index))) return false;
+        hasDigits = true;
+      }
+      return hasDigits && ch !== '_';
     }
-    return hasDigits && ch !== '_';
   }
 
-  // base 10 (except 0) or base 60
+  // base 10 (except 0)
 
   // value should not start with `_`;
   if (ch === '_') return false;
 
   for (; index < max; index++) {
     ch = data[index];
     if (ch === '_') continue;
-    if (ch === ':') break;
     if (!isDecCode(data.charCodeAt(index))) {
       return false;
     }
@@ -96,15 +100,11 @@ function resolveYamlInteger(data) {
   // Should have digits and should not end with `_`
   if (!hasDigits || ch === '_') return false;
 
-  // if !base60 - done;
-  if (ch !== ':') return true;
-
-  // base60 almost not used, no needs to optimize
-  return /^(:[0-5]?[0-9])+$/.test(data.slice(index));
+  return true;
 }
 
 function constructYamlInteger(data) {
-  var value = data, sign = 1, ch, base, digits = [];
+  var value = data, sign = 1, ch;
 
   if (value.indexOf('_') !== -1) {
     value = value.replace(/_/g, '');
@@ -122,25 +122,8 @@ function constructYamlInteger(data) {
 
   if (ch === '0') {
     if (value[1] === 'b') return sign * parseInt(value.slice(2), 2);
-    if (value[1] === 'x') return sign * parseInt(value, 16);
-    return sign * parseInt(value, 8);
-  }
-
-  if (value.indexOf(':') !== -1) {
-    value.split(':').forEach(function (v) {
-      digits.unshift(parseInt(v, 10));
-    });
-
-    value = 0;
-    base = 1;
-
-    digits.forEach(function (d) {
-      value += (d * base);
-      base *= 60;
-    });
-
-    return sign * value;
-
+    if (value[1] === 'x') return sign * parseInt(value.slice(2), 16);
+    if (value[1] === 'o') return sign * parseInt(value.slice(2), 8);
   }
 
   return sign * parseInt(value, 10);
@@ -158,7 +141,7 @@ module.exports = new Type('tag:yaml.org,2002:int', {
   predicate: isInteger,
   represent: {
     binary:      function (obj) { return obj >= 0 ? '0b' + obj.toString(2) : '-0b' + obj.toString(2).slice(1); },
-    octal:       function (obj) { return obj >= 0 ? '0'  + obj.toString(8) : '-0'  + obj.toString(8).slice(1); },
+    octal:       function (obj) { return obj >= 0 ? '0o'  + obj.toString(8) : '-0o'  + obj.toString(8).slice(1); },
     decimal:     function (obj) { return obj.toString(10); },
     /* eslint-disable max-len */
     hexadecimal: function (obj) { return obj >= 0 ? '0x' + obj.toString(16).toUpperCase() :  '-0x' + obj.toString(16).toUpperCase().slice(1); }

diff --git a/support/demo_template/index.html b/support/demo_template/index.html
@@ -118,7 +118,6 @@ <h4 class="subheader"><a href="#" id="permalink">permalink</a> Edit source here:
   canonical: 6.8523015e+5
   exponentioal: 685.230_15e+03
   fixed: 685_230.15
-  sexagesimal: 190:20:30.15
   negative infinity: -.inf
   not a number: .NaN
 
@@ -127,10 +126,9 @@ <h4 class="subheader"><a href="#" id="permalink">permalink</a> Edit source here:
 int:
   canonical: 685230
   decimal: +685_230
-  octal: 02472256
+  octal: 0o2472256
   hexadecimal: 0x_0A_74_AE
   binary: 0b1010_0111_0100_1010_1110
-  sexagesimal: 190:20:30
 
 # http://yaml.org/type/merge.html ---------------------------------------------#
 

diff --git a/test/issues/0027.js b/test/issues/0027.js
@@ -0,0 +1,48 @@
+'use strict';
+
+
+var assert = require('assert');
+var yaml = require('../../');
+
+
+describe('Should load numbers in YAML 1.2 format', function () {
+  it('should not parse base60', function () {
+    // previously parsed as int
+    assert.strictEqual(yaml.load('1:23'), '1:23');
+    // previously parsed as float
+    assert.strictEqual(yaml.load('1:23.45'), '1:23.45');
+  });
+
+  it('should allow leading zero in int and float', function () {
+    assert.strictEqual(yaml.load('01234'), 1234);
+    assert.strictEqual(yaml.load('00999'), 999);
+    assert.strictEqual(yaml.load('-00999'), -999);
+    assert.strictEqual(yaml.load('001234.56'), 1234.56);
+    assert.strictEqual(yaml.load('001234e4'), 12340000);
+    assert.strictEqual(yaml.load('-001234.56'), -1234.56);
+    assert.strictEqual(yaml.load('-001234e4'), -12340000);
+  });
+
+  it('should parse 0o prefix as octal', function () {
+    assert.strictEqual(yaml.load('0o1234'), 668);
+    // not valid octal
+    assert.strictEqual(yaml.load('0o1289'), '0o1289');
+  });
+});
+
+
+describe('Should dump numbers in YAML 1.2 format', function () {
+  it('should dump in different styles', function () {
+    assert.strictEqual(yaml.dump(123, { styles: { '!!int': 'binary' } }), '0b1111011\n');
+    assert.strictEqual(yaml.dump(123, { styles: { '!!int': 'octal' } }), '0o173\n');
+    assert.strictEqual(yaml.dump(123, { styles: { '!!int': 'hex' } }), '0x7B\n');
+  });
+
+  it('should quote all potential numbers', function () {
+    var tests = '1:23 1:23.45 01234 0999 -01234 01234e4 01234.56 -01234.56 0x123 0o123';
+
+    tests.split(' ').forEach(function (sample) {
+      assert.strictEqual(yaml.dump(sample, { noCompatMode: false }), "'" + sample + "'\n");
+    });
+  });
+});
diff --git a/test/samples-common/construct-float.js b/test/samples-common/construct-float.js
@@ -6,7 +6,6 @@ var expected = {
   canonical: 685230.15,
   exponential: 685230.15,
   fixed: 685230.15,
-  sexagesimal: 685230.15,
   'negative infinity': Number.NEGATIVE_INFINITY,
   'not a number': NaN
 };
@@ -18,7 +17,6 @@ function testHandler(actual) {
   assert.strictEqual(actual['canonical'],         expected['canonical']);
   assert.strictEqual(actual['exponential'],       expected['exponential']);
   assert.strictEqual(actual['fixed'],             expected['fixed']);
-  assert.strictEqual(actual['sexagesimal'],       expected['sexagesimal']);
   assert.strictEqual(actual['negative infinity'], expected['negative infinity']);
 
   assert(Number.isNaN(actual['not a number']));

diff --git a/test/samples-common/construct-float.yml b/test/samples-common/construct-float.yml
@@ -1,6 +1,5 @@
 canonical: 6.8523015e+5
 exponential: 685.230_15e+03
 fixed: 685_230.15
-sexagesimal: 190:20:30.15
 negative infinity: -.inf
 not a number: .NaN
diff --git a/test/samples-common/construct-int.js b/test/samples-common/construct-int.js
@@ -5,6 +5,5 @@ module.exports = {
   decimal: 685230,
   octal: 685230,
   hexadecimal: 685230,
-  binary: 685230,
-  sexagesimal: 685230
+  binary: 685230
 };
diff --git a/test/samples-common/construct-int.yml b/test/samples-common/construct-int.yml
@@ -1,6 +1,5 @@
 canonical: 685230
 decimal: +685_230
-octal: 02472256
+octal: 0o2472256
 hexadecimal: 0x_0A_74_AE
 binary: 0b1010_0111_0100_1010_1110
-sexagesimal: 190:20:30