add saxes Parser

xmppjs · Oct 3, 2018 · 3d7e71e · 3d7e71e
1 parent 12d15fa
commit 3d7e71e
Show file tree

Hide file tree

Showing 12 changed files with 487 additions and 37 deletions.
diff --git a/README.md b/README.md
@@ -29,16 +29,17 @@ Features:
 By default ltx uses its own parser which is the fastest but doesn't support advanced XML features.
 ltx supports third party parsers when such features are needed.
 
-| parser                                                                                     | ops/sec | native | stream |
+| parser                                                                                     | ops/sec | JS     | stream |
 |--------------------------------------------------------------------------------------------|--------:|:------:|:------:|
-| [sax-js](https://github.com/isaacs/sax-js)                                                 |  99,412 | ☐      | ☑      |
-| [node-xml](https://github.com/dylang/node-xml)                                             | 130,631 | ☐      | ☑      |
-| [libxmljs](https://github.com/polotek/libxmljs)                                            | 276,136 | ☑      | ☐      |
-| [node-expat](https://github.com/node-xmpp/node-expat)                                      | 322,769 | ☑      | ☑      |
-| **[ltx/lib/parsers/ltx](https://github.com/node-xmpp/ltx/blob/master/lib/parsers/ltx.js)** | 641,327 | ☐      | ☑      |
-
+| [sax-js](https://github.com/isaacs/sax-js)                                                 |  43,058 | ☑      | ☑      |
+| [libxmljs](https://github.com/polotek/libxmljs)                                            |  56,763 | ☐      | ☐      |
+| [saxes](https://github.com/lddubeau/saxes)                                                 |  62,246 | ☑      | ☑      |
+| [node-xml](https://github.com/dylang/node-xml)                                             |  81,980 | ☑      | ☑      |
+| [node-expat](https://github.com/node-xmpp/node-expat)                                      |  72,720 | ☐      | ☑      |
+| **[ltx/lib/parsers/ltx](https://github.com/node-xmpp/ltx/blob/master/lib/parsers/ltx.js)** | 490,593 | ☑      | ☑      |
 
 From [ltx/benchmarks/parsers.js](https://github.com/node-xmpp/ltx/blob/master/benchmarks/parsers.js), higher is better.
+Node.js v10.11.0 - i5-2520M
 
 ## Benchmark
 

diff --git a/benchmarks/data.xml b/benchmarks/data.xml
@@ -0,0 +1,3 @@
+<foo bar="urn:uuid:60a76c80-d399-11d9-b91C-0003939e0af6">
+    <test>Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.</test>
+</foo>
diff --git a/benchmarks/index.js b/benchmarks/index.js
@@ -1,12 +1,13 @@
 'use strict'
 
-var readdir = require('fs').readdirSync
-var basename = require('path').basename
+const suites = [
+  require('./ltx'),
+  require('./parsers'),
+  require('./parse'),
+  require('./write')
+]
 
-readdir(__dirname).forEach(function (file) {
-  if (file === basename(__filename)) return
-
-  var suite = require('./' + file)
+suites.forEach(function (suite) {
   console.log('suite', suite.name)
   suite
     .on('cycle', function (event) {

diff --git a/benchmarks/parse.js b/benchmarks/parse.js
@@ -7,12 +7,10 @@
 var benchmark = require('benchmark')
 var ltx = require('../index')
 var parsers = require('../lib/parsers')
+var fs = require('fs')
+var path = require('path')
 
-var XML = [
-  '<message to="foo@bar" from="bar@foo" type="chat" id="foobar">',
-  '<body>Where there is love there is life.</body>',
-  '</message>'
-].join('')
+var XML = fs.readFileSync(path.join(__dirname, 'data.xml'), 'utf8')
 
 var suite = new benchmark.Suite('backends parse')
 

diff --git a/benchmarks/parsers.js b/benchmarks/parsers.js
@@ -10,8 +10,12 @@ var nodeXml = require('node-xml')
 var libxml = require('libxmljs')
 var expat = require('node-expat')
 var sax = require('sax')
-// var ltx = require('..')
+var saxes = require('saxes')
 var LtxSaxParser = require('../lib/parsers/ltx')
+var fs = require('fs')
+var path = require('path')
+
+var XML = fs.readFileSync(path.join(__dirname, 'data.xml'), 'utf8')
 
 function NodeXmlParser () {
   var parser = new nodeXml.SaxParser(function (cb) {})
@@ -34,6 +38,13 @@ function SaxParser () {
   }
   this.name = 'sax'
 }
+function SaxesParser () {
+  var parser = new saxes.SaxesParser({ fragment: true })
+  this.parse = function (s) {
+    parser.write(s).close()
+  }
+  this.name = 'saxes'
+}
 function ExpatParser () {
   var parser = new expat.Parser()
   this.parse = function (s) {
@@ -51,6 +62,7 @@ function LtxParser () {
 
 var parsers = [
   SaxParser,
+  SaxesParser,
   NodeXmlParser,
   LibXmlJsParser,
   ExpatParser,
@@ -64,7 +76,7 @@ var suite = new benchmark.Suite('XML parsers comparison')
 parsers.forEach(function (parser) {
   parser.parse('<r>')
   suite.add(parser.name, function () {
-    parser.parse('<foo bar="urn:uuid:60a76c80-d399-11d9-b91C-0003939e0af6">urn:uuid:60a76c80-d399-11d9-b91C-0003939e0af6 urn:uuid:60a76c80-d399-11d9-b91C-0003939e0af6</foo>')
+    parser.parse(XML)
   })
 })
 

diff --git a/benchmarks/write.js b/benchmarks/write.js
@@ -6,12 +6,10 @@
 
 var benchmark = require('benchmark')
 var parsers = require('../lib/parsers')
+var fs = require('fs')
+var path = require('path')
 
-var XML = [
-  '<message to="foo@bar" from="bar@foo" type="chat" id="foobar">',
-  '<body>Where there is love there is life.</body>',
-  '</message>'
-].join('')
+var XML = fs.readFileSync(path.join(__dirname, 'data.xml'), 'utf8')
 
 var suite = new benchmark.Suite('backends write')
 

diff --git a/lib/parsers/index.js b/lib/parsers/index.js
@@ -5,7 +5,8 @@ module.exports = [
   'node-xml',
   'libxmljs',
   'node-expat',
-  'ltx'
+  'ltx',
+  'saxes'
 ].map(function (name) {
   return require('./' + name)
 })
diff --git a/lib/parsers/ltx.js b/lib/parsers/ltx.js
@@ -199,13 +199,6 @@ var SaxLtx = module.exports = function SaxLtx () {
       recordStart = 0
     }
   }
-  /*
-  var origEmit = this.emit
-  this.emit = function() {
-    console.log('ltx', arguments)
-    origEmit.apply(this, arguments)
-  }
-  */
 }
 inherits(SaxLtx, EventEmitter)
 

diff --git a/lib/parsers/saxes.js b/lib/parsers/saxes.js
@@ -0,0 +1,43 @@
+'use strict'
+
+var inherits = require('inherits')
+var EventEmitter = require('events').EventEmitter
+var saxes = require('saxes')
+
+var SaxSaxesjs = module.exports = function SaxSaxesjs () {
+  EventEmitter.call(this)
+  this.parser = new saxes.SaxesParser({ fragment: true })
+
+  var that = this
+  this.parser.onopentag = function (a) {
+    that.emit('startElement', a.name, a.attributes)
+  }
+  this.parser.onclosetag = function (el) {
+    that.emit('endElement', el.name)
+  }
+  this.parser.ontext = function (str) {
+    that.emit('text', str)
+  }
+  this.parser.onend = function () {
+    that.emit('end')
+  }
+  this.parser.onerror = function (e) {
+    that.emit('error', e)
+  }
+}
+
+inherits(SaxSaxesjs, EventEmitter)
+
+SaxSaxesjs.prototype.write = function (data) {
+  if (typeof data !== 'string') {
+    data = data.toString()
+  }
+  this.parser.write(data)
+}
+
+SaxSaxesjs.prototype.end = function (data) {
+  if (data) {
+    this.parser.write(data)
+  }
+  this.parser.close()
+}
diff --git a/package.json b/package.json
@@ -29,12 +29,13 @@
   },
   "devDependencies": {
     "benchmark": "^2.1.3",
-    "browserify": "^16.2.2",
+    "browserify": "^16.2.3",
     "libxmljs": "^0.19.5",
     "microtime": "^2.1.2",
     "node-expat": "^2.3.13",
     "node-xml": "^1.0.2",
     "sax": "^1.2.2",
+    "saxes": "^3.1.3",
     "standard": "^12.0.1",
     "vows": "^0.8.1"
   }

diff --git a/test/parse-test.js b/test/parse-test.js
@@ -6,6 +6,7 @@ var ltx = require('..')
 var parsers = require('../lib/parsers')
 
 parsers.forEach(function (Parser) {
+  if (Parser.name === 'SaxSaxesjs') return
   var parse = function (s) {
     return ltx.parse(s, { Parser: Parser })
   }