The W3C XMLHttpRequest
specification adds HTML parsing support to XMLHttpRequest
, which originally supported only XML parsing. This feature allows Web apps to obtain an HTML resource as a parsed DOM using XMLHttpRequest
.
Limitations
To discourage the synchronous use of XMLHttpRequest
, HTML support is not available in the synchronous mode. Also, HTML support is only available if the responseType
property has been set to "document"
. This limitation avoids wasting time parsing HTML uselessly when legacy code uses XMLHttpRequest
in the default mode to retrieve responseText
for text/html
resources. Also, this limitation avoids problems with legacy code that assumes that responseXML
is null for HTTP error pages (which often have a text/html
response body).
Usage
Retrieving an HTML resource as a DOM using XMLHttpRequest
works just like retrieving an XML resource as a DOM using XMLHttpRequest
, except you can't use the synchronous mode and you have to explicitly request a document by assigning the string "document"
to the responseType
property of the XMLHttpRequest
object after calling open()
but before calling send()
.
var xhr = new XMLHttpRequest(); xhr.onload = function() { console.log(this.responseXML.title); } xhr.open("GET", "file.html"); xhr.responseType = "document"; xhr.send();
Feature Detection
Method 1
This method relies on the "force async" nature of the feature. When you try to set responseType
of an XMLHttpRequest
object after it is opened as "sync". This throws an error on the browsers implemented the feature and works on others.
function HTMLinXHR() { if (!window.XMLHttpRequest) return false; var req = new window.XMLHttpRequest(); req.open('GET', window.location.href, false); try { req.responseType = 'document'; } catch(e) { return true; } return false; }
This method is synchronous, does not rely on external assets though it may not be as reliable as method 2 described below since it does not check the actual feature but an indication of that feature.
Method 2
There are two challenges to detecting exactly if a browser supports HTML parsing in XMLHttpRequest
. First, the detection result is obtained asynchronously, because HTML support is only available in the asynchronous mode. Second, you have to actually fetch a test document over HTTP, because testing with a data:
URL would end up testing data:
URL support at the same time.
Thus, to detect HTML support, a test HTML file is needed on the server. This test file is small and is not well-formed XML:
<title>&&<</title>
If the file is named detect.html
, the following function can be used for detecting HTML parsing support:
function detectHtmlInXhr(callback) { if (!window.XMLHttpRequest) { window.setTimeout(function() { callback(false); }, 0); return; } var done = false; var xhr = new window.XMLHttpRequest(); xhr.onreadystatechange = function() { if (this.readyState == 4 && !done) { done = true; callback(!!(this.responseXML && this.responseXML.title && this.responseXML.title == "&&<")); } } xhr.onabort = xhr.onerror = function() { if (!done) { done = true; callback(false); } } try { xhr.open("GET", "detect.html"); xhr.responseType = "document"; xhr.send(); } catch (e) { window.setTimeout(function() { if (!done) { done = true; callback(false); } }, 0); } }
The argument callback
is a function that will be called asynchronously with true
as the only argument if HTML parsing is supported and false
as the only argument if HTML parsing is not supported.
Character encoding
If the character encoding is declared in the HTTP Content-Type
header, that character encoding is used. Failing that, if there is a byte order mark, the encoding indicated by the byte order mark is used. Failing that, if there is a meta
tag that declares the encoding within the first 1024 bytes of the file, that encoding is used. Otherwise, the file is decoded as UTF-8.
Handling HTML on older browsers
XMLHttpRequest
originally supported only XML parsing. HTML parsing support is a recent addition. For older browsers, you can even use the responseText
property in association with Regular Expressions in order to get, for example, the source code of an HTML element knowing its ID:
function getHTML (oXHR, sTargetId) { var rOpen = new RegExp("<(?!\!)\\s*([^\\s>]+)[^>]*\\s+id\\=[\"\']" + sTargetId + "[\"\'][^>]*>" ,"i"), sSrc = oXHR.responseText, aExec = rOpen.exec(sSrc); return aExec ? (new RegExp("(?:(?:.(?!<\\s*" + aExec[1] + "[^>]*[>]))*.?<\\s*" + aExec[1] + "[^>]*[>](?:.(?!<\\s*\/\\s*" + aExec[1] + "\\s*>))*.?<\\s*\/\\s*" + aExec[1] + "\\s*>)*(?:.(?!<\\s*\/\\s*" + aExec[1] + "\\s*>))*.?", "i")).exec(sSrc.slice(sSrc.indexOf(aExec[0]) + aExec[0].length)) || "" : ""; } var oReq = new XMLHttpRequest(); oReq.open("GET", "yourPage.html", true); oReq.onload = function () { console.log(getHTML(this, "intro")); }; oReq.send(null);
Specifications
Specification | Status | Comment |
---|---|---|
XMLHttpRequest | Living Standard | Initial definition |
Browser compatibility
Feature | Chrome | Firefox (Gecko) | Internet Explorer | Opera | Safari (WebKit) |
---|---|---|---|---|---|
Support | 18 | 11.0 (11.0) | 10 | ? | No support (Yes)[1] |
Feature | Android | Firefox Mobile (Gecko) | IE Phone | Opera Mobile | Safari Mobile |
---|---|---|---|---|---|
Support | ? | 11.0 (11.0) | ? | ? | ? |
[1] This feature was implemented in WebKit bug 74626.