Skip to content

John McEleney / casperjs

Go to a project
Toggle navigation
Toggle navigation pinning
Commit 10495da5 10495da54a17f73579dd1c7f5ffb4202ab71d54a by Nicolas Perriault
Options

fixes #178 - plain text version for non-html contents.

1 parent 7d9f0e18
Inline Side-by-side
Showing 2 changed files with 3 additions and 29 deletions
......@@ -943,36 +943,10 @@ Casper.prototype.getPageContent = function getPageContent() {
"use strict";
this.checkStarted();
var contentType = utils.getPropertyPath(this, 'currentResponse.contentType');
if (!utils.isString(contentType)) {
if (!utils.isString(contentType) || contentType.indexOf("text/html") !== -1) {
return this.page.frameContent;
}
// for some reason (qt)webkit/Gecko will always enclose non text/html body contents within an html
// structure like this:
// webkit: <html><head></head><body><pre style="(...)">content</pre></body></html>
// gecko: <html><head><link rel="alternate stylesheet" type="text/css" href="resource://gre-resources/plaintext.css" title="..."></head><body><pre>document.write('foo');\n</pre></body></html>
var sanitizedHtml = this.evaluate(function checkHtml() {
var head = __utils__.findOne('head'),
body = __utils__.findOne('body');
if (!head || !body) {
return null;
}
// for content in Webkit
if (head.childNodes.length === 0 &&
body.childNodes.length === 1 &&
__utils__.findOne('body pre[style]')) {
return __utils__.findOne('body pre').textContent.trim();
}
// for content in Gecko
if (head.childNodes.length === 1 &&
body.childNodes.length === 1 &&
head.childNodes[0].localName === 'link' &&
head.childNodes[0].getAttribute('href') === 'resource://gre-resources/plaintext.css' &&
body.childNodes[0].localName === 'pre' ) {
return body.childNodes[0].textContent.trim();
}
return null;
});
return sanitizedHtml ? sanitizedHtml : this.page.frameContent;
return this.page.framePlainText;
};
/**
......
......@@ -13,7 +13,7 @@ casper.test.begin("Casper.getPageContent() text/html content", 1, function(test)
casper.test.begin("Casper.getPageContent() non text/html content", 1, function(test) {
casper.start("tests/site/dummy.js", function() {
test.assertEquals(this.getPageContent(), "document.write('foo');",
test.assertEquals(this.getPageContent(), "document.write('foo');\n",
"Casper.getPageContent() retrieves non text/html content");
}).run(function() {
test.done();
......