public-offering
/
pad.pub0.org
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

use cheerio instead of jsdom

pull/2341/head
John McLear 2014-11-25 17:26:09 +00:00
parent 97068b562d
commit 60d44cd3df
3 changed files with 11 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
src/node/utils/ImportHtml.js
View File

@ -14,22 +14,18 @@
* limitations under the License. * limitations under the License.
*/ */
var jsdom = require('jsdom-nocontextifiy').jsdom; var jsdom = require('jsdom').jsdom;
var log4js = require('log4js'); var log4js = require('log4js');
var Changeset = require("ep_etherpad-lite/static/js/Changeset"); var Changeset = require("ep_etherpad-lite/static/js/Changeset");
var contentcollector = require("ep_etherpad-lite/static/js/contentcollector"); var contentcollector = require("ep_etherpad-lite/static/js/contentcollector");
var cheerio = require("cheerio");
function setPadHTML(pad, html, callback) function setPadHTML(pad, html, callback)
{ {
var apiLogger = log4js.getLogger("ImportHtml"); var apiLogger = log4js.getLogger("ImportHtml");
// Parse the incoming HTML with jsdom var $ = cheerio.load(html);
try{ var doc = $('html')[0];
var doc = jsdom(html.replace(/>\n+</g, '><'));
}catch(e){
apiLogger.warn("Error importing, possibly caused by malformed HTML");
var doc = jsdom("<html><body><div>Error during import, possibly malformed HTML</div></body></html>");
}
apiLogger.debug('html:'); apiLogger.debug('html:');
apiLogger.debug(html); apiLogger.debug(html);
@ -38,7 +34,7 @@ function setPadHTML(pad, html, callback)
// using the content collector object // using the content collector object
var cc = contentcollector.makeContentCollector(true, null, pad.pool); var cc = contentcollector.makeContentCollector(true, null, pad.pool);
try{ // we use a try here because if the HTML is bad it will blow up try{ // we use a try here because if the HTML is bad it will blow up
cc.collectContent(doc.childNodes[0]); cc.collectContent(doc);
}catch(e){ }catch(e){
apiLogger.warn("HTML was not properly formed", e); apiLogger.warn("HTML was not properly formed", e);
return; // We don't process the HTML because it was bad.. return; // We don't process the HTML because it was bad..

2
src/package.json
View File

@ -25,7 +25,7 @@
"formidable" : "1.0.9", "formidable" : "1.0.9",
"log4js" : "0.6.6", "log4js" : "0.6.6",
"nodemailer" : "0.3.x", "nodemailer" : "0.3.x",
"jsdom-nocontextifiy" : "0.2.10", "cheerio" : "0.18.0",
"async-stacktrace" : "0.0.2", "async-stacktrace" : "0.0.2",
"npm" : "1.4.x", "npm" : "1.4.x",
"ejs" : "0.6.1", "ejs" : "0.6.1",

5
src/static/js/contentcollector.js
View File

@ -54,10 +54,14 @@ function makeContentCollector(collectStyles, browser, apool, domInterface, class
}, },
nodeNumChildren: function(n) nodeNumChildren: function(n)
{ {
if(n.childNodes == null) return 0;
return n.childNodes.length; return n.childNodes.length;
}, },
nodeChild: function(n, i) nodeChild: function(n, i)
{ {
if(n.childNodes.item == null){
return n.childNodes[i];
}
return n.childNodes.item(i); return n.childNodes.item(i);
}, },
nodeProp: function(n, p) nodeProp: function(n, p)
@ -66,6 +70,7 @@ function makeContentCollector(collectStyles, browser, apool, domInterface, class
}, },
nodeAttr: function(n, a) nodeAttr: function(n, a)
{ {
if(n.getAttribute == null) return null;
return n.getAttribute(a); return n.getAttribute(a);
}, },
optNodeInnerHTML: function(n) optNodeInnerHTML: function(n)