backup/pywb
1
0
Fork 0
mirror of https://github.com/webrecorder/pywb.git synced 2025-03-24 06:59:52 +01:00
Code Issues Projects Releases Activity

client-side rewrite: for rewrite_html(), pre-rewrite problematic tags (FRAME/TD/TH/TR) that are filtered out if standalone, improves #138

Browse Source
This commit is contained in:
Ilya Kreymer 2017-07-21 12:01:40 -07:00
parent 64d05aca45
commit 9d86601aab
1 changed files with 35 additions and 22 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

57
pywb/static/wombat.js
View File

@ -1415,7 +1415,12 @@ var _WBWombat = function($wbwindow, wbinfo) {
string = write_buff + string; string = write_buff + string;
write_buff = ""; write_buff = "";
} }
var inner_doc = new DOMParser().parseFromString(string, "text/html");
var orig_string = string;
string = string.replace(/<(\/?)(FRAME|TD|TR|TH)\b/ig, "<$1PYWB_$2");
var inner_doc = new DOMParser().parseFromString(string, "text/html");
if (!inner_doc) { if (!inner_doc) {
return string; return string;
@ -1434,34 +1439,42 @@ var _WBWombat = function($wbwindow, wbinfo) {
changed = rewrite_elem(inner_doc.all[i]) || changed; changed = rewrite_elem(inner_doc.all[i]) || changed;
} }
if (!changed) { function get_new_html() {
return string; var new_html;
}
var new_html = ""; // if original had <html> tag, add full document HTML
if (string && string.indexOf("<html") >= 0) {
new_html = inner_doc.documentElement.outerHTML;
} else {
// otherwise, just add contents of head and body
new_html = inner_doc.head.innerHTML;
new_html += inner_doc.body.innerHTML;
// if original had <html> tag, add full document HTML if (check_end_tag) {
if (string && string.indexOf("<html") >= 0) { if (inner_doc.all.length > 3) {
new_html = inner_doc.documentElement.outerHTML; var end_tag = "</" + inner_doc.all[3].tagName.toLowerCase() + ">";
} else { if (ends_with(new_html, end_tag) && !ends_with(string, end_tag)) {
// otherwise, just add contents of head and body new_html = new_html.substring(0, new_html.length - end_tag.length);
new_html = inner_doc.head.innerHTML; }
new_html += inner_doc.body.innerHTML; } else if (string[0] != "<" || string[string.length - 1] != ">") {
write_buff += string;
if (check_end_tag) { return;
if (inner_doc.all.length > 3) {
var end_tag = "</" + inner_doc.all[3].tagName.toLowerCase() + ">";
if (ends_with(new_html, end_tag) && !ends_with(string, end_tag)) {
new_html = new_html.substring(0, new_html.length - end_tag.length);
} }
} else if (string[0] != "<" || string[string.length - 1] != ">") {
write_buff += string;
return;
} }
} }
return new_html;
} }
return new_html; if (changed) {
string = get_new_html();
}
if (string && string != orig_string) {
string = string.replace(/<(\/?)PYWB_(FRAME|TD|TR|TH)\b/ig, "<$1$2");
}
return string;
} }
//============================================ //============================================