aboutsummaryrefslogtreecommitdiff
path: root/bg
diff options
context:
space:
mode:
authorhackademix <giorgio@maone.net>2018-07-15 23:15:34 +0200
committerhackademix <giorgio@maone.net>2018-07-17 17:45:27 +0200
commitdbe1ddbc1ddd9b40c6e0e5c896768f91cb6994e0 (patch)
treec3eb5a2b9bdbf59901d9da2602a0635ea3cc00c7 /bg
parenteb929a0aff3a23a53f750ab3c866fbec57107fbb (diff)
Refactor and fix HTTP response filtering to touch only scripts, either external or embedded inside HTML documents and sub-documents.
Diffstat (limited to 'bg')
-rw-r--r--bg/ResponseMetaData.js82
-rw-r--r--bg/ResponseProcessor.js110
2 files changed, 192 insertions, 0 deletions
diff --git a/bg/ResponseMetaData.js b/bg/ResponseMetaData.js
new file mode 100644
index 0000000..40ca3f3
--- /dev/null
+++ b/bg/ResponseMetaData.js
@@ -0,0 +1,82 @@
+/**
+* GNU LibreJS - A browser add-on to block nonfree nontrivial JavaScript.
+*
+* Copyright (C) 2018 Giorgio Maone <giorgio@maone.net>
+*
+* This file is part of GNU LibreJS.
+*
+* GNU LibreJS is free software: you can redistribute it and/or modify
+* it under the terms of the GNU General Public License as published by
+* the Free Software Foundation, either version 3 of the License, or
+* (at your option) any later version.
+*
+* GNU LibreJS is distributed in the hope that it will be useful,
+* but WITHOUT ANY WARRANTY; without even the implied warranty of
+* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+* GNU General Public License for more details.
+*
+* You should have received a copy of the GNU General Public License
+* along with GNU LibreJS. If not, see <http://www.gnu.org/licenses/>.
+*/
+
+/**
+ This class parses HTTP response headers to extract both the
+ MIME Content-type and the character set to be used, if specified,
+ to parse textual data through a decoder.
+*/
+
+class ResponseMetaData {
+ constructor(request) {
+ let {responseHeaders} = request;
+ this.headers = {};
+ for (let h of responseHeaders) {
+ if (/^\s*Content-(Type|Disposition)\s*$/i.test(h.name)) {
+ let propertyName = h.name.split("-")[1].trim();
+ propertyName = `content${propertyName.charAt(0).toUpperCase()}${propertyName.substring(1).toLowerCase()}`;
+ this[propertyName] = h.value;
+ this.headers[propertyName] = h;
+ }
+ }
+ this.forcedUTF8 = false;
+ }
+
+ get charset() {
+ let charset = "";
+ if (this.contentType) {
+ let m = this.contentType.match(/;\s*charset\s*=\s*(\S+)/);
+ if (m) {
+ charset = m[1];
+ }
+ }
+ Object.defineProperty(this, "charset", { value: charset, writable: false, configurable: true });
+ return charset;
+ }
+
+ get isUTF8() {
+ return /^utf-8$/i.test(this.charset);
+ }
+
+ forceUTF8() {
+ if (!(this.forcedUTF8 || this.isUTF8)) {
+ let h = this.headers.contentType;
+ if (h) {
+ h.value = h.value.replace(/;\s*charset\s*=.*|$/, "; charset=utf8");
+ this.forcedUTF8 = true;
+ } // if the header doesn't exist the browser should default to UTF-8 anyway
+ }
+ return this.forcedUTF8;
+ }
+
+ createDecoder() {
+ if (this.charset) {
+ try {
+ return new TextDecoder(this.charset);
+ } catch (e) {
+ console.error(e);
+ }
+ }
+ return new TextDecoder("utf-8");
+ }
+};
+
+module.exports = { ResponseMetaData };
diff --git a/bg/ResponseProcessor.js b/bg/ResponseProcessor.js
new file mode 100644
index 0000000..3f3151b
--- /dev/null
+++ b/bg/ResponseProcessor.js
@@ -0,0 +1,110 @@
+/**
+* GNU LibreJS - A browser add-on to block nonfree nontrivial JavaScript.
+*
+* Copyright (C) 2018 Giorgio Maone <giorgio@maone.net>
+*
+* This file is part of GNU LibreJS.
+*
+* GNU LibreJS is free software: you can redistribute it and/or modify
+* it under the terms of the GNU General Public License as published by
+* the Free Software Foundation, either version 3 of the License, or
+* (at your option) any later version.
+*
+* GNU LibreJS is distributed in the hope that it will be useful,
+* but WITHOUT ANY WARRANTY; without even the implied warranty of
+* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+* GNU General Public License for more details.
+*
+* You should have received a copy of the GNU General Public License
+* along with GNU LibreJS. If not, see <http://www.gnu.org/licenses/>.
+*/
+
+/**
+ An abstraction layer over the StreamFilter API, allowing its clients to process
+ only the "interesting" HTML and script requests and leaving the other alone
+*/
+
+let {ResponseMetaData} = require("./ResponseMetaData.js");
+
+let listeners = new WeakMap();
+let webRequestEvent = browser.webRequest.onHeadersReceived;
+
+class ResponseProcessor {
+
+ static install(handler, types = ["main_frame", "sub_frame", "script"]) {
+ if (listeners.has(handler)) return false;
+ let listener =
+ request => new ResponseTextFilter(request).process(handler);
+ listeners.set(handler, listener);
+ webRequestEvent.addListener(
+ listener,
+ {urls: ["<all_urls>"], types},
+ ["blocking", "responseHeaders"]
+ );
+ return true;
+ }
+
+ static uninstall(handler) {
+ let listener = listeners.get(handler);
+ if (listener) {
+ webRequestEvent.removeListener(listener);
+ }
+ }
+}
+
+class ResponseTextFilter {
+ constructor(request) {
+ this.request = request;
+ let {type, statusCode} = request;
+ let md = this.metaData = new ResponseMetaData(request);
+ this.canProcess = // we want to process html documents and scripts only
+ (statusCode < 300 || statusCode >= 400) && // skip redirections
+ !md.disposition && // skip forced downloads
+ (type === "script" || /\bhtml\b/i.test(md.contentType));
+ }
+
+ process(handler) {
+ if (!this.canProcess) return {};
+ let metaData = this.metaData;
+ let {requestId, responseHeaders} = this.request;
+ let filter = browser.webRequest.filterResponseData(requestId);
+ let buffer = [];
+
+ filter.ondata = event => {
+ buffer.push(event.data);
+ };
+
+ filter.onstop = async event => {
+ let decoder = metaData.createDecoder();
+ let params = {stream: true};
+ let text = this.text = buffer.map(
+ chunk => decoder.decode(chunk, params))
+ .join('');
+ let editedText = null;
+ try {
+ let response = {
+ request: this.request,
+ metaData,
+ text,
+ };
+ editedText = await handler(response);
+ } catch(e) {
+ console.error(e);
+ }
+ if (metaData.forcedUTF8 ||
+ editedText !== null && text !== editedText) {
+ // if we changed the charset, the text or both, let's re-encode
+ filter.write(new TextEncoder().encode(editedText));
+ } else {
+ // ... otherwise pass all the raw bytes through
+ for (let chunk of buffer) filter.write(chunk);
+ }
+
+ filter.disconnect();
+ }
+
+ return metaData.forceUTF8() ? {responseHeaders} : {};
+ }
+}
+
+module.exports = { ResponseProcessor };