add much needed documentation about whats going on in the background script (most importantly, the cookie setting code)

This commit is contained in:
Peter Snyder 2017-10-14 17:39:27 -05:00
parent 259789ad81
commit ca1eb8cf5a

View file

@ -65,6 +65,12 @@
}, 1000);
// Listen for updates to the domain rules from the config page.
// The two types of messages that are sent to the background page are
// "rulesUpdate", which comes from the config page, indicating the domain
// blocking / matching rules have changed, and the "rulesForDomains"
// message, which comes from the browserAction popup, and is a request
// for information about "here are the domains of the frames on the
// current page, which rules are being used to match them".
rootObject.runtime.onMessage.addListener(function (request, ignore, sendResponse) {
const [label, data] = request;
@ -95,6 +101,18 @@
const requestOptions = ["blocking", "responseHeaders"];
// Inject the blocking settings for each visited domain / frame.
// This needs to be done syncronously, so that the DOM of the visited
// page can be instrumented at "document_start" time. This means we
// can't do any of the "obvious" techniques for loading the "what should"
// be blocked in this frame" information (ie using the storage API).
// So, instead, we halt at the http query point, match the domain being
// loaded against the current rule set, pack the set of standards
// that should be blocked into a base64 encoded bitfield, and then
// push that to the page as a cookie.
//
// The page then reads the information about what standards to block
// out of the cookie (by decoding and unpacking the bitfield), and then
// deletes the cookie, so nothing is left behind.
rootObject.webRequest.onHeadersReceived.addListener(function (details) {
const url = details.url;