From c4277c4e09bb0fda01d147c9608f43da2c0b6d7c Mon Sep 17 00:00:00 2001
From: Jonathan Kingston <jkingston@duckduckgo.com>
Date: Fri, 27 Oct 2023 14:58:00 +0100
Subject: [PATCH] Expire storage and hash hostname

---
 shared/js/background/broken-site-report.js | 42 +++++++++++++++++++---
 shared/js/background/events.js             |  5 +++
 2 files changed, 42 insertions(+), 5 deletions(-)

diff --git a/shared/js/background/broken-site-report.js b/shared/js/background/broken-site-report.js
index 668628e382..f206720a4f 100644
--- a/shared/js/background/broken-site-report.js
+++ b/shared/js/background/broken-site-report.js
@@ -88,7 +88,17 @@ const requestCategoryMapping = {
     'ignore-user': 'ignoredByUserRequests'
 }
 
-function computeLastSentDay (urlString) {
+async function digestMessage (message) {
+    const msgUint8 = new TextEncoder().encode(message)
+    const hashBuffer = await crypto.subtle.digest('SHA-256', msgUint8)
+    const hashArray = Array.from(new Uint8Array(hashBuffer))
+    const hashHex = hashArray
+        .map((b) => b.toString(16).padStart(2, '0'))
+        .join('')
+    return hashHex
+}
+
+async function computeLastSentDay (urlString) {
     const url = new URL(urlString)
     const time = new Date()
     // Round time to nearest day
@@ -96,16 +106,38 @@ function computeLastSentDay (urlString) {
     // Output time as a string in the format YYYY-MM-DD
     const dayOutput = time.toISOString().split('T')[0]
 
+    // Use a sha256 hash prefix of the hostname so that we don't store the full hostname
+    const hash = await digestMessage(url.hostname)
+    const hostnameHashPrefix = hash.slice(0, 6)
+
     const reportTimes = settings.getSetting('brokenSiteReportTimes') || {}
-    const lastSentDay = reportTimes[url.hostname]
+    const lastSentDay = reportTimes[hostnameHashPrefix]
 
     // Update existing time
-    reportTimes[url.hostname] = dayOutput
+    reportTimes[hostnameHashPrefix] = dayOutput
     settings.updateSetting('brokenSiteReportTimes', reportTimes)
 
     return lastSentDay
 }
 
+/**
+ * Clears any expired broken site report times
+ * Called by an alarm every hour to remove entries older than 30 days
+ */
+export async function clearExpiredBrokenSiteReportTimes () {
+    await settings.ready()
+    const brokenSiteReports = settings.getSetting('brokenSiteReportTimes') || {}
+    // Expiry of 30 days
+    const expiryTime = new Date().getTime() - 30 * 24 * 60 * 60 * 1000
+    for (const hashPrefix in brokenSiteReports) {
+        const reportTime = new Date(brokenSiteReports[hashPrefix])
+        if (reportTime.getTime() < expiryTime) {
+            delete brokenSiteReports[hashPrefix]
+        }
+    }
+    settings.updateSetting('brokenSiteReportTimes', brokenSiteReports)
+}
+
 /**
  * Given an optional category and description, create a report for a given Tab instance.
  *
@@ -121,7 +153,7 @@ function computeLastSentDay (urlString) {
  * @prop {string | undefined} arg.category - optional category
  * @prop {string | undefined} arg.description - optional description
  */
-export function breakageReportForTab ({
+export async function breakageReportForTab ({
     tab, tds, remoteConfigEtag, remoteConfigVersion,
     category, description
 }) {
@@ -155,7 +187,7 @@ export function breakageReportForTab ({
     const debugFlags = tab.debugFlags.join(',')
     const errorDescriptions = JSON.stringify(tab.errorDescriptions)
     const httpErrorCodes = tab.httpErrorCodes.join(',')
-    const lastSentDay = computeLastSentDay(tab.url)
+    const lastSentDay = await computeLastSentDay(tab.url)
 
     const brokenSiteParams = new URLSearchParams({
         siteUrl,
diff --git a/shared/js/background/events.js b/shared/js/background/events.js
index cbdfe512c3..b35021e531 100644
--- a/shared/js/background/events.js
+++ b/shared/js/background/events.js
@@ -17,6 +17,7 @@ import {
 import tdsStorage from './storage/tds'
 import httpsStorage from './storage/https'
 import ATB from './atb'
+import { clearExpiredBrokenSiteReportTimes } from './broken-site-report'
 const utils = require('./utils')
 const experiment = require('./experiments')
 const settings = require('./settings')
@@ -433,6 +434,8 @@ browserWrapper.createAlarm('clearExpiredHTTPSServiceCache', { periodInMinutes: 6
 browserWrapper.createAlarm('rotateUserAgent', { periodInMinutes: 24 * 60 })
 // Rotate the sessionKey
 browserWrapper.createAlarm('rotateSessionKey', { periodInMinutes: 24 * 60 })
+// Expire site breakage reports
+browserWrapper.createAlarm('clearExpiredBrokenSiteReportTimes', { periodInMinutes: 24 * 60 })
 
 browser.alarms.onAlarm.addListener(async alarmEvent => {
     // Warning: Awaiting in this function doesn't actually wait for the promise to resolve before unblocking the main thread.
@@ -459,6 +462,8 @@ browser.alarms.onAlarm.addListener(async alarmEvent => {
         await utils.resetSessionKey()
     } else if (alarmEvent.name === REFETCH_ALIAS_ALARM) {
         fetchAlias()
+    } else if (alarmEvent.name === 'clearExpiredBrokenSiteReportTimes') {
+        await clearExpiredBrokenSiteReportTimes()
     }
 })