420380 - allow gethash responses that don't match table/chunk id

Status: RESOLVED FIXED

(Toolkit :: Safe Browsing, defect, P1)

Description

[Dave Camp (:dcamp)]

Right now we require gethash responses to match table and chunk id to consider the match valid.  This is necessary to get the caching right, but in a few rare cases it will cause false negatives:

a) If a url was removed from one list and added to another since we last updated
b) If a url was removed and readded to the list since we last updated
c) Google screws up and sends the wrong table/chunk ids because their gethash server is out of sync with their update server (the case that actually prompted this bug report).

Attached patch will accept a gethash match with an incorrect table/chunk id, but won't cache the result.

Comment 1

[Dave Camp (:dcamp)]

Attachment: fix

Oops, forgot to attach the patch.

Comment 2

[Dave Camp (:dcamp)]

Attachment: test updates

... and some test updates.

Comment 3

[Dave Camp (:dcamp)]

Checking in src/nsUrlClassifierDBService.cpp;
/cvsroot/mozilla/toolkit/components/url-classifier/src/nsUrlClassifierDBService.cpp,v  <--  nsUrlClassifierDBService.cpp
new revision: 1.69; previous revision: 1.68
done
Checking in tests/unit/test_partial.js;
/cvsroot/mozilla/toolkit/components/url-classifier/tests/unit/test_partial.js,v  <--  test_partial.js
new revision: 1.7; previous revision: 1.6
done