Closed Bug 1214454 Opened 9 years ago Closed 9 years ago

Switching the TP list should trigger a list download on next startup

Categories

(Firefox :: Settings UI, defect, P1)

defect

Tracking

()

VERIFIED FIXED
Firefox 44
Iteration:
44.2 - Oct 19
Tracking Status
firefox44 --- verified

People

(Reporter: francois, Assigned: past)

References

Details

(Whiteboard: [fxprivacy])

Attachments

(1 file)

Bug 1175562 changed the timing around list updates in two important ways:

1. restarting the browser doesn't force a list update, it honors the "nextupdatetime" for each list
2. the initial list download doesn't happen automatically anymore and could be delayed by up to 5 minutes

The way that #2 works is that when the browser starts for the first time (fresh profile), it doesn't have these prefs:

  browser.safebrowsing.provider.mozilla.lastupdatetime
  browser.safebrowsing.provider.mozilla.nextupdatetime

and when these prefs are missing, the initial update could be delayed by up
to 5 minutes (a requirement of the Safe Browsing protocol).

Unfortunately the logs don't actually mention that because the message is suppressed for the initial update (see bug 1214448).

In order to force an immediate update on startup, we should set:

  browser.safebrowsing.provider.mozilla.nextupdatetime = 42

at the same time as we set the urlclassifier.trackingTable pref.

Note: 42 is not a magic value, but rather it's a non-zero timestamp that is definitely in the past.
Assignee: nobody → past
Status: NEW → ASSIGNED
Priority: -- → P1
Whiteboard: [fxprivacy]
Flags: qe-verify?
Whiteboard: [fxprivacy] → [fxprivacy] [triage]
Flags: qe-verify? → qe-verify+
Attached patch Patch v1Splinter Review
This seems to do the trick.
Attachment #8673548 - Flags: review?(francois)
Attachment #8673548 - Flags: review?(francois) → review+
https://hg.mozilla.org/mozilla-central/rev/9c53d7a22d54
Status: ASSIGNED → RESOLVED
Closed: 9 years ago
Resolution: --- → FIXED
Target Milestone: --- → Firefox 44
Iteration: --- → 44.2 - Oct 19
Whiteboard: [fxprivacy] [triage] → [fxprivacy]
Hi Florin, can a QA contact be assigned for verification of this resolved bug.
Blocks: 1188565
Flags: needinfo?(florin.mezei)
Flags: needinfo?(florin.mezei)
QA Contact: paul.silaghi
https://bugzilla.mozilla.org/show_bug.cgi?id=1177085#c68
(In reply to Paul Silaghi, QA [:pauly] from comment #68)
> I'm not sure I completely understand the implications of this feature.
> In private browsing, on cnn.com with "disconnect.me basic protection" =>
> control center says: "nightly is blocking parts of the page that may track
> your browsing", TP shield displayed.
> With "disconnect.me strict protection" => control center says: "no tracking
> elements detected on this page", TP shield is not displayed.
Now the TP shield is not displayed on cnn.com with "disconnect.me basic protection" on a clean profile, when the "browser.safebrowsing.provider.mozilla.lastupdatetime" and browser.safebrowsing.provider.mozilla.nextupdatetime" prefs are missing.
The TP shield icon appears after switching to "disconnect.me strict protection".
44.0a1 (2015-10-18) Win 7
Thoughts?
Flags: needinfo?(past)
I would guess that problem fixes itself after 5 minutes, when the DB is actually downloaded? Or if you switch to strict and back to basic?

If we are confident in our own server side we can ship the prefs lastupdatetime/nextupdatetime preset to 42, but note that there's still a timing window in which you can get the same result - if you visit the site before the update completes.
(In reply to Gian-Carlo Pascutto [:gcp] from comment #6)
> I would guess that problem fixes itself after 5 minutes, when the DB is
> actually downloaded? Or if you switch to strict and back to basic?
Confirmed, in both cases. 44.0a1 (2015-10-19)
gcp knows that part better than I do. We might have to live with this edge case in 42.
Flags: needinfo?(past)
I'm marking this bug as verified based on comments 5,6,7. Feel free to reopen if you consider the above behavior as not expected.
Status: RESOLVED → VERIFIED
(In reply to Panos Astithas [:past] from comment #8)
> gcp knows that part better than I do. We might have to live with this edge
> case in 42.

Fx42 and 43 are not affected by this 5-min delay since that got added in 44.

When gcp said "42" in comment 6, I think he meant defaulting to "browser.safebrowsing.provider.mozilla.nextupdatetime = 42" in Firefox 44, which is probably a good idea if we want to make TP work out of the box.

I filed bug 1216604 for this.
(In reply to François Marier [:francois] from comment #10)
> When gcp said "42" in comment 6, I think he meant defaulting to
> "browser.safebrowsing.provider.mozilla.nextupdatetime = 42" in Firefox 44,
> which is probably a good idea if we want to make TP work out of the box.

D'oh, that makes more sense!
Depends on: 1216604
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: