Vivaldi-Snapshot vs Vivaldi-Stable Command Surprise.
-
I run all my browsers in Firejail. Nothing has changed re that practice recently. [Some of] My long-time launchers are respectively:
1. firejail -- vivaldi-snapshot --disk-cache-dir=/tmp/vivaldi-snapshot-cache 2. firejail -- vivaldi-stable --disk-cache-dir=/tmp/vivaldi-stable-cache 3. firejail -- chromium --disk-cache-dir=/tmp/chromium-cache
Another of my long-time installed applications is
Net Activity Viewer
. I don't use it very often, probably only every ~10^-2 Halley's Comet [standard temporal units, donchaknow?]. Today was its time to shine once again.NAV reports not only stuff like protocol, local host, local address, local port, remote address, remote port, remote host... but also program & command. When i launch each of the above browsers, NAV reports for those latter two parameters:
-
vivaldi-bin
&/opt/vivaldi-snapshot/vivaldi-bin --type=utility --field-trial-handle=******************,*******************,****** --lang=en-GB --running-vivaldi --service-sandbox-type=network --service-request-channel-token=******************** --shared-files=v8_natives_data:100,v8_snapshot_data:101
-
vivaldi-bin
&/opt/vivaldi/vivaldi-bin --disk-cache-dir=/tmp/vivaldi-stable-cache
-
chromium
&/usr/lib/chromium/chromium --disk-cache-dir=/tmp/chromium-cache
My rarely-reliable memory dimly suggests to me that in prior times V-SS used to report in NAV with the more intuitive command
/opt/vivaldi-snapshot/vivaldi-bin --disk-cache-dir=/tmp/vivaldi-snapshot-cache
.My info above shows that C & V-S continue to NAV-report commands consistent with my custom launchers [which seems sensible & appropriate]. Very evidently however, V-SS has gone wildly off into the weeds with its reported command... with some "interesting" names & numbers [each replaced above with same-number of asterisks for privacy given i do not know what these numbers are doing] within its mega-long command.
What gives here nowadays with Snapshot?
UPDATE 19/4/19: Don't understand how i missed it originally, but in fact both V-Stable & Chromium DO also manifest unique-per-installation numbers with
field-trial-handle
... see my later post #4. -
-
@Steffie Many of my VMs also have Vivaldi/s installed. In one of them i have just now compiled & also installed
Net Activity Viewer
, then launched V-SS. For this V-SS instance, per NAV:/opt/vivaldi-snapshot/vivaldi-bin --type=utility --field-trial-handle=*******************,********************,131072 --lang=en-AU --running-vivaldi --service-sandbox-type=network --service-request-channel-token=******************** --shared-files=v8_natives_data:100,v8_snapshot_data:101
.Comparison reveals that the 1st, 2nd & 4th numbers are unique per V-SS installation, whereas the 3rd is common [hence i've now revealed it].
Sooooo, this uniqueness all seems rather... interesting
-
It appears that my preceding posts, whilst deliberately heavy on insinuation, might have been a tad too subtle. Hence now, gloves off. Does anyone know:
- Why Snapshot is now running with a command that includes the strange
--field-trial-handle
option / switch [it did not used to]? What does this mean? - What the implications would be from Snapshot now including apparently three unique numbers [per installation] with that preceding option / switch?
Btw, using NAV is not the sole way to see these complex command strings. They also show clearly in
htop
& KDEKSysGuard
, for instance......albeit they do NOT show in Vivaldi Snapshot's Help - About:
Vivaldi 2.5.1511.4 (Official Build) snapshot (64-bit) Revision 6b05515843d8a6856ab94290cf183a64f13dc69d OS Linux JavaScript V8 7.4.288.13 Flash (Disabled) User Agent Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/74.0.3729.55 Safari/537.36 Vivaldi/2.5.1511.4 Command Line /usr/bin/vivaldi-snapshot --disk-cache-dir=/tmp/vivaldi-snapshot-cache --flag-switches-begin --allow-insecure-localhost --ignore-gpu-blacklist --flag-switches-end --save-page-as-mhtml Executable Path /opt/vivaldi-snapshot/vivaldi-snapshot Profile Path /home/steffie/.config/vivaldi-snapshot/Default
- Why Snapshot is now running with a command that includes the strange
-
I wondered if maybe this apparent tracking ID# stamping might somehow be originating not from the Vivaldi code itself, but somehow as a consequence of the Arch packaging done by the maintainer of the
herecura
repo. Hence i looked for a way to independently test that conjecture.Launched my Mint 19 Cinnamon VM [ie, not Arch, not Manjaro, not
herecura
], which already had V-Stable installed, but not Snapshot. To my surprise, its Stable also manifests these unique-per-installation numbers withfield-trial-handle
, gasp.Installed V-SS from the Mint repos, & as now pretty much expected , it too manifests unique-per-installation numbers with
field-trial-handle
. To eliminate any chance of this being caused by Mint's repo packaging, i then uninstalled SS, downloaded direct from Vivaldi site, & installed. Once again, unique-per-installation numbers withfield-trial-handle
.What's going on here please, Vivaldi?
Oh deary me, my head is just exploding now. I've discovered that even Chromium, in Mint, is manifesting its own unique-per-installation numbers with
field-trial-handle
.Idea: In Mint C VM, all my browsers are running naked [no Firejail], whereas in my real OS they all run in FJ... so might this difference be a factor? Launched my real OS' Chromium, in FJ, reconfirmed NO "tracking". Closed it, relaunched C naked... whoa, now lots of the tracking. Closed it, relaunched in FJ... well blow me down, now also has all the tracking. Strewth.
Tested that idea again with V-Stable in real OS, & discovered that my OP contains an error. Somehow i seem to have not noticed originally, but in fact even in my real Manjaro OS, regardless of running naked or in FJ, BOTH V-Snapshot AND V-Stable are manifesting their own unique-per-installation numbers with
field-trial-handle
.Given today's discovery re Chromium, it seems pretty clear that it is not Vivaldi [Stable & Snapshot] doing this of their own volition, but rather it seems to be an inherited behavioural trait from Chromium. Sigh.
-
@Steffie For what it's worth, on macOS, Vivaldi has been launching with
--field-trial-handle=
(followed by comma-separated numbers) going all the way back to version 1.11. -
@Steffie I can see that you're concerned about this and I'll try to clarify what I can.
"Field Trials" is a Chrome/Chromium thing. Google uses them primarily for A/B testing and for when they want to do a staged rollout of new code or a new feature to a certain percentage of the population. To the best of my knowledge, they are not used for tracking. I know that Field Trials are enabled in Chrome but I don't know about Chromium builds. You also can't generally tell whether or not you've been enrolled in a field trial but you can reset to a certain extent by launching Chrome with the
--reset-variation-state
option.Field trials are considered a "bad thing" from a Vivaldi perspective because enabling experimental Chromium code could potentially break Vivaldi, and that breakage would be random and would only affect some of the population. I'll have to confirm this with the devs but as far as I know, Vivaldi does whatever it can to disable Field Trials when they build the browser.
-
@xyzzy Many thanks. Yes i am concerned about this, actually initially i was a bit shocked. I'd never heard of this before, & per my OP, afaik in "older" days my V's never did this, nor my C [though given i use that much less often, i can't be sure].
You said
To the best of my knowledge, they are not used for tracking
...but i hope you can appreciate that to at least some of us, discovery of unique Google ID numbers attached to our individual V's [& C's] is not even slightly conducive to a feeling of comfort.
you can reset to a certain extent by launching Chrome with the --reset-variation-state option
...is that something we now should do with V?
enabling experimental Chromium code could potentially break Vivaldi, and that breakage would be random and would only affect some of the population
...well that is incredibly significant!
Vivaldi does whatever it can to disable Field Trials when they build the browser
...& yet as my chance discovery then subsequent tests revealed, as well as your own
on macOS, Vivaldi has been launching with --field-trial-handle= (followed by comma-separated numbers) going all the way back to version 1.11
, somehow our V's are still afflicted with this [IMO] Google blight. -
@Steffie I don't think that there's anything that you need to do because, as far as I know, Vivaldi has Field Trials disabled internally. (About a year ago, when Google was selectively enabling "Strict Site Isolation" in Chrome, the Vivaldi team definitely didn't want this to get enabled in Vivaldi because, at the time, Strict Site Isolation and enabling Out-of_Process iframes broke Vivaldi's code in some really bad ways.)
There's also no need to invoke Vivaldi with --reset-variation-state. (unless, Heaven forbid!, field trials actually did get enabled somehow and an unwanted field trial got activated.) I only mentioned that parameter so that you would have another search term for your own research on Field Trials.
As for the privacy aspects, even if the Field Trial code was active, I don't think that it can be used to track you. Besides, there are far more effective ways to track you, especially if you take security/privacy protections and countermeasures to an extreme and end up giving your web browser a very unique and distinct fingerprint.
-
@xyzzy OK, thanks. I really appreciate your helpful info.
-
-