Closed klack closed 1 month ago
Why would you expect it to be effective for font fingerprinting? It isn't.
Why would you expect it to be effective for font fingerprinting? It isn't.
We can read both the feature and the page linked, together:
"Scale the output values of Canvas::measureText() with a randomly selected factor in the range -0.0003% to 0.0003%, which are recomputed on every document initialization."
"Font fingerprinting techniques rely on measuring the dimensions of HTML elements filled with text or single Unicode glyphs."
I suppose what I am seeking is something else entirely, so I will close this.
I think the confusion is that there are multiple different methods of font fingerprinting and that flag only deals with canvas fingerprinting. So the test you'd want to use is: https://browserleaks.com/canvas
The test you were using is for font enumeration, and there aren't any easy ways of dealing with that. Some browsers tried to deal with that by supplying a small set of fonts that are the only ones allowed to be used but that ends up breaking a lot of sites.
Some browsers tried to deal with that by supplying a small set of fonts that are the only ones allowed to be used but that ends up breaking a lot of sites.
AFAIK Cromite does just that with the end-result being all websites use the same set of fonts and the web becomes "tasteless" (except for websites that don't use local fonts and rely on remote ones instead), but I didn't notice any real breakages, so that is definitely doable, but as usual it costs the convenience and usability. And yes, if you wonder, it can effortlessly pass the test mentioned.
OS/Platform
Debian, Ubuntu, and derivatives
Installed
Flatpak
Version
129.0.6668.70
Have you tested that this is not an upstream issue or an issue with your configuration?
--user-data-dir
command line argument and it could not be reproduced thereDescription
Canvas::measureText() fingerprint deception does not appear to be working
How to Reproduce?
chrome://flags/#fingerprinting-canvas-measuretext-noise
Actual behaviour
The font fingerprint does not change
Expected behaviour
A new hash should be calculated on refresh
Relevant log output
No response
Additional context
No response