Privacy: investigate privacy implications of HTML canvas usage

[jayaddison]

Describe the bug
During testing of #1 in Firefox 140.8.0 with the fingerprint-resistance feature enabled, noise is introduced to the screenshots captured by this component.

We receive some valid screenshots to our internal (Google-hosted) corporate infrastructure where we read feedback.

We definitely have some test cases where we've captured a test screenshot in production to check that the feedback and screenshot functionality are still working; let's use this data as part of an investigation into whether the screenshot feature compromises user privacy.

To Reproduce
N/A - research/investigation required

Expected behavior
Determine whether the screenshot-capture functionality included in this component causes a compromise of privacy for users who include screenshots in their feedback.

Screenshots

[jayaddison]

Attached at the end of this comment is a screenshot received during testing of the feedback feature in production.

• If this does not contain significant fingerprinting information, then that is unfortunately not sufficient to ensure that the feedback mechanism cannot be used to fingerprint users -- we'd need to test other browsers/GPUs/etc in addition to gain confidence.
• If it does contain fingerprinting information, then that is sufficient to determine that there is definitely a problem to resolve here.

I wasn't initially able to determine much from the image -- other than screen resolution and colour depth; I do admit that those are factors in device fingerprinting.

[jayaddison]

I think that it may be worth adding a note/message to this component along the lines of Note: submitting a screenshot will include some information about your computing environment - for example, display size and colour - in your feedback., to inform users about the information disclosure.