Reviews for CanvasBlocker
CanvasBlocker by kkapsner
Response by kkapsner
Developer response
posted 5 years agoThe hashes change for me if I hit "Re-test your browser". Depending on the CB settings the hashes may stay the same (Random number generator). Please open an issue at https://github.com/kkapsner/CanvasBlocker/issues with your settings. Then we investigate the problem further.
Edit in response to your edit: if the hash does not change for you with CanvasBlocker it might be a bug and I want to investigate it further. But this review page is a very bad communication channel for such things (e.g. I did not get any notification that you edited your review and you have no way to actually respond). That's why you should create an issue at Github. It's not about "shuffling people off". It's about better communication and actually solving the problem and/or getting better insight.
The EFF investigation is 10 years old (CB is "only" 5 years old) and I think some improvements were made in the meantime. I can only repeat myself: on https://panopticlick.eff.org/ I get different hashes when I click "Re-test your browser".
I have two ideas why you get the same hash:
1. you have privacy.resistFingerprinting enabled which also protects canvas (see https://github.com/kkapsner/CanvasBlocker/issues/158 and https://github.com/ghacksuserjs/ghacks-user.js/issues/767 for further information)
2. you have the persistent random number generator enabled (stealth preset)
In response to your edit2: it's OK if you do not want to report a Github. At the moment I have no issue open at Github were the protection is not working - I try to solve them as quick as possible. Sometimes it's a CB bug and sometimes it's a misconfiguration or misunderstanding. If you know of other persons with failed protection maybe they want to help me to solve this issue. Without the complete details of the system and some sort of reproduction scenario I have no way of knowing what is going wrong.
I do not know which bad experiences you had with Github but it's simply a development platform and the kind of responses/interaction can vary very much between repositories. It all depends on people.
I had several issue with exactly the same symptom (hash not changing - like https://github.com/kkapsner/CanvasBlocker/issues/199 with Panopticlick or the most recent one being https://github.com/kkapsner/CanvasBlocker/issues/425). All of them were resolved or I had do close them because I did not get an answer for my following up questions that I have to prompt to solve the issue.
I do not say that Panopticlick does not maintain accurate results. I simply say that some people try to solve the problem in the last 10 years and made progress (the most know is the TOR Browser and the Firefox uplift privacy.resistFingerprinting). I know that CanvasBlocker is not perfect and there will be other fingerprinting techniques and attack vectors in the future that CB does not cover at the moment. But at the moment (and especially with the Canvas hashes on Panopticlick) it's working fine.
The "random number generator" is a setting withing the CanvasBlocker settings - I could show you a screenshot of where to find it if we were on Github...
It is set to "persistent" if you selected the "stealth" preset upon installation.
I do not see a point why Mozilla would want to buy my add-on. If they would want to incorporate it into their browser they simply could to it as it's open source and the licence allows the usage for non commercial usage (and they do charge money for Firefox). There are loads of good add-ons that work and that are not integrated into Firefox. That's the idea and beauty of Firefox: you can customize it.
Edit in response to your edit: if the hash does not change for you with CanvasBlocker it might be a bug and I want to investigate it further. But this review page is a very bad communication channel for such things (e.g. I did not get any notification that you edited your review and you have no way to actually respond). That's why you should create an issue at Github. It's not about "shuffling people off". It's about better communication and actually solving the problem and/or getting better insight.
The EFF investigation is 10 years old (CB is "only" 5 years old) and I think some improvements were made in the meantime. I can only repeat myself: on https://panopticlick.eff.org/ I get different hashes when I click "Re-test your browser".
I have two ideas why you get the same hash:
1. you have privacy.resistFingerprinting enabled which also protects canvas (see https://github.com/kkapsner/CanvasBlocker/issues/158 and https://github.com/ghacksuserjs/ghacks-user.js/issues/767 for further information)
2. you have the persistent random number generator enabled (stealth preset)
In response to your edit2: it's OK if you do not want to report a Github. At the moment I have no issue open at Github were the protection is not working - I try to solve them as quick as possible. Sometimes it's a CB bug and sometimes it's a misconfiguration or misunderstanding. If you know of other persons with failed protection maybe they want to help me to solve this issue. Without the complete details of the system and some sort of reproduction scenario I have no way of knowing what is going wrong.
I do not know which bad experiences you had with Github but it's simply a development platform and the kind of responses/interaction can vary very much between repositories. It all depends on people.
I had several issue with exactly the same symptom (hash not changing - like https://github.com/kkapsner/CanvasBlocker/issues/199 with Panopticlick or the most recent one being https://github.com/kkapsner/CanvasBlocker/issues/425). All of them were resolved or I had do close them because I did not get an answer for my following up questions that I have to prompt to solve the issue.
I do not say that Panopticlick does not maintain accurate results. I simply say that some people try to solve the problem in the last 10 years and made progress (the most know is the TOR Browser and the Firefox uplift privacy.resistFingerprinting). I know that CanvasBlocker is not perfect and there will be other fingerprinting techniques and attack vectors in the future that CB does not cover at the moment. But at the moment (and especially with the Canvas hashes on Panopticlick) it's working fine.
The "random number generator" is a setting withing the CanvasBlocker settings - I could show you a screenshot of where to find it if we were on Github...
It is set to "persistent" if you selected the "stealth" preset upon installation.
I do not see a point why Mozilla would want to buy my add-on. If they would want to incorporate it into their browser they simply could to it as it's open source and the licence allows the usage for non commercial usage (and they do charge money for Firefox). There are loads of good add-ons that work and that are not integrated into Firefox. That's the idea and beauty of Firefox: you can customize it.