Closed yuccai closed 2 years ago
Why is randomization so problematic for Pica? And is it possible to have a workaround? Such as a warning, a non strict mode or a degraded result?
Because we don't control the browser the user is using 🤔
This is the function that detects anti fingerprinting randomization:
https://github.com/nodeca/pica/blob/34e82a5ddd0519ac7eb118889ae4f36360618d98/lib/utils.js#L139-L162
When anti fingerprinting randomization is detected by Pica, an error is thrown:
https://github.com/nodeca/pica/blob/34e82a5ddd0519ac7eb118889ae4f36360618d98/index.js#L653-L658
More about anti fingerprinting:
Here is nothing to fix. If FF fingerprint protection enabled, pica become completely unusable. And throws error as expected. It's programmer's responsibility to check for success of resize call, and decide what to do without resize.
We do the best of possible - provided descriptive error of failure.
We experienced a possible false positive of the fingerprinting detection. It was reported by a user on iOS with WKWebView.
However I dont see how that could happen based on the code above when not using fingerprinting detection.
Hello,
Currently, when pica detects that fingerprinting protection is enabled, it throws an error. Could it be possible to only throw a warning instead of blocking whole process ?