All the privacy.clearOnShutdown.* used on Desktop doesn't work on mobile(Currently, we are violating the disk avoidance property).
On mobile we need to use the privacy.clear flag with the following attributes:
And the code that uses it. Android just updates the prefs when the user manually clicks in the "Set" button. We can use the distribution preferences to forcefully update them.
However, we have other problem. If the user closes the browser without clicking in the quit button in the TBA menu, the sanitize method in the browser.js is not called and the TBA doesn't delete any session information.
This may be a silly question, but can you simply refactor the quitAndClear() method into a quit() and clearPrivateData()? And then call clearPrivateData() in onCreate()?
quitAndClear and the clear private data logic are different because we are calling different events Browser:Quit and Sanitize:ClearData, respectively. That said, I created a new java method that builds the arguments used by both methods.
This may be a silly question, but can you simply refactor the quitAndClear() method into a quit() and clearPrivateData()? And then call clearPrivateData() in onCreate()?
Hm, for some reason the preferences.json file is not visible in the .apk. I have the same issue with #27111 (moved) but I am quite sure that it at some point got included (I even have a .apk lying around that shows this). I need to debug this first before making progress on this bug...
Hm, for some reason the preferences.json file is not visible in the .apk. I have the same issue with #27111 (moved) but I am quite sure that it at some point got included (I even have a .apk lying around that shows this). I need to debug this first before making progress on this bug...
I guess setting TB_BUILD_WITH_DISTRIBUTION fell through the cracks while transitioning to tor-browser-build...
I pushed a fixup commit to tor-browser-builds master branch that cases care of the TB_BUILD_WITH_DISTRIBUTION issue (commit f38b27734129d7a83f068e09ea1a1d07327622f5).
Looks okay to me. I pushed the patches to tor-browser-60.3.0esr-8.5-1 (commits b03bfc5100f3d255ae0116ab18653759f87fe241, 57ef4c8201764f36e50347e913825c99ba2dc8bd, and 977e468c7f83b78cf920a7aa4943147eda195c45).
I agree with sysrqb that we should file an upstream bug for that and get it properly fixed. igt0 could you do that?
I am not sure yet what "ideally" exactly means in this case but I feel looking at the other Tor Browser platforms something like when I close the session my private data is gone sounds like a good idea. What is Fennec's PBM on Android doing here?
Trac: Status: needs_review to closed Resolution: N/Ato fixed