Navigation History Deletion on iOS |
|||||
Issue descriptionUsers are often surprised that after deleting their browsing history, a list of visited sites remains in the navigation history of a tab and pressing the back button resurfaces visited websites. This issue was fixed on Android and Desktop by propagating history deletion to tabs. The same should be done on iOS. Related bug for Android and Desktop: https://crbug.com/407074
,
Aug 2
Sorry, I'm currently not working on iOS issues
,
Sep 27
This seems to be a Navigation History Issue not related to the History UI (Which is cleared after CBD) eugenbut@ WDYT?
,
Sep 27
This is not related to history UI, but rather to clearing browsing data. https://crrev.com/796851 fixed the bug for Desktop and Android, but nothing was done for iOS. Martin, is this something that your team can take?
,
Sep 27
Clearing browser data in iOS today involves deleting the WKWebView. With go/bling-navigation-experiment, I actually have to add special logic to make sure navigation history of the tab is NOT deleted. This is a source of complexity and bugs because it uses the client-side restore logic. If user actually wants to see the navigation history cleared, we can delete the special logic and it'll be an overall tech debt win for iOS.
,
Sep 27
I think that "Cached Images and Files" deletion also deletes WKWebView. But clearing cashes should not remove back-forward navigation history.
,
Sep 27
Ah OK. If we can differentiate between these two cases in Bling's code and only restore history in clearing cache case, that would be nice.
,
Sep 27
We could move cookie clearing code to ios/web layer ( crbug.com/884341 ). In that case ios/web could differentiate between different types of cleared data.
,
Sep 27
So to summarize my understanding: WKWebView already automatically deletes navigation history when some datatypes are deleted. We have a code in Chrome which restores it. We want to remove that code from the history deletion codepath. We don't want to remove that code from cache deletion codepath. Is that correct? Re #4, I'm happy to put this on our roadmap, given that our team owned this on Desktop and Android too.
,
Sep 27
,
Sep 27
Comment #9 sounds like a great summary. If we follow this approach we will diverge from /content implementation, but that's probably fine given that solution will be much cleaned for iOS. |
|||||
►
Sign in to add a comment |
|||||
Comment 1 by benhenry@chromium.org
, Aug 1