1651062 - File icon used for reader mode (about:reader)

Status: NEW

(Firefox :: Site Identity, defect, P3)

Description

[Tom S [:evilpie]]

Bug 1570678 replaced the neutral (i) info icon with a file icon when activating the reader mode even for http(s) pages. This seems wrong as the content of the page is not actually a file. Maybe the icon should just be whatever the page uses without reader mode, e.g. a lock?

Comment 1

[:Gijs (he/him)]

Dupe of bug 1165592?

Comment 2

[:Gijs (he/him)]

wtf bugzilla, I did not close this as fixed...

Comment 3

[(inactive) Nihanth Subramanya [:nhnt11]]

I don't think it's a dupe, but thanks for pointing out that bug. It seems like reader mode was never getting "proper" treatment by the identity UI, but the file icon is still a regression (we used the "i" icon before bug 1570678).

I think we should try and get things back to how they were so that the issue in bug 1165592 is self-contained.

Comment 4

[(inactive) Nihanth Subramanya [:nhnt11]]

Mmmmmm, this is an interesting one actually, maybe the file icon treatment is not worse than what we had before....

I need to think about this. I'm now wondering if it's easier to just take the opportunity to give about:reader proper treatment in the identity UI than to un-regress.

Comment 5

[BugBot [:suhaib / :marco/ :calixte]]

Set release status flags based on info from the regressing bug 1570678

Comment 6

[Kim Moir [:kmoir] ET]

nhnt11: Triaging as REO for 79, could you set a priority and severity for this bug?

Comment 8

[Tom S [:evilpie]]

This is still reproducible. When clicking the icon it says "This page is stored on your computer".

Comment 9

[:Gijs (he/him)]

(In reply to Tom S [:evilpie] from comment #8)

This is still reproducible. When clicking the icon it says "This page is stored on your computer".

Yes, fixing it is non-trivial, and the impact seems... really limited. See also bug 1165592. The identity panel doesn't have access to "whatever the page uses without reader mode" anymore when reader mode loads - and there might or might not have been a page loaded before (the alternative is reader mode uses an XHR request to fetch the article contents). Reusing the identity state from the previously-loaded page (if one exists) may also be wrong in the case of e.g. mixed content which may or may not be present on the reader-ized version (e.g. ads).

The actual page itself is stored on your computer, but all the content is loaded from somewhere else. If an http page did the same thing, we'd also show the lock-with-strike-through icon, even if almost all the content travelled over https. So like, yes, in an ideal world we'd give users better information here, but getting to that world is pretty difficult, and I don't think that the current state is so awful that we absolutely must prioritize the significant amount of work it'd be to fix this properly.