588548 - opening a (single?) background tab, then navigating to it, then closing it should bring you back to where you started

Status: RESOLVED DUPLICATE of bug 533232

(Firefox :: Tabbed Browser, defect)

Description

[Joe Drew (not getting mail)]

I open a tab in the background off of the one I was just reading, and I switch to it. I finish reading whatever it was I wanted to read in that tab, then I close the tab, and it switches me to the tab right of the one I just closed. In my mind, I feel like the tab I go to is the one I switched from.

Right now this is not the case because we throw away tab parenting information on manual tab switch, on purpose. Certainly, should I switch to a tab not in the set of tabs just opened in the background, the parenting information should be thrown away; my thought is that maybe we should hold on to it if I'm reading a bunch of stuff in background tabs as an offshoot of reading the current tab, and I'll want to return to the current tab after reading the background tabs.

Sorry if this is hard to understand.

Comment 1

[Paul O'Shannessy [:zpao] (not bugmail, email directly)]

I know Frank was looking into this or something like it. Perhaps he has a bug to dupe this too (though things are a bit different with Tab Candy) or can at least add his insights.

Comment 3

[Mike Beltzner [:beltzner, not reading bugmail]]

Usually we use A b c c1 to represent four tabs, A (with focus), b, c, and c1 which was opened by c.

What Joe is saying is for a strip:

a b C d

if he forces a link from C to open in a new window:

a b C c1 d

then switches to it:

a b c C1 d

then closes it, he expects:

a b C d

but gets:

a b c D

For the case where there is a single c1 opened by C, we should return to the opener. If there are multiple c1 c2 c3 this would not apply (as it's harder to know the right thing to do) and if focus goes from C to d or a, then this would not apply.

Comment 4

[Alex Limi (:limi) — Firefox UX Team]

(In reply to comment #1)
> I know Frank was looking into this or something like it.

Yes, he has a good spec for what the order should be in bug 578327 (check out the link to the wiki in "related links")