Ed Lee :Mardak Reporter
	Description • 14 years ago

Instead of undo close tab opening any tab from any group (and adding the undo-closed-tab to the current group), the action should only look for tabs to undo-close of the current tab group.

	Raymond Lee [:raymondlee]
	Comment 1 • 14 years ago

Bug 587557 is related to this.  Does this bug contain the behaviour we want to implement?

	Ian Gilman [:iangilman]
	Comment 2 • 14 years ago

Right now, sessionstore keeps a list of the last 5 tabs closed (per window? or for the whole app?). It should instead keep track of the last 5 tabs closed per group. 

This will require sessionstore to know about Tab Candy groups, which will be a lot easier once I have a working version of the TabViewGroups jsm that I'm working for bug 578512. Let's hold off on this until then. 

We may need the help of someone familiar with sessionstore to do a part of this; CCing zpao, ehsan, and dietrich.

	Paul O'Shannessy [:zpao] (not bugmail, email directly)
	Comment 3 • 14 years ago

(In reply to comment #2)
> Right now, sessionstore keeps a list of the last 5 tabs closed (per window? or
> for the whole app?). It should instead keep track of the last 5 tabs closed per
> group. 

Per window. The number is based on the pref "browser.sessionstore.max_tabs_undo", which has a default value of 10.

	(no longer active)
	Comment 4 • 14 years ago

(In reply to comment #2)
> We may need the help of someone familiar with sessionstore to do a part of
> this; CCing zpao, ehsan, and dietrich.

What kind of help would you need?  Paul, can you help with this?

	Aza Raskin [:aza]
	Comment 5 • 14 years ago

For reference, the spec is available here: https://wiki.mozilla.org/Firefox/Projects/TabCandy/Design/UndoCloseGroup

	Paul O'Shannessy [:zpao] (not bugmail, email directly)
	Comment 6 • 14 years ago

(In reply to comment #4)
> (In reply to comment #2)
> > We may need the help of someone familiar with sessionstore to do a part of
> > this; CCing zpao, ehsan, and dietrich.
> 
> What kind of help would you need?  Paul, can you help with this?

Minimally right now. I've got other beta 5 blockers.

I think it would be best to just do this as part of (or after) bug 588217, otherwise it'd just be building another hack on.

The "hack" is to listen for TabClose, SS.getTabState(tab), associate with a group, reuse the max_tabs_undo pref for limiting, and then make tabbrowser stop using SS for data & tell it to look at tab candy instead.

	Michael Yoshitaka Erlewine [:mitcho]
	Comment 7 • 14 years ago

See also bug 566798, which describes an early behavior spec by Aza regarding this (and other) issues (I think).

	Aza Raskin [:aza]
	Comment 8 • 14 years ago

I'm punting this to the future as I think we can ship without it.

	bhearsum@mozilla.com (:bhearsum)
	Comment 9 • 14 years ago

(In reply to comment #8)
> I'm punting this to the future as I think we can ship without it.

Is there a different bug tracking "undo close tab opens tab in current tab group, rather than original", then? That's the current behaviour, and doesn't seem like something we should ship with, to me.

	bhearsum@mozilla.com (:bhearsum)
	Comment 10 • 14 years ago

(In reply to comment #9)
> (In reply to comment #8)
> > I'm punting this to the future as I think we can ship without it.
> 
> Is there a different bug tracking "undo close tab opens tab in current tab
> group, rather than original", then? That's the current behaviour, and doesn't
> seem like something we should ship with, to me.

I filed this in bug 607411.

	Ioana (away)
	Comment 11 • 12 years ago

Mozilla/5.0 (Windows NT 6.1; rv:14.0) Gecko/20100101 Firefox/14.0

Currently, when the user closes a tab from group1, then selects "Undo close tab" from group2, the tab is reopened in group2.

	Ioana (away)
	Comment 12 • 12 years ago

(In reply to Ioana Budnar [QA] from comment #11)
> Mozilla/5.0 (Windows NT 6.1; rv:14.0) Gecko/20100101 Firefox/14.0
> 
> Currently, when the user closes a tab from group1, then selects "Undo close
> tab" from group2, the tab is reopened in group2.

Practicly, bug 607411 has not been fixed (as specified in the bug itself). Are there any chances that it could get fixed?