Programs should exit gracefully on session end
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
One Hundred Papercuts |
Won't Fix
|
Undecided
|
Unassigned | ||
Ubuntu |
Invalid
|
Undecided
|
Unassigned |
Bug Description
This one might be a bit tenuous, but bear with me.
When the user requests Gnome to shut down or log out, the X session is immediately TERMinated, along with its contents. This is what all window managers have done for eons, but IMO the user experience would be vastly improved if instead, programs were given the option of exiting gracefully, and requesting user input if appropriate. This behaviour would then be similar to that found on other desktop operating systems.
A few use cases to demonstrate what I mean:
1) A user logs in and starts upFirefox as part of their standard workflow. When done, they log out. The next day, they log in again, and start up Firefox.
What you expect: Firefox asks the user what it should do before logging out on day one, or restores tabs automatically (depending on user settings).
What actually happens: Firefox immediately exits on day one, and on day two recovers tabs as if it had crashed, informing the user that "a crash occurred".
2) A user is working on a document in OpenOffice, and then switches workspace to browse the internet before saving. When done with browsing, the user chooses to shut down, forgetting that they have an unsaved document open on another workspace.
What you expect: OpenOffice should ask the user whether they'd like to save the document or discard changes, much like it does when pressing the close button or hitting Alt+F4. Since the workspace has changed, this request should either appear on the active workspace, or the user should be switched to OpenOffice's workspace (perhaps even have all workspaces collapse to a single workspace).
What actually happens: OpenOffice immediately exits. The next time OpenOffice is opened, the user has to walk through the recovery process to get their document back. Potentially worse, the user sends the document in an email or copies it to another medium, unaware that the file does not contain the newest changes.
3) The user starts a large file transfer to a USB stick, then absent-mindedly logs out instead of locking the screen before the file transfer is complete.
What you expect: Either a) the user is informed that a file transfer is in progress, and is given the option of stopping the transfer or interrupting the log out procedure or b) the file transfer is gracefully interrupted, without leaving potentially corrupted files on either file system
What actually happens: the file transfer is immediately interrupted, with a potential for corruption on the destination media.
Changed in ayatana: | |
status: | New → Invalid |
tags: | added: ayatana |
affects: | dead-ayatana → libindicator |
This may or may not be something small to fix, but definitely worth fixing