328810 - shutting down calendar doesn't wait for in-progress writes to finish

Status: RESOLVED FIXED

(Calendar :: Internal Components, defect)

Description

[Dan Mosedale (:dmosedale, :dmose)]

This was spun off of bug 218983.

Comment 1

[Joey Minta]

Attachment: proposal

This is how I see this bug possibly being fixed.  It's really tough to test and I'm not sure I understand all the nuances.  Another option is to piggy-back off the tryToClose window function in the main UI.  This would make it a bit easier to prompt to tell the user what's going on.  Feedback?

Comment 2

[Matthew (lilmatt) Willis]

(In reply to comment #1)
> Another option is to piggy-back off the tryToClose window function in the
> main UI.

Note that closing the main window does NOT quit the app on Mac, so that approach may not work across the board

Comment 3

[Joey Minta]

(In reply to comment #2)
> Note that closing the main window does NOT quit the app on Mac, so that
> approach may not work across the board
> 
There's still hiddenWindow.xul that we could use.

dmose and I wanted to get mconnor's opinion on this, as he knows this code better.

Comment 4

[Joey Minta]

mconnor points to: http://lxr.mozilla.org/mozilla/source/xpfe/components/startup/public/nsIAppStartup.idl#88

mail seems to use this quite extensively: http://lxr.mozilla.org/mozilla/source/mail/components/nsMailDefaultHandler.js#82

Comment 5

[Joey Minta]

Attachment: i will survive

Patch successfully keeps our app around until we finish the write.  Tested by delaying writing with a really big for-loop.

Comment 6

[Michiel van Leeuwen (email: mvl+moz@)]

Comment on attachment 229313 [details] [diff] [review]
i will survive

>                 try  {
>+                    appStartup.enterLastWindowClosingSurvivalArea();
>                 } catch (ex) {
>+                    appStartup.exitLastWindowClosingSurvivalArea();
>                 }

Looks like you want a finally block there, instead of putting the other call to exitLastWindowClosingSurvivalArea in a completly different place in the code.

Comment 7

[Joey Minta]

(In reply to comment #6)
> Looks like you want a finally block there, instead of putting the other call to
> exitLastWindowClosingSurvivalArea in a completly different place in the code.
> 
The other call is made after an observer is notified following an async call.  (after the write actually finished) |finally| has no way, that I know of, to handle that.

Comment 8

[Michiel van Leeuwen (email: mvl+moz@)]

Comment on attachment 229313 [details] [diff] [review]
i will survive

whoops, you are right. async code still manages to confuse me...
r=mvl

Comment 9

[Joey Minta]

patch checked in.