636110 - Update signature of duplicateTabIn() stub to match match Firefox 4.0.

Status: RESOLVED FIXED

(SeaMonkey :: Tabbed Browser, defect)

Description

[Philip Chee]

duplicateTabIn() is the Firefox equivalent of our OpenSessionHistoryIn(). After I implemented a stub duplicateTabIn() for extension compatibility, Dao decided that our API was saner. This bug just updates our stub.

Comment 1

[Philip Chee]

Attachment: Patch v1.0 inc SessionRestore fix.

>    duplicateTab: function sss_duplicateTab(aWindow, aTab, aDelta, aRelated) {
>      if (!aTab.ownerDocument || !aTab.ownerDocument.defaultView.__SSi ||
>          aWindow && !aWindow.getBrowser)
>        throw (Components.returnCode = Components.results.NS_ERROR_INVALID_ARG);
>  
>      var tabState = this._collectTabData(aTab, true);
>      var sourceWindow = aTab.ownerDocument.defaultView;
>      this._updateTextAndScrollDataForTab(sourceWindow, aTab.linkedBrowser, tabState, true);
> -    this._sendWindowStateEvent(aWindow, "Busy");
>      tabState.index += aDelta;
> +    tabState.index = Math.max(1, Math.min(tabState.index, tabState.entries.length));
>  
>      if (aWindow) {
> +      this._sendWindowStateEvent(aWindow, "Busy");

_sendWindowStateEvent falls over when aWindow is null. Introduced in:

http://hg.mozilla.org/comm-central/rev/c18dddea67de
Bug 633722 - Port Bug 615394 [Session Restore should notify when it is beginning and ending a restore]
Oops.

Comment 2

[Misak Khachatryan]

Comment on attachment 514710 [details] [diff] [review]
Patch v1.0 inc SessionRestore fix.

Perhaps we should modify it to use current window if aWindow is null ...

Comment 3

[[e-mail]]

(In reply to comment #2)
> Perhaps we should modify it to use current window if aWindow is null ...
No, because in that case we're creating a new window, and therefore restoreWindow will call _sendWindowStateEvent on the new window for us, right?

Incidentally I notice that undo close tab calls _sendWindowStateEvent even when using the tabbrowser to restore the tab.

Comment 4

[Philip Chee]

> > Perhaps we should modify it to use current window if aWindow is null ...
> No, because in that case we're creating a new window, and therefore
> restoreWindow will call _sendWindowStateEvent on the new window for us, right?
Also my reading of the code indicates that there should be a matching _sendWindowStateEvent(aWindow, "Ready") on the same window so it's no use sending only half of the notifications.

Comment 5

[Philip Chee]

Pushed to comm-central.
http://hg.mozilla.org/comm-central/rev/aa25d31c8221

Comment 6

[[e-mail]]

Comment on attachment 514710 [details] [diff] [review]
Patch v1.0 inc SessionRestore fix.

>-  OpenSessionHistoryIn(aWhere, delta, aTab)
>+  OpenSessionHistoryIn(aWhere, aDelta, aTab)
[Bah, I failed to notice the missing semicolon]

Comment 7

[Philip Chee]

> [Bah, I failed to notice the missing semicolon]

Added semicolon:
http://hg.mozilla.org/comm-central/rev/8310d0fd89a6