255332 - Warning: reference to undefined property Components.classes['@mozilla.org/winhooks;1']

Status: RESOLVED WONTFIX

(SeaMonkey :: General, defect)

Description

[timeless]

Warning: reference to undefined property
Components.classes['@mozilla.org/winhooks;1']
Source File: chrome://communicator/content/pref/platformPrefOverlay.xul
Line: 98

Steps:
1. don't build winhooks
2. run mozilla
3. open prefs

The code should use a try block, it could also be written:
if (("nsIWindowsHooks" in Components.interfaces) &&
    ("@mozilla.org/winhooks;1" in Components.classes))
try {
var winhooksService =
Components.classes["@mozilla.org/winhooks;1"].getService(Components.interfaces.nsIWindowsHooks);
parent.isTurboEnabled = winhooksService.isOptionEnabled("-turbo");
document.getElementById("enableTurbo").checked = parent.isTurboEnabled;
} catch (e) {};

Each instance of code that checks for winhooks should be audited.

xpfe/browser/resources/content/navigator.js is ok
xpfe/communicator/resources/content/nsContextMenu.js is not ok
xpfe/components/prefwindow/resources/content/pref-winhooks.js is silly (it null
checks something which will not be null - an exception is throw instead...)
xpfe/components/prefwindow/resources/content/pref-advanced.xul is not ok
xpfe/components/winhooks/nsSetDefaultBrowser.js does not sound quite right

Comment 1

[Frank Wein [:mcsmurf]]

Attachment: Patch, try #1

Comment 2

[Frank Wein [:mcsmurf]]

Attachment: Patch try #2

Comment 3

[[e-mail]]

Comment on attachment 155938 [details] [diff] [review]
Patch try #2

>-        var isWin = navigator.appVersion.indexOf("Windows") != -1;
>+        var isWin = (("nsIWindowsHooks" in Components.interfaces) && ("@mozilla.org/winhooks;1" in Components.classes));
Surely even timeless can't lose his windows hooks in between showing the
menuitem and clicking on it?
>-      var winhooks = Components.classes[ "@mozilla.org/winhooks;1" ].
>-                       getService(Components.interfaces.nsIWindowsHooks);
>+      if (("nsIWindowsHooks" in Components.interfaces) &&
>+          ("@mozilla.org/winhooks;1" in Components.classes)) {

>-      if (navigator.platform != "Win32") {
>+      if (navigator.platform != "Win32" || !(("nsIWindowsHooks" in Components.interfaces) &&
>+         ("@mozilla.org/winhooks;1" in Components.classes))) {
You don't need to test for Win32 here; other platforms won't have the winhooks
component.

>-      var winhooksService = Components.classes["@mozilla.org/winhooks;1"].getService(Components.interfaces.nsIWindowsHooks);
>-      if (winhooksService) {
>-        parent.isTurboEnabled = winhooksService.isOptionEnabled("-turbo");
>-        document.getElementById("enableTurbo").checked = parent.isTurboEnabled;
>+      if (("nsIWindowsHooks" in Components.interfaces) &&
>+          ("@mozilla.org/winhooks;1" in Components.classes)) {
How is timeless going to lose his windows hooks now? By stepping through with
Venkman?

>+      if (("nsIWindowsHooks" in Components.interfaces) &&
>+          ("@mozilla.org/winhooks;1" in Components.classes)) {
How does timeless manage to install winhooks/nsSetDefaultBrowser.js without the
rest of winhooks?

sr=me on what's left of the patch ;-)

Comment 4

[Frank Wein [:mcsmurf]]

ok, someone else can work on a patch ;-) (really), the problems of timeless are
"silly" :)

Comment 5

[Serge Gautherie (:sgautherie)]

Filter "spam" on "guifeatures-nobody-20080610".

Comment 6

[Frank Wein [:mcsmurf]]

Will get fixed by new shell service ui.

Comment 7

[Serge Gautherie (:sgautherie)]

I assume this is "R.Fixed" by bug 441050 and bug 364168.

Frank (, Neil), can you verify this resolution ?
And that there was nothing in the patch that would still be wanted.

Comment 8

[[e-mail]]

gerv said wontfix is better.

Comment 9

[[e-mail]]

In theory you could morph this bug into the shell service equivalent - currently nsContextMenu assumes that the shell service exists on and only on Windows - but I think we're better off with a new bug.