Comment 1 neil@parkwaycc.co.uk 2012-07-31 14:13:03 PDT

Created attachment 647678 [details] [diff] [review]
Proposed patch

Comment 2 Ian Neal 2012-08-04 06:06:03 PDT

Comment on attachment 647678 [details] [diff] [review]
Proposed patch

>   // build the file name
>-  nsCAutoString filePath(PR_GetEnv("HOME"));
>+  nsCString filePath(PR_GetEnv("HOME"));
Part of another patch?

> NS_IMETHODIMP
> nsGNOMEShellService::GetDesktopBackgroundColor(PRUint32 *aColor)
> {

>+  if (background.IsEmpty())
>+    return NS_ERROR_FAILURE;
Firefox does *aColor = 0; return NS_OK here - should we be doing the same?

Comment 3 neil@parkwaycc.co.uk 2012-08-04 09:49:16 PDT

(In reply to Ian Neal from comment #2)
> (From update of attachment 647678 [details] [diff] [review])
> >-  nsCAutoString filePath(PR_GetEnv("HOME"));
> >+  nsCString filePath(PR_GetEnv("HOME"));
> Part of another patch?
No, just random cleanup (nsStringAPI typdefs nsCAutoString to nsCString, and this also helps us avoid the mass nsCAutoString -> nsAutoCString change that has been bandied about).

> >+  if (background.IsEmpty())
> >+    return NS_ERROR_FAILURE;
> Firefox does *aColor = 0; return NS_OK here - should we be doing the same?
The idea is that this will cause us to hide the colour picker, so that we don't try to change the background colour, since it seems likely that it would fail too.

Comment 4 Philip Chee 2012-08-05 09:50:56 PDT

Pushed to comm-central:
http://hg.mozilla.org/comm-central/rev/872744c69ef8