Related to bug 933933, Mike?

I don't think so - I think this is just us not drawing that separator when using Lion fullscreen (which I guess we detect with the inFullscreen attribute...why that's distinct from sizemode="fullscreen" is not something I'm sure of.) This patch draws that separator, and also gives the navigator-toolbox a titlebar texture when in fullscreen mode. Also removes unnecessary padding from the top of the window when in fullscreen with a lw-theme. How do you feel reviewing this one, Gijs?

Comment on attachment 8334232 [details] [diff] [review] Patch v1 Review of attachment 8334232 [details] [diff] [review]: ----------------------------------------------------------------- ::: browser/themes/osx/browser.css @@ +2559,5 @@ > /* > * Draw the bottom border of the tabstrip when core doesn't do it for us: > */ > +#main-window:-moz-any([privatebrowsingmode=temporary],[sizemode="fullscreen"], > + [inFullscreen],[customizing],[customize-exiting]) #TabsToolbar::after, Somewhat orthogonal: Is it intentional that "[customize-exiting]" is left out a few blocks above this one? It seems like that could also cause a flicker of the border.

Comment on attachment 8334232 [details] [diff] [review] Patch v1 Review of attachment 8334232 [details] [diff] [review]: ----------------------------------------------------------------- ::: browser/themes/osx/browser.css @@ -3581,5 @@ > - /* This additional padding matches the change in height in the pseudo-element > - * above. The rules combined force the top 22px of the background image to > - * be hidden, so there image doesn't jump around with the loss of the titlebar */ > - padding-top: 11px; > - background-position: right -11px; (In reply to Mike Conley (:mconley) from comment #2) > Also removes unnecessary padding > from the top of the window when in fullscreen with a lw-theme. That's actually the opposite of what I think we should do as it's not very usable with the tabs at the top of the screen with the system menubar popping in all of the time. See bug 839888 where this should have been made consistent with the non-LWT case. We also have bug 878436 to update these offsets.

This is pretty subtle.