Re: Trunk emacs infelicity with linum mode

[Stefan Monnier]

>> This brings a new concept to me: how can a face be invalid in a specific
>> frame?
> I think it's because the face was not yet realized for that frame.
> Before we create a new frame, we realize its basic faces, but not the
> rest.  I think.

Yes, I kind of guessed this part, but it would be good to try and not
expose those details to Elisp.  IOW I think it's a bug that Elisp gets
to know about "not yet realized faces".

I guess in the present case, the underlying problem is that we run
window-configuration-change-hook too early.


        Stefan