Re: [evolution-patches] Mailer Context Menus Patch

[Rodney Dawes <dobey novell com>]

Hrmm. Well, I changed e-popup.c back. What exactly is a "sub-item" here?
With the current e-popup.c, it's trying to create separators in
submenus, that don't exist, so the code is aborting. I don't see any way
to make it not do this, outside of changing e-popup.c. Perhaps I don't
understand the code, or what you mean here, but it seems impossible to
do what we want with the context menus, in terms of hiding submenus when
right-clicking on the message itself, with the current e-popup.c code.

-- dobey

On Sat, 2005-01-29 at 07:57 +0800, Not Zed wrote:
> 
> No, you can't change e-popup.c, it is like that for a reason.  As the
> comment says, the flags need to be different for separators, you need
> to include all the bits of all the sub-items it bounds.  Otherwise
> there is no way to selectively turn on a bar only if any of the items
> inside it are used (the mask is not actually an 'enable visibility'
> mask, it is a 'disable visibility' mask).
> 
> This should be easy to fix with just some different bit-sets on the
> visibility mask of the BAR items.
> 
> Otherwise it looks much better, thanks.
> 
> 
> On Fri, 2005-01-28 at 16:12 -0500, Rodney Dawes wrote: 
> > Here is a new patch that moves everything to a single menu and uses
> > flags. I had to patch e-popup.c to get this to work properly, since
> > it was wanting to always create separators in the submenus for some
> > reason.
> > 
> > -- dobey
> > 
> > 
> > On Fri, 2005-01-28 at 11:10 +0800, Not Zed wrote:
> > > 
> > > Instead of using separate arguments for boolean variables to the popup
> > > function, you should add new flags to em-popup, and control it that
> > > way.  You shouldn't need to create three separate popup lists at all,
> > > it can all be controlled via the flags on the one list, and that gives
> > > the same capabilities to all plugins, otherwise they will not have
> > > access to these differentiators - particularly since the menu is given
> > > the same id in all cases.
> > > 
> >