Re: Unbreaking the gnome clipboard

[Andrew Sobala <aes gnome org>]

On Sun, 2003-06-15 at 20:59, Philip Van Hoof wrote:
> On Sun, 2003-06-15 at 19:08, Mark Finlay wrote:
> > > I'd be more than happy to mentor you (or anyone else pursuing this) on
> > > the gnome-settings-daemon. I used to work on the control center so I
> > > have some experience with it. Let me know if you're interested.
> > 
> > A kind offer Rachel, hope someone takes you up on it :)
> > 
> > After looking more at gcm I'm starting to think that it would be much better 
> > to start from scratch. Gcm is very complicated, IMHO all we need in gnome is:
> 
> Well take a look at 
> 
> http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/gcm/Gcm.Net/Daemon/ClipboardManager.cs?rev=1.9
> 
> In C# its so easy that it even impressed me, the author of it. Yet, it
> works.
> 
> This daemon is everything you guys need. That is what you are asking
> for. If you run it .. it will do exactly what you want: maintain a copy
> of the clipboard for when the clipboard-owning application dies...
> 
> The only thing that needs to be added is the possibility to limit the
> amount of clipboards that you want to keep in memory... if you set that
> to "1" then this is 100% want you want.
> 
> The plans for Gcm.Net are to create a lightweight XML interface so that
> external applications can tell the manager-deamon to do things (like set
> a specific clipboard or give a preview of such a collected clipboard).
> Gcm.Net will also include end-user applications like a TrayIcon (or
> PanelApplet) and a Hotkey-menu (For example Ctrl+Shift+c will open a
> menu with all possible Clipboards to paste).
> 
> ps. Two other C# dudes have already showed interest in helping me
> getting this application working.
> 
> I have been thinking about writing the daemon in C because then the
> happy hippo people of gnome.org would not start wining about Mono as
> dependancy. Well, maybe thats an option for later. For now I will go for
> the C# version of this Clipboard Daemon. (Mainly because then I don't
> have to worry about pointers and free-ing memory and all that sh*t
> nobody really wants to care about)
> 

It does look nice and simple... _but_ if we're including it in GNOME we
would have to:

1) have a long conversation about Mono (which is still too immature
IMHO, because I'm a happy hippo) or
2) rewrite it in C :-)

> 
> > 1. The ability to press CTRL+C in a selection in an app, close the app, open another 
> > app, then press CTRL+V
> 
> 
> 
> > 2. The ability to make a selection, and use the middle mouse button to paste the selection
> > even after the selection is lost. ie. the middle mouse button pastes the last selection
> > instead of being limited to the current one.
> 
> 
> Well, you will have a VERY hard time getting that working :-). The
> mouse-selection a.k.a. the PRIMARY selection is not very flexible
> because selecting text happens to be one of the standard things a common
> user does all the time, even when he or she is not planning to do
> CopyPasting with it. If you steel the ownership of the PRIMARY... then
> right after one selects some text, it will get unselected. If you create
> a poller (this is what GCM does when you enable the mouse-selection in
> the preferences) then you will have to compare the last
> PRIMARY-clipboard with the current one to know if you have to collect it
> or not. I do this by comparing the clipboard-owner (wich does not take
> very much CPU) but then .. if you want to collect a new clipboard (from
> the PRIMARY) you first have to make another Widget the clipboard
> owner... 

Urgh. It sounds like we'll never have PRIMARY working the way I would
"expect" it to, then.

-- 
Andrew Sobala <aes gnome org>

"We made GNOME-VFS support smb: and nfs: URIs. And we made OOo support
GNOME-VFS. Booyakasha!" -- nat