gnustep-dev
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: DirectFB backend, I'm going for it!


From: Fred Kiefer
Subject: Re: DirectFB backend, I'm going for it!
Date: Mon, 12 Jan 2004 15:04:41 +0100
User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030821

Alexander Malmberg wrote:
Chad Hardin wrote:
Speaking of a new deamon, last time I looked,  [NSWorkspace
-launchedApplications] only returned the applications which were
launched *after* the calling app waas launched.  (It's supposed to
return all launched applications).  This looks like daemon material.


Again, this depends on the backend/'desktop', and can't be solved in
-gui alone. It's something of an unresolved architecture problem in
-gui. I've been talking about a solution (or at least resolution :) in
#GNUstep for a while. I'll write something for the list.


I don't know, what you did suggest on #GNUstep as I never go there, but some months ago I did suggest here on this list to implement this feature based on the information gdomap already has about the running applications. That way GSServiceManager would ask gdomap for the names of all registered objects and could than get a handle onto them. If an object is still active and also represents a GSServiceManager, this is a connection to a running application. All of this could be packaged up in a method on GSServiceManager [-(NSArray*)runningApplications]. This than could be used for further operations, eg to get the data for launchedApplications or to implement hide/unhiding of other applications.

The benefit of this solution is, that we would not need another GNUstep deamon or any additional protocol, just a few extensions on GSServiceManager. This also doesn't require an extension of gdomap, we just would reuse the code from there, that handles the -N parameter (Although a higher level interface would of course be welcome. NSSocketPortNameServer springs to mind). On a system with a lot of none GUI applications running this implementation may be a bit slower than a dedicated service for this task, but I doubt, that anyone would notice this difference in a typical case.





reply via email to

[Prev in Thread] Current Thread [Next in Thread]