Re: galf once more: application launch feedback

[Peter Åstrand <astrand lysator liu se>]

> I do appreciate that using X is a more lightweight solution, but the stumbling
> block in the design of a feedback mechanism is the need to cater for
> applications using non-compliant toolkits.
> In your proposal, if I understand it correctly, you are relying on GTK+ to allow
> the application to signal the feedback server that it has started.  This
> wouldn't allow applications using other toolkits to signal the server.  You
> could, I suppose, use a default timeout for such applications but that doesn't
> really address the problem.

No, unless I have forgotten something, my proposal should work with other
toolkits than GTK+, and with non-compliant TK:s.

For non-compliant toolkits: The binary wrapper is the component to figure
out whether the toolkit is compliant or not (eg. if mappingmode should be
used or not). It can do this by looking at desktop-files and/or by using
command line options (like --mappingmode in Xalf). It may also be possible
to use ldd to check if the application is linked against a compliant
toolkit, but this is probably a too ugly solution. Anyway, the wrappers
decision is included in the message it sends to the feedback server. This
is no different from a CORBA solution. As I said, changing the method by
which to send messages does not have to affect the contents of the
messages.

Regarding other toolkits than GTK+: XClientMessageEvents is a very general
mechanism and sending a message with it could be done by other toolkits.
As well as setting an X window property, which was your proposal.


-- 
/Peter Åstrand <astrand lysator liu se>