Re: Will Gnome _EVER_ be stable?
- From: "Stuart D. Gathman" <stuart bmsi com>
- To: "Brandon S. Allbery KF8NH" <allbery ece cmu edu>
- Cc: Henry Katz <henry katz iscs-i com>, Gnome Mailing List <gnome-list gnome org>
- Subject: Re: Will Gnome _EVER_ be stable?
- Date: Fri, 22 Mar 2002 09:51:31 -0500 (EST)
On 21 Mar 2002, Brandon S. Allbery KF8NH wrote:
> On Thu, 2002-03-21 at 20:45, Henry Katz wrote:
> Actually, it's more to do with the freakish way some settings belong to
> the application, some to the session manager, and some to the window
> manager, and the nasty synchronization issues that result.
>
> (Could this be solved by having gconfd pass appropriate settings to the
> session mgr and/or window mgr, thus reducing the problem from "all apps
> must deal with nasty synchronization issues" to "gconfd must deal with
> synchronization"?)
I am a Gnome user and haven't done any Gnome programming, so I apologize
if this is a "duh", but in other programming projects, I have found that
a library is often better than a daemon for things like this.
Doesn't gtk+ have a configuration framework that takes care of stuff
like this?
An old article titled "Worse is Better" described how it was very tricky
to save kernel state across task switches when implementing a simple
thing like 'read()'. Unix punted and simply returns a "Try Again" error
code to let the application recover. What was not mentioned is that
the "application" needn't deal with this. A wrapper around 'read()' can
check for this error and retry as needed. All state is saved on the
user mode stack. While it is easy to invent, too bad such a wrapper wasn't
standardized (or did it miss it?).
--
Stuart D. Gathman <stuart bmsi com>
Business Management Systems Inc. Phone: 703 591-0911 Fax: 703 591-6154
"Confutatis maledictis, flamis acribus addictis" - background song for
a Microsoft sponsored "Where do you want to go from here?" commercial.
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]