Comment 131 for bug 390508

Il 07/07/2010 12:02, Holger Berndt ha scritto:

> There's a "should", a recommendation. The spec does not demand that the
> expire timeout parameter is respected. (In fact, if it did, it would be
> a fishy spec - an implementation could just as well chose (or offer
> config parameters to let the user chose) to not display any non-critical
> messages at all, but just log them to a file. Timeouts of a few seconds
> don't make any sense in a log file.)

Is this the case? Is notify-osd specifically intended for file logging?
No, it isn't, so why is a good thing not to follow this recommendation?
After tons of messages still lacks a good answer to this question. And
no, "uniformity" is not a good answer.

> As pointed out before, the spec gives a clear recommendation as to what
> clients should expect: "Clients should try and avoid making assumptions
> about the presentation and abilities of the notification server."
> Obviously, you're not following that recommendation. That's your choice,
> and also your problem.

This has nothing to do with the design choices made in the server, it
doesn't explain why is good not to honor the clients timeout
(independently of the client expectations).