|
From: | Martin Pala |
Subject: | Re: The checksum statement ++ |
Date: | Fri, 15 Aug 2003 21:20:41 +0200 |
User-agent: | Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030714 Debian/1.4-2 |
Jan-Henrik Haukeland wrote:
It was exactly my first though :) Then i figured out, that probably 'timeout' event is simplier and will do the same work.Martin Pala <address@hidden> writes:And what about the first proposal, send today 12:43 (generalization of timeout statement and timeout event hard error classification + its broadcasting through dependency tree)?It's not bad but I feel it's a bit of a workaround to connect it to timeout. But as you mentioned having an event (especially checksum) propagate upwards in the dependency tree is important. Maybe using a new STOP_MONITORING event or something is better (cleaner)? I don't know.
Yes, it is possible, no problem. This could involve lot of changes and it is not critical for 4.0 => we don't need to implement it now. Though i preffer to solve it in 4.0 - it is more logical then present behavior + it can involve (small) language change and it could be better to do it at once. I can implement it as soon as i'll be back. Please choose what you like better :)I need to investigate your proposal and the control.c code more to see it better. But right now, I think we could doucment us out of the problem and maybe refactor the code in a later 4.1 version. What do you think?
[Prev in Thread] | Current Thread | [Next in Thread] |