On Sat, Mar 01, 2008 at 05:35:52AM +0100, Krzysztof Oledzki wrote:
> On Fri, 29 Feb 2008, Willy Tarreau wrote:
> >>BTW: there is one more unhandled patch: "Prevent redispatcher from
> >>selecting the same server, version #3". ;)
> >OK, I had some confusion initially about this one because I thought it
> >changed the behaviour to what you described in the second part of the
> >changelog, which it apparently does not. I'll merge it.
> There are two different patches I'm afraid you have possibly mistaken:
> - "Don't increment server connections too much + fix retries"
> - "Prevent redispatcher from selecting the same server, version #3"
both applied now.
> The "second part of the changelog" comment applies to the fist one, not
> the one I mentioned above. Sorry for the confusion, it seems I send too
> many patches on once.
I think that what tends to confuse me is when there is a mix of real patches and RFCs, because sometimes I overlook a patch thinking it's an RFC or that it depends on an RFC. Since I sometimes remain a few days not reviewing patches or some mails, I can quickly get out of context.
So the clearer the distinction between patches and RFCs and the more testing you run on your patch, the less the risk of a patch being confused.
What can also help is when you send a series of patches which depend one on another is simply to number them (git send-email may even help you run the mass mailing). That way, if we have to discuss one of them, it'll get easier to resend the full series from this one and ignore the old one.
willy Received on 2008/03/04 06:28
This archive was generated by hypermail 2.2.0 : 2008/03/04 06:30 CET