Date: Mon, 23 Feb 1998 10:46:05 +1030 From: Greg Lehey <grog@lemis.com> To: Chuck Robey <chuckr@glue.umd.edu> Cc: "Jordan K. Hubbard" <jkh@time.cdrom.com>, Nate Williams <nate@mt.sri.com>, Terry Lambert <tlambert@primenet.com>, current@FreeBSD.ORG Subject: Re: More breakage in -current as a result of header frobbing. Message-ID: <19980223104605.49121@freebie.lemis.com> In-Reply-To: <Pine.BSF.3.96.980222174223.14406Y-100000@localhost>; from Chuck Robey on Sun, Feb 22, 1998 at 05:50:31PM -0500 References: <19980223084422.20994@freebie.lemis.com> <Pine.BSF.3.96.980222174223.14406Y-100000@localhost>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, 22 February 1998 at 17:50:31 -0500, Chuck Robey wrote: > On Mon, 23 Feb 1998, Greg Lehey wrote: > >> On Sun, 22 February 1998 at 10:31:35 -0800, Jordan K. Hubbard wrote: >>> Nate Williams wrote: >>>> We can't have it both ways. Either we have an *enforced* policy of >>>> actually punishing offenders (which may offend them and cause them to >>>> run away), or we don't have any policy at all and so claiming to have >>> >>> I'd rather not have a policy based on punishment. We're supposed to >>> be doing this for fun and the joy of making a good OS, and hitting >>> people with sticks is not conducive to that. >> >> Agreed. But a policy based on punishement isn't the same thing as a >> policy which includes punishment in exceptional circumstances. > > I think Jordan and Greg both have points they are correct about. What > about making a web page that the offender gets put on ... nothing > incredibly horrible, just a rogues gallery (I broke current on 00/00/00 > with a commit to " ... ". You figure that the list is pretty quickly > going to get some names on it, so it won't be so embarrassing as to turn > people off, but most people wouldn't want to show up on it, so they'd > avoid doing things to get themselves so honored. > > Folks like John Dyson would show up a lot ... most of us know why that'd > be, and I _wouldn't_ want to unduly embarrass John ... but if John could > take it as sort of a left-handed compliment, it'd probably be a real > deterrent to most of the rest of us. Think of it as minimal punishment. > If it turns out to be too negative, a little editing of the web page could > either soften (or harden) the impact, as needed. Hmmm. I think this is a little drastic. I certainly wouldn't recommend anything unless the person in question is a repeat offender. And in John's case, as I said before, I don't think he has offended. There's a difference between committing a file that won't build and committing a change that causes the VM system to die in certain very specific circumstances. > I don't want to be as drastic as I think Greg is proposing, but I think > ignoring it isn't probably the best way to go either. Wait, what are you talking about? I didn't propose *any* particular punishment in my suggestion. I'm just trying to find a way to keep -CURRENT usable at least some of the time by channeling the commits. I know my suggestion isn't perfect, and the objections people have to it are valid. I just wish that somebody would come up with something better. Nate (I think) brought up the fact that people can't always commit when it would be suitable for my model. Would anybody care to think about a two-phase commit model: commit to the bleeding edge whenever you want, commit things that roughly work within a week? Greg To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19980223104605.49121>