Date: Thu, 28 Mar 2002 09:38:57 -0700 (MST) From: "M. Warner Losh" <imp@village.org> To: mi@aldan.algebra.com Cc: sobomax@FreeBSD.org, cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: Re: cvs commit: ports/Mk bsd.port.mk Message-ID: <20020328.093857.81408996.imp@village.org> In-Reply-To: <200203281628.g2SGSFp0016178@aldan.algebra.com> References: <1017310574.580.20.camel@notebook> <200203281628.g2SGSFp0016178@aldan.algebra.com>
next in thread | previous in thread | raw e-mail | index | archive | help
In message: <200203281628.g2SGSFp0016178@aldan.algebra.com>
Mikhail Teterin <mi@aldan.algebra.com> writes:
: On 28 Mar, Maxim Sobolev wrote:
:
: > The whole affair (the problem spotted by knu and this one) showed that
: > bento testing, is important, but should not be used as a substitute
: > for ordinary peer review and testing, because bento's bsd.port.mk
: > usage pattern is quite poor. Unfortunately, I suspect that even
: > though my patch was available for more that two weeks but only few
: > individuals actually gave it a try. That's the real problem, which
: > needs to be addressed somehow.
:
: That's fine, actually. Only what you commit gets tested -- and fixed.
: Seems like a normal worflow to me -- as long as the fixing part comes
: promptly.
I have found that there's an increasisng group of people that test
things. When I put patches up, I'll get 2 or 3 people telling me
"looks good" or "tested it and XXXX is right or wrong." If I have a
series of patches, I'll put out maybe 10 or 12 versions of the patches
and get maybe 8 people to test it.
Then when I commit it to -current, I get about 500 people testing it
within a week (at least for pccard stuff). I find out within a week
if I've broken something on somebody's machine. So when I commit more
than once a week, it can be a little hard to sort out which thing
broke stuff, but with small commits it is easy for the complainer to
do the binary search for me.
When I commit to -stable, I get about 100000 people testing things
right away. I find out that things break within a few days, even
after letting things mellow in -current for a while. They have more
whacked out hardware than the folks that run -current, generally
speaking. However, since I batch things into -stable, it can be a
whole lot harder to find out what, exactly, I did to break things,
which has left a few laptops non-functional to this day....
Warner
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020328.093857.81408996.imp>