Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 03 Dec 1999 18:17:45 -0700
From:      Warner Losh <imp@village.org>
To:        obrien@FreeBSD.org
Cc:        cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org
Subject:   Re: cvs commit: src/sys/kern subr_bus.c 
Message-ID:  <199912040117.SAA14552@harmony.village.org>
In-Reply-To: Your message of "Fri, 03 Dec 1999 16:57:13 PST." <19991203165713.A1939@dragon.nuxi.com> 
References:  <19991203165713.A1939@dragon.nuxi.com>  <Pine.BSF.4.05.9912030858490.12054-100000@semuta.feral.com> <199912031701.KAA11135@harmony.village.org> 

next in thread | previous in thread | raw e-mail | index | archive | help
In message <19991203165713.A1939@dragon.nuxi.com> "David O'Brien" writes:
: Often one does not need to do a `make world' to catch syntax errors.  A
: simple make of program (or kernel) would do.  A `make world' is [mostly]
: only needed to test interactions in the build process.

One problem is that of forgotten checkins.  I almost always checkin
building code.  However that is "building on my machine" rather than
"building with a fresh set of sources."  Several times I've had the
problem of neglecting to checkin a change to a file that has
changed/been added.  Many other people have had this problem as well,
and I know that the software compiles for them.

There are many times new source will be imported which doesn't
compile, but the import must be on a vendor branch unmodified.  The
files are then pulled off the vendor branch and made to compile
shortly after the initial commit.

Finally, there have been instances where I change files foo and bar
and make my commit.  Bob changes files baz and boo and makes his
commit.  The confluance of these changes causes a problem, but each
change taken singlely doesn't have the problem.  This has happened in
the past as well.

Automating a 'it must compile to be committed' rule is hard to do,
despite its desitrability.  The simple fact of life is that the
FreeBSD will be unbuildable from time to time and history has shown
that these periods rarely last more than a few hours (although
serveral days in a row snapshots may be broken by different people
each day).  The more mainstream the breakage (eg no kernels can be
built) gets faster attention than the less mainstream (eg make release
fails) due to the relative numbers of people doing each one.

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?199912040117.SAA14552>