Date: Sun, 05 Jan 1997 22:00:43 -0800 From: "Jordan K. Hubbard" <jkh@time.cdrom.com> To: Stephen McKay <syssgm@devetir.qld.gov.au> Cc: CVS-committers@freefall.freebsd.org, cvs-all@freefall.freebsd.org, cvs-sys@freefall.freebsd.org Subject: Re: cvs commit: src/sys/i386/isa/bs bs.c bs_isa.c bs_pisa.c bsfunc.c bsfunc.h bshw.c bshw.h bshw.lst bshw_dma.c bshw_pdma.c bsvar.h ccbque.h dvcfg.h scsi_dvcfg.h Message-ID: <12738.852530443@time.cdrom.com> In-Reply-To: Your message of "Mon, 06 Jan 1997 13:14:01 %2B1000." <199701060314.NAA07311@troll.devetir.qld.gov.au>
next in thread | previous in thread | raw e-mail | index | archive | help
> >I recognise the *potential* benefit of such expansion, but balanced
> >against the *clear and obvious annoyances* that they cause, I don't
> >think they're worth it at all.
>
> My first impression is "Bollocks!", but that's too harsh in a public forum. :
-)
My reply to Jason Thorpe was *definitely* not something for a public
forum. :-)
> I add "$Id" lines to all my code, at home and at work, and normally to be
SM: "Hello, my name is Stephen"
Group: "Hello Stephen!"
SM: "I.. I.. Add $Id$ lines to my code! At home and (*sob*) at work
during the day too!"
Group: "That's OK, we love you Stephen!"
Therapist: "Group hug!!"
:-)
Toning down what I said to Jason for more public expression, let me
simply state my opinion that $Id$ was a really bad, stupid, genuinely
EVIL idea which was, like all truly bad ideas, promptly enshrined in
SCCS and RCS and preserved to haunt and torment us through the ages,
like an egyptian mummy's eternal curse.
I want my sources to represent what's currently needed by myself and
the compiler to read and compile the code, nothing more. Change log
information, obscure version numbers, lengthy API documentation, there
are other places for that sort of information to go, places where I'd
much *prefer* it to be, and it doesn't bloody well need to go anywhere
near my code! :-)
At no stage during this 4 year project has the $Id$ information ever
been of any use to me, nor have I ever received an email which said
"hey, are you running that binary with version 1.4.6.9.1 or version
1.19.3.7 of foo.c?" We just don't communicate that way - we send
diffs or we point to CVS log information which bears little relation
whatsoever to whatever version of foo.c I happen to have checked out,
or the binaries I have lying around.
I'm not saying there's no conceivable use for such a feature, I'm
simply saying that I've never needed it.
As to the annoyances, it's simply the constant barrage of garbage
diffs they generate when you're trying to examine differences between
files. I don't care when NetBSD bumped its bloody version numbers,
for example, I only care about the actual changes they made. Likewise
doing CVS merges between branches just generates extraneous cruft from
the $Id$ strings which you then have to edit out or simply smash down
on top of the old in your patch run, something which offends my
delicate sensibilities. :-)
I say it's the wrong way of solving this problem and we have all the
version information we need already in the CVS repository without
polluting the source with it. If it's not easy to say "what version
are you, file?" then we should make that easy. Above all, we should
try to stop doing silly things because "we've always done them this
way." :-)
Jordan
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?12738.852530443>