Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 22 Aug 1995 06:19:35 +0100
From:      Gary Palmer <gary@palmer.demon.co.uk>
To:        "Julian Stacey" <jhs@freebsd.org>
Cc:        ports@freebsd.org
Subject:   Re: bsd.ports.mk checksum 
Message-ID:  <6794.809068775@palmer.demon.co.uk>
In-Reply-To: Your message of "Tue, 22 Aug 1995 02:01:13 %2B0200." <199508220001.CAA12973@vector.eikon.e-technik.tu-muenchen.de> 

next in thread | previous in thread | raw e-mail | index | archive | help
In message <199508220001.CAA12973@vector.eikon.e-technik.tu-muenchen.de>, "Juli
an Stacey <jhs@freebsd.org>" writes:
>That's sounds a usable compromise, but I'd still prefer md5 didnt forcibly
>access distfiles, for ports that had already extracted succesfully.  
>Could you ask Gary P for his views please ?

Why does Satoshi have to ask my opinion? I really don't see the
relevance of that question. My opinions have been stated already in
public, which youself, Satoshi and everyone on the ports list have
seen.

>Earlier you made a false assertion that I was advancing a minority view 
>(re. md5), & based your intransigence to change on that assertion.  
>If I had not debunked that assertion with a bit of gentle sarcasm,
>nothing would have changed.

I would have to say it is the minority view. The ports mechanism is
NOT designed to have the sources left lying around, in EXTRACTED
form. It CANNOT handle this... (think about it - if it goes to the
next version, there will still be a .extract_done or whatever lying
around, and the system won't dare change anything, so you'll still
have the old sources extracted).

The ports system is decidedly compile, install and clean-up in one go,
or be burnt. Making it otherwise would require taking bsd.port.mk to
such a level of complexity that it probably would fail more often than
it would work.

Yes, I agree that the checksum routine should not be run at every
step, but I think that your reasons are mis-founded.

And throughout this discussion, you have YET to suggest a fix to this
situation! Where's the patch to fix this problem? Why haven't I seen
it? The ports system is a voluntary effort (probably more so than the
main code tree), and I think it is unreasonable to go flinging mud
(like you did in the message I am replying to) AND still expect us to
go digging and impliment the fixes you want, when you have shown
little inclination to do so youself.

>Questions:
>	Was the extension of md5 checking to the `all' target proposed
>	& discussed first, or just commited ?

I have no idea - I don't have the cvs log for bsd.port.mk locally.

You have an account on freefall, why ask such a pointless question
when it'd be easier for you to go find out the answer yourself instead
of making other people do the leg work?

>	Was the author, reviewer, & committer one person ?

See my last answer.

And if it was, it wouldn't surprise me. Most of my work to bsd.port.mk
has been done locally, and then committed to the tree to get feedback.
To date, no-one has had a problem with any of my changes (other than
my initial choice of names for the checksum stuff).

>	Should FreeBSD ports changes have a minimum of 2 people ?

If we did follow this standard, no work would get done. Very few
people who are `knowledgable' enough about the ports system to review
such changes (myself included) have the time to review every change to
the ports repository. I'm more than happy for someone to go an commit
stuff, as it'd be their problem to clear up.

>	Is anyone exempt from the 2 person rule ?

As far as I am concerned, for the ports tree, everyone is (perhaps
excepting bsd.port.mk, which is too easy to screw up unless you know
what you are doing). Since their name will be in the Makefile, the
commit logs, and probably the $Id$ line, they are easy to chase up,
and they only hurt themselves as their work will be viewed with some
suspicion thereafter.

And I have to say, this is fast becoming something that I'd prefer NOT
seeing on any of the FreeBSD mailing lists (i.e. a slanging
match). Either calm down, or take the next round off the freebsd-ports
list. If I reeive another e-mail on this subject which still has the
signature of a flamethrower, and it isn't in private, I won't answer
it. Sorry, but I will not be responsible for propogating this in
public.

Yours

Gary



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?6794.809068775>