Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 16 Jun 1998 07:23:17 -0600 (MDT)
From:      Atipa <freebsd@atipa.com>
To:        "Jordan K. Hubbard" <jkh@time.cdrom.com>
Cc:        "Justin M. Seger" <jseger@freebsd.scds.com>, ports@FreeBSD.ORG, asami@cs.berkeley.edu
Subject:   Re: Size of a port... 
Message-ID:  <Pine.BSF.3.96.980616065354.14521A-100000@altrox.atipa.com>
In-Reply-To: <4693.897983741@time.cdrom.com>

next in thread | previous in thread | raw e-mail | index | archive | help


On Tue, 16 Jun 1998, Jordan K. Hubbard wrote:

> > Speaking of sizes and ports, is anyone else totally sick of the ungodly
> > number of files in /usr/ports? It to me is absolutely disgusting. It takes
> > several minutes to to any type of "find" in /usr because of them.
> 
> Then don't run your find in /usr/ports. :-)

Well, what about "du -k" then?

It still to me is a kuldge. You know the
I/O is way too bound when untarring the ports can only be done
at 15kb/sec when sysinstalling via 100MBit/sec ethernet on a P2-400.

> I hope you'll pardon me for saying so, but this has got to be one of
> the stupidest posts I've ever read and rivals the "it has too many
> notes!" scene in Amadeus. 

Actually, this is one of my pet peeves, probably to an irrational level.
Those ports maintiners are just too damn good! Try this:

% time ls -R /usr/ports | wc -l
(standard input)
time: command terminated abnormally
	143.24 real	1.02 user	4.30 sys

That was on a P-166 IDE disks 2.2.5-RELEASE.

On a Pro-200-512k w/ 256M RAM and 7200RPM disks, it did:
% time ls -R /usr/ports | wc -l
   60693
real	2m26.05s
user    0m1.241s
sys     0m3.95s

And you think I am crazy? Sheesh. Sixty thousand inodes is not excessive,
when about only 1300 are needed? It took 2.5 minutes just to do a damn
ls -R on that tree. du, find, etc., are all very slow because of the
current structure.

> I can only hope that the fine folks at
> Atipa are just having a Really Bad Day and will send a follow-up
> posting shorting admitting that Kevin probably should have been kept
> away from the keyboard after 5:00pm and simply gotten dead drunk
> instead. :-)

Well, I may have been a bit loopy, but not off my rocker. Hopefully
softupdates will help out, but that won't really cahnge the problem.

> The ports collection has a very deliberate design with many advantages
> if one actually understands it (and few do, with many more only
> thinking that they do :), and its layout is neither disgusting nor
> ridiculous.  It is now time for Kevin to go lie down. :-)

I can see several advantages, but none that _require_ that layout. For
instance, CVS-ability, indexing, describing, patching, building, etc.,
could all be maintined by a "ports" command and about 4 or 5 files
external to the .tgz files for each port. 

If I am crazy, try looking at the number of files generated when a SQL
server maintains 60,000 records. Probably about 3.

Am I making any sense? I have yet to drink my coffee... :)

Kevin


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-ports" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.3.96.980616065354.14521A-100000>