Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 21 Oct 2005 17:39:58 -0500
From:      linimon@lonesome.com (Mark Linimon)
To:        "Michael C. Shultz" <ringworm01@gmail.com>
Cc:        freebsd-ports@freebsd.org
Subject:   Re: [SUGGEST] Reform eclipse and eclipse related ports
Message-ID:  <20051021223958.GA19955@soaustin.net>
In-Reply-To: <200510211519.47370.ringworm01@gmail.com>
References:  <43522953.6050700@ebs.gr> <200510211454.41789.ringworm01@gmail.com> <20051021220910.GA18988@soaustin.net> <200510211519.47370.ringworm01@gmail.com>

next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Oct 21, 2005 at 03:19:47PM -0700, Michael C. Shultz wrote:
> Seems like the quantity of ports available will eventually hit a plateau
> with the current two level directory structure.  No one is afraid to update
> the basic OS when its needed, even when it means using an entirly different 
> file system ( ie. UFS1 -=> 2 ),  why be so scared when it comes to the ports 
> system?

Then PLEASE SUBMIT PATCHES.  Tested ones.  Involving portsmon.  Involving
the build cluster.  Involving marcusom tinderbox.  Involving FreshPorts.
Involving everything in bsd.*.mk.  Involving fixing up all the dependencies
after all the thousands of repocopies.

You will be submitting thousands, if not tens of thousands, of lines of
patches to do so, invoving sh, awk, sed, perl, python, and SQL -- that I
know of.  There are probably others.

Now: I am not going to discuss this issue any further until I see those
patches.

People, you just have No Idea how much work you are talking about here,
just to fiddle around with organizing ports into directories on a physical
disk, which I will continue to restate my opinion until I am blue in the
face that is the wrong problem to solve _anyway_.

The _right_ problems to solve are searching and browsing.  If you solve
those problems correctly, the physical layout on disk becomes hidden as
an implementation detail and no one but hardcore ports developers ever
has to think about it again.

And you don't have to regression test thousands of lines of patches to
do so.

This is at least the 20th time this particular idea has been floated.
It hasn't gotten any better the last 19 times.  Please go back and read
the archives.  I'm done discussing it.

mcl



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20051021223958.GA19955>