X-MimeOLE: Produced By Microsoft Exchange V6.5
Received: by onstor-exch02.onstor.net 
	id <01C77BBD.44471AC0@onstor-exch02.onstor.net>; Tue, 10 Apr 2007 15:12:02 -0700
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Content-class: urn:content-classes:message
Subject: RE: Code review request
Date: Tue, 10 Apr 2007 15:12:02 -0700
Message-ID: <BB375AF679D4A34E9CA8DFA650E2B04E02215718@onstor-exch02.onstor.net>
In-Reply-To: <20070410135729.1f2d3198@ripper.onstor.net>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: Code review request
Thread-Index: Acd7stnjOTstjbU9QiGO0IKtEYVorgACT3zg
From: "Larry Scheer" <larry.scheer@onstor.com>
To: "Andy Sharp" <andy.sharp@onstor.com>

See comments in-line

-----Original Message-----
From: Andy Sharp=20
Sent: Tuesday, April 10, 2007 1:57 PM
To: Larry Scheer
Subject: Re: Code review request

We should talk more about this, but you don't need to hold up the
checkin for that discussion.  Go ahead and check this stuff in for now.

In general going forward what I want you to think about redesigning
as much of the makefile system as you want, including 100%.  The defs.mk
is a great place to start because it will help define a whole lot of
other things.

I think this is really important and probably the only chance you'll
ever get to redesign this stuff to your heart's content.  And in my
opinion it needs a lot of it.  For instance, I really don't like the
multiple, iterative self referencing target stuff.  Where a piece
of the target is parsed off, a freakin' environment variable set, and
then make calls itself again, and starts the same process all over
again.  Potentially calling itself 2 or 3 times before it actually
starts making anything.  That's just sick and wrong.  That whole design
needs to be completely replaced, IMHO.

>>> You must be referring to what I call the "convenience targets"=20
>>> such as make dbg-ssc-bc. I didn't want to do them but was threatened
>>> with being voted off the island unless I provided them. I put them
in
>>> at the very last minute before my contract expired in 2004.
>>> That is why I still invoke make as "make PROD=3Dbc VARIANT=3Ddbg
TARGET=3DSSC"
>>> Actually I rarely provide TARGET unless I want to override the
default.

Anyway, stuff to think about, and let's talk about it some more.

>>> Okay, it is good to have someone else around who understands this
>>> stuff beyond a superficial level.

Cheers,

a

On Mon, 9 Apr 2007 18:41:58 -0700 "Larry Scheer"
<larry.scheer@onstor.com> wrote:

> Comments in line below maked with '--'.
>=20
>=20
> -----Original Message-----
> From: Andy Sharp
> Sent: Mon 4/9/2007 4:49 PM
> To: Larry Scheer
> Subject: Re: Code review request
> =20
> On Mon, 9 Apr 2007 14:34:36 -0700 "Larry Scheer"
> <larry.scheer@onstor.com> wrote:
>=20
> > Andy,
> >    When you have a moment can you review the following files. They
> > are in my home dir so you should have access to them.
> >=20
> > Script to create cross packages and comvert them to rpms:
> > =09
> > /homes/larrys/perforce/trees/cougar/linux/pkg-tools/deb2cross2rpm
>=20
> I don't know if it works, but it looks OK. I think if you invoke it
> without arguments it just silently exits, but it that's what you want
> it to do, then that's OK too.
>=20
> -- Yea, I need to add a usage message and exit when there are no args.
> -- Musta forgot that...
>=20
> > Global makefiles
> > =09
> > /homes/larrys/perforce/trees/cougar/nfx-tree/Tools/defs.mk
>=20
> I don't think this is ready yet.  We should sit down and discuss the
> overall plan for the Makefiles and includes and stuff.  I just want to
> hear what you have in mind and so forth.  I'm not clear on where this
> file fits in and at what stages and so forth.  And I'm not too
> thrilled with the massive number of ifdefs (ifeqs in makefiles).
>=20
> -- My changes are not that much to this file. This is an Onstor
> Original, -- 90% of what you are looking at is legacy.
> -- I have been talking to Jeff about splitting this file apart.
> Making -- massive design changes to this file has always been risky
> since -- everything uses it, even the prom builds use it, so I am
> told.=20
>=20
> > /homes/larrys/perforce/trees/cougar/nfx-tree/Tools/tool-chain-pkg.mk
>=20
> Fill in the first TODO ~:^)
>=20
> I think you're getting there, or maybe you're already there, but let's
> sit down (stand up?) tomorrow and you can run me through this stuff a
> little bit before you check it in if that's OK.  Tomorrow morning is
> fine if that works for you. Maybe grab a room with a whiteboard.
>=20
> Line 59 has trailing whitespace.
>=20
> I just want to make sure that I understand the large strokes of the
> overall flow of the design, and what stage we think we're in and what
> direction we're going to be going in as things progress.
>=20
> -- Ok, I need to see what tomorrow brings, but maybe we can do a=20
> -- whiteboard talk before lunch.
>=20
> Cheers,
>=20
> a
>=20
