From smr at sumost.ca  Sun Nov 12 22:09:01 2006
From: smr at sumost.ca (smr@sumost)
Date: Sun, 12 Nov 2006 21:09:01 -0600
Subject: [MINC-development] MINC 1.4.1
Message-ID: <000301c706d1$12539eb0$6401a8c0@Bayes>

Hello,

I noticed minc-1.4.1.tar.gz on packages.bic.mni.mcgill.ca/tgz (dated 30 
October) so I packaged it for debian.

While doing so, however, I noticed that this tarball is not from the tip of 
the MINC 1.x branch
so lots of the fixes for the past year are missing.  In addition, there is 
no minc-1-4-1 tag on
CVS.

Is this an oversight or intentional?

-Steve

From a.janke at gmail.com  Mon Nov 13 00:18:40 2006
From: a.janke at gmail.com (Andrew Janke)
Date: Mon, 13 Nov 2006 00:18:40 -0500
Subject: [MINC-development] MINC 1.4.1
In-Reply-To: <000301c706d1$12539eb0$6401a8c0@Bayes>
References: <000301c706d1$12539eb0$6401a8c0@Bayes>
Message-ID: <d747ed50611122118i892ed87rd3606bb5d053ce63@mail.gmail.com>

Hi Steve,

This was an abomination released upon the world by Claude in a moment
of weakness.  No it isn't an official release and no I didn't sanction
it! :)

At the moment I am a tad stuck trying to figure out just which is the
HEAD 1.x version.  Happen to have a magic CVS command that will get me
an uptodate 1.x checkout?

ta


a

On 11/12/06, smr at sumost <smr at sumost.ca> wrote:
> Hello,
>
> I noticed minc-1.4.1.tar.gz on packages.bic.mni.mcgill.ca/tgz (dated 30
> October) so I packaged it for debian.
>
> While doing so, however, I noticed that this tarball is not from the tip of
> the MINC 1.x branch
> so lots of the fixes for the past year are missing.  In addition, there is
> no minc-1-4-1 tag on
> CVS.
>
> Is this an oversight or intentional?
>
> -Steve
> _______________________________________________
> MINC-development mailing list
> MINC-development at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-development
>


-- 
Andrew Janke   (a.janke at gmail.com || http://a.janke.googlepages.com/)
Canada->Montreal                                   Cell: +1 (514) 924 2012

From smr at sumost.ca  Mon Nov 13 09:58:23 2006
From: smr at sumost.ca (Steve Robbins)
Date: Mon, 13 Nov 2006 08:58:23 -0600
Subject: [MINC-development] MINC 1.4.1
In-Reply-To: <d747ed50611122118i892ed87rd3606bb5d053ce63@mail.gmail.com>
References: <000301c706d1$12539eb0$6401a8c0@Bayes>
	<d747ed50611122118i892ed87rd3606bb5d053ce63@mail.gmail.com>
Message-ID: <20061113085823.a13n31n6y7pcs8k8@sumost.ca>

Quoting Andrew Janke <a.janke at gmail.com>:

> This was an abomination released upon the world by Claude in a moment
> of weakness.  No it isn't an official release and no I didn't sanction
> it! :)

Oh.  Well I'd be more than happy to roll a 1.4.2 release from head of  
CVS.  It would come with up-to-date autoconfigury, which is a bonus  
from my perspective as Debian maintainer as the Debian patch will thus  
be a lot smaller.

I just need someone to advise me on what to do with the .tar.gz once built.


> At the moment I am a tad stuck trying to figure out just which is the
> HEAD 1.x version.  Happen to have a magic CVS command that will get me
> an uptodate 1.x checkout?

   cvs -d .... checkout -r minc-1-X-branch minc


For future reference, do a "cvs stat -v ChangeLog" to see all existing  
tags and look for one of type "branch" (rather than "revision").

Regards,
-Steve


From a.janke at gmail.com  Tue Nov 14 23:10:56 2006
From: a.janke at gmail.com (Andrew Janke)
Date: Tue, 14 Nov 2006 23:10:56 -0500
Subject: [MINC-development] MINC 2.1
Message-ID: <d747ed50611142010l7f35b111r1e424f9d00c5f87e@mail.gmail.com>

Hi all,

I think it is high time that we got together and released MINC 2.1,
what changes would I like to see in this release?

   1. As per Sebas and Jon Harlaps suggestion I would like to propose
that we do away with the default writing of MINC 1 files  and instead
write MINC 2 by default.  This would mean that we would add a -1 flag
to all binaries and make the default -2.

  2. Change the build process such that --enable-minc-2 or whatever it
is called is on by default.

Thoughts?

So the next question is who would like to volunteer to do this? :)



-- 
Andrew Janke   (a.janke at gmail.com || http://a.janke.googlepages.com/)
Canada->Montreal                                   Cell: +1 (514) 924 2012

From baghdadi at sickkids.ca  Wed Nov 15 11:08:52 2006
From: baghdadi at sickkids.ca (Leila Baghdadi)
Date: Wed, 15 Nov 2006 11:08:52 -0500
Subject: [MINC-development] MINC 2.1
In-Reply-To: <d747ed50611142010l7f35b111r1e424f9d00c5f87e@mail.gmail.com>
References: <d747ed50611142010l7f35b111r1e424f9d00c5f87e@mail.gmail.com>
Message-ID: <1163606932.9019.325.camel@localhost.localdomain>

Hi Andrew,

I would love to volunteer except I am busy writing image processing
code.

I could still take care of it if nobody else volunteered within the next
two weeks.

Leila

On Tue, 2006-14-11 at 23:10 -0500, Andrew Janke wrote:
> Hi all,
> 
> I think it is high time that we got together and released MINC 2.1,
> what changes would I like to see in this release?
> 
>    1. As per Sebas and Jon Harlaps suggestion I would like to propose
> that we do away with the default writing of MINC 1 files  and instead
> write MINC 2 by default.  This would mean that we would add a -1 flag
> to all binaries and make the default -2.
> 
>   2. Change the build process such that --enable-minc-2 or whatever it
> is called is on by default.
> 
> Thoughts?
> 
> So the next question is who would like to volunteer to do this? :)
> 
> 
> 


From a.janke at gmail.com  Mon Nov 20 03:36:36 2006
From: a.janke at gmail.com (Andrew Janke)
Date: Mon, 20 Nov 2006 03:36:36 -0500
Subject: [MINC-development] For the bored or morbidly interested
Message-ID: <d747ed50611200036r6ca70882s323ae16a55b1c077@mail.gmail.com>

May I present shell-pipe 0.1.

yes, it is currently flaky, yes it only sort of works, yes I do like
to re-invent the wheel.

whatisit? yet another way of running a MINC pipe that handles
dependencies, I know that steve robbins had a hack at this a while ago
and he even wrote his own make replacement from what I remember.

so what _really_ is it? a bunch of simple shell scripts that do some
basic pre-processing on a series of MINC files and a Makefile to put
it all together.

Good stuff:

   * you can now just add another subject and type make.

   * you can update a script, type make and "the right thing" will happen

   * you can replace a native file with a new one and "the right
thing" will happen -- unless of course you are evil and change the
mtime.

   * you can add another script and type make.

So, yes I have tried to make it as fault tolerant and extensible as I
can, but it still amounts to a few shell scripts and a Makefile as
opposed to yet another rppl/civet/etc. My plan here is not to take
over the world (yet) but instead to try to provoke a bit of thought in
a slightly different direction to what we usually seem to.   That and
the fact that Claude now builds his whole quarantine from a Makefile
so it will integrate _very_ nicely with this. (for the binaries
anyhow).

Now if I could just bend autoconf to my needs i could make (ha pun!) a
version in automake/conf.

Without further ado, get it here (or attached):

   ~rotor/shell-pipe-0.1.tar.gz

or in minc_dev/shell-pipe for the CVS types.

As i said there are still a few bugs so be sure to use make -k as
opposed to just make.  You will also have to run make twice for some
esoteric reason that I cannot fathom that is to do with vpath
directives.

There is a rudimentary README in there as well.


-- 
Andrew Janke   (a.janke at gmail.com || http://a.janke.googlepages.com/)
Canada->Montreal                                   Cell: +1 (514) 924 2012
-------------- next part --------------
A non-text attachment was scrubbed...
Name: shell-pipe-0.1.tar.gz
Type: application/x-gzip
Size: 6814 bytes
Desc: not available
Url : http://www.bic.mni.mcgill.ca/pipermail/minc-development/attachments/20061120/7a1233a7/shell-pipe-0.1.tar.bin

From smr at sumost.ca  Mon Nov 20 10:11:17 2006
From: smr at sumost.ca (Steve Robbins)
Date: Mon, 20 Nov 2006 09:11:17 -0600
Subject: [MINC-development] [geeks] For the bored or morbidly interested
In-Reply-To: <d747ed50611200036r6ca70882s323ae16a55b1c077@mail.gmail.com>
References: <d747ed50611200036r6ca70882s323ae16a55b1c077@mail.gmail.com>
Message-ID: <20061120091117.mo817gxpj1cko0sw@sumost.ca>

Quoting Andrew Janke <a.janke at gmail.com>:

> whatisit? yet another way of running a MINC pipe that handles
> dependencies, I know that steve robbins had a hack at this a while ago
> and he even wrote his own make replacement from what I remember.

Yes, I wrote a tool in perl that computed dependencies and the like.   
Later on, I gave it all up and used a simple Makefile as you are doing.

If I had to do it again, I'd use scons: a make replacement where you  
can write rules in python.  I don't have any direct experience with it  
yet, but scons seems to be designed to support parallelism very well.

-Steve
P.S.  I just learned that cons (a perl-based predecessor to scons)  
already existed at the time I had written my perl tool.