From steve at sumost.ca  Sat Jan  3 16:20:10 2009
From: steve at sumost.ca (Steve M. Robbins)
Date: Sat, 3 Jan 2009 15:20:10 -0600
Subject: [MINC-development] "Amusing" mincblur issue...
In-Reply-To: <d747ed50811261607x3b0aa554r9e0e684f70e4b0dd@mail.gmail.com>
References: <d747ed50811232205m1085669fpb18201a34826c56a@mail.gmail.com>
	<d747ed50811261607x3b0aa554r9e0e684f70e4b0dd@mail.gmail.com>
Message-ID: <20090103212010.GL5408@sumost.ca>

On Thu, Nov 27, 2008 at 11:07:42AM +1100, Andrew Janke wrote:
> Hrm, seems like this is more endemic than I originally thought...

It just bit me trying to use rawtominc.

Amusingly, your proposed change exactly undoes revision 6.9 of
ParseArgv.c -- which was done this summer as part of 'large number of
changes from Claude (64 bit and updates)'.

I used your test code to create a new unit test (test_arg_parse.c and
run_test_arg_parse.sh) which exhibits the bug.  Then I reverted change
6.9 of ParseArgv.c and the new tests pass.  I think the code is now
correct, but it would be good for you and Claude to have a look to see
whether change 6.9 was fixing a problem and, if so, to augment
test_arg_parse.c.

I did a few other minor cleanups, including bumping to revision
2.0.17.

Regards,
-Steve

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://www2.bic.mni.mcgill.ca/pipermail/minc-development/attachments/20090103/8bcfbf4e/attachment.pgp 

From a.janke at gmail.com  Sat Jan  3 23:41:01 2009
From: a.janke at gmail.com (Andrew Janke)
Date: Sun, 4 Jan 2009 15:41:01 +1100
Subject: [MINC-development] "Amusing" mincblur issue...
In-Reply-To: <20090103212010.GL5408@sumost.ca>
References: <d747ed50811232205m1085669fpb18201a34826c56a@mail.gmail.com>
	<d747ed50811261607x3b0aa554r9e0e684f70e4b0dd@mail.gmail.com>
	<20090103212010.GL5408@sumost.ca>
Message-ID: <d747ed50901032041g65b2b53lc9a2d8f255120391@mail.gmail.com>

Thanks Steve,

Given that this was a revert of Claudes original changes (that I
committed) I was sort of waiting for a corroboration that what I had
done was "right".  In any case given that we both now get the same
result it would seem high time to release 2.0.17 and get this tested
in a few more places before other things break.

I will also do a few commits for the latest EBTKS + N3 + latest
gcc/libc + C++ problems that rear their heads on Ubuntu Intrepid and
Debian testing. (I forget the current release name). Oh how I love old
C++ code.....   (it currently barfs during the N3 build although it is
really a problem in EBTKS).


a

Steve M. Robbins <steve at sumost.ca>:
> On Thu, Nov 27, 2008 at 11:07:42AM +1100, Andrew Janke wrote:
>> Hrm, seems like this is more endemic than I originally thought...
>
> It just bit me trying to use rawtominc.
>
> Amusingly, your proposed change exactly undoes revision 6.9 of
> ParseArgv.c -- which was done this summer as part of 'large number of
> changes from Claude (64 bit and updates)'.
>
> I used your test code to create a new unit test (test_arg_parse.c and
> run_test_arg_parse.sh) which exhibits the bug.  Then I reverted change
> 6.9 of ParseArgv.c and the new tests pass.  I think the code is now
> correct, but it would be good for you and Claude to have a look to see
> whether change 6.9 was fixing a problem and, if so, to augment
> test_arg_parse.c.
>
> I did a few other minor cleanups, including bumping to revision
> 2.0.17.

From steve at sumost.ca  Sun Jan  4 00:19:20 2009
From: steve at sumost.ca (Steve M. Robbins)
Date: Sat, 3 Jan 2009 23:19:20 -0600
Subject: [MINC-development] "Amusing" mincblur issue...
In-Reply-To: <d747ed50901032041g65b2b53lc9a2d8f255120391@mail.gmail.com>
References: <d747ed50811232205m1085669fpb18201a34826c56a@mail.gmail.com>
	<d747ed50811261607x3b0aa554r9e0e684f70e4b0dd@mail.gmail.com>
	<20090103212010.GL5408@sumost.ca>
	<d747ed50901032041g65b2b53lc9a2d8f255120391@mail.gmail.com>
Message-ID: <20090104051920.GR5408@sumost.ca>

On Sun, Jan 04, 2009 at 03:41:01PM +1100, Andrew Janke wrote:
> Thanks Steve,
> 
> Given that this was a revert of Claudes original changes (that I
> committed) I was sort of waiting for a corroboration that what I had
> done was "right".  In any case given that we both now get the same
> result it would seem high time to release 2.0.17 and get this tested
> in a few more places before other things break.

Yes, a new release is a good idea, IMHO.  

By the way, the bug also caused mincheader and mincdiff to fail, which
I noticed some time ago [1] but didn't diagnose the cause of until
today.  I've already uploaded a fixed version of minc to Debian.


Cheers,
-Steve

[1] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=506276
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://www2.bic.mni.mcgill.ca/pipermail/minc-development/attachments/20090103/06d9d4f5/attachment.pgp 

From a.janke at gmail.com  Tue Jan 20 07:50:06 2009
From: a.janke at gmail.com (Andrew Janke)
Date: Tue, 20 Jan 2009 23:50:06 +1100
Subject: [MINC-development] "Amusing" mincblur issue...
In-Reply-To: <20090104051920.GR5408@sumost.ca>
References: <d747ed50811232205m1085669fpb18201a34826c56a@mail.gmail.com>
	<d747ed50811261607x3b0aa554r9e0e684f70e4b0dd@mail.gmail.com>
	<20090103212010.GL5408@sumost.ca>
	<d747ed50901032041g65b2b53lc9a2d8f255120391@mail.gmail.com>
	<20090104051920.GR5408@sumost.ca>
Message-ID: <d747ed50901200450u26b07b3bw834ca24931cc012@mail.gmail.com>

> Yes, a new release is a good idea, IMHO.
>
> By the way, the bug also caused mincheader and mincdiff to fail, which
> I noticed some time ago [1] but didn't diagnose the cause of until
> today.  I've already uploaded a fixed version of minc to Debian.

See packages.bic.mni.mcgill.ca/tgz for minc-2.0.17


a