[MINC-users] minc-2.0.15 build error

[EJ Nikelski]

Hi Andrew (and List),

   I've built using 32-bit on Leopard (other than the larger flat
addressing space, I'm at a loss with regard to why 64-bit should be
any faster).  In addition, I'm also using netcdf-3.6.2 and hdf5-1.6.5.
 The use of hdf5-1.6.5 was required to solve some EMMA/minc2-related
problems (hdf5-1.6.5 is the version that Matlab R2007a ships with, and
using a different version causes nasty problems).  At any rate, other
than the failure of the minc2 make test, everything is working well
(even EMMA writing minc2 volumes), so I'm hoping that this failure is
limited to the test.  Sadly, I have little time to throw at this ...
however, if you have a suggestion WRT putting some debug code
somewhere, I will happily recompile and re-test.

-Jim



On Thu, May 22, 2008 at 10:14 AM, Andrew Janke <a.janke at gmail.com> wrote:
> Hi Jim,
>
> I have tried to reproduce this error on another leopard machine but to
> no avail...  'make check' works fine.
>
> Have you made any progress on this yourself?  The machine I am using
> has a 64bit build of hdf/netcdf that Alexandre did, are you also
> building 64 bit of 32 bit?
>
>
> a
>
> On Sun, May 18, 2008 at 1:57 AM, EJ Nikelski <nikelski at bic.mni.mcgill.ca> wrote:
>>
>>   For reasons that I would rather not get into, I found myself
>> rebuilding minc2 on my Intel mac (os x 10.5).  Although I'm certain
>> that this worked previously, I now get an error during "make check" --
>> specifically:
>>
>> Testing icv_dim1
>> Testing icv_fillvalue
>> PASS: run_test2.sh
>> PASS: xfmconcat_01.sh
>> PASS: xfmconcat_02.sh
>> Assertion failed: (extent != 0), function px_get, file posixio.c, line 412.
>> /bin/sh: line 1: 66090 Abort trap              ${dir}$tst
>> FAIL: mincapi
>> ==================================
>> 1 of 5 tests failed
>> Please report to a.janke at gmail.com
>> ==================================
>>
>>   I've done some snooping, and the problem is related to this bit of
>> code in mincapi.c  ...
>>
>>  /* Try to open the file for write - should fail.
>>   */
>>  fd2 = miopen(ip->name, NC_WRITE);
>>  if (fd2 >= 0) {
>>    FUNC_ERROR("miopen");
>>  }
>>
>>   The call to miopen does indeed fail, but instead of returning, it
>> fails on the assertion in posixio.c.   Andrew, given that you've
>> dutifully attached your name to the error message, I suppose that you
>> get first kick at the can.  What do you think?  BTW - I'm building
>> minc2 with hdf5-1.6.7 and netcdf-3.6.2.
>
>
>
> --
> Andrew Janke (a.janke at gmail.com || http://a.janke.googlepages.com/)
> Canberra->Australia +61 (402) 700 883
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www2.bic.mni.mcgill.ca/mailman/listinfo/minc-users
>



-- 
=================================
Jim Nikelski, Ph.D.
Postdoctoral Research Fellow
Bloomfield Centre for Research in Aging
Lady Davis Institute for Medical Research
Sir Mortimer B. Davis - Jewish General Hospital
McGill University
Tel: (514) 340-8222 x 2298
Fax: (514) 340-8295