[MINC-development] test failures: minc2-label-test

[Robert D. Vincent]

Hi Steve,

I just checked in a set of changes that seems to fix the minc2-label-test
problem on my emulated MIPS64 system. HDF5 does not provide converters
between enumerated types and native types, so we needed to supply one.
Also, we were not setting up the mtype_id field correctly when opening
label volumes.

    -bert

On Wed, Oct 7, 2015 at 12:10 PM, Robert D. Vincent <bert at phalarope.com>
wrote:

> Hi Steve,
>
> Well, whatever is happening, I can reproduce it on a MIPS64 emulator now,
> so I should be able to fix it.
>
>     -bert
> Thanks, Bert,
>
> > My memory here is a little weak, but I think we tried to create parallel
> > HDF5 types for volumes, such that the "native" type would be in the
> > machine's byte order, but the file type would always be little endian. My
> > guess is that the H5Tenum_insert() function does not do any translation
> of
> > its argument, so that we needed to swap bytes to keep the label values
> > consistent on big-endian architectures.
>
> Did you write that backwards?  The test failures coincide with big-endian
> machines -- the little endian ones all pass (except mipsel, which fails on
> a
> different test).
>
> What is even more confusing to me is that the test code's function
> create_label_image() actually defines the labels and then tests them --
> with
> no failures -- using the same miget_label_value() call that later fails at
> Line 172.  The difference I can see is that the first test operates on a
> volume created in memory and the second operates on a volume read from
> disk.
> Does that mean one is using 'mtype' and the other 'ftype'?  [That doesn't
> appear to be the case -- the code of miget_label_value() seems to use only
> mtype.]
>
>
>
> >
> > Once upon a time, we regularly tested on big-endian machines, but that is
> > rare these days. It's certainly possible that there are assumptions about
> > either the byte order or type sizes.
> >
> > Do you happen to know the HDF5 version in use?
>
> All use 1.8.15, except ppc64 which runs 1.8.13.
>
> -Steve
>
> _______________________________________________
> MINC-development mailing list
> MINC-development at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-development
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.bic.mni.mcgill.ca/pipermail/minc-development/attachments/20151009/f16aa57c/attachment.html>