octave-bug-tracker
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Octave-bug-tracker] [bug #31471] Problem saving ranges with HDF5 format


From: David Bateman
Subject: [Octave-bug-tracker] [bug #31471] Problem saving ranges with HDF5 format
Date: Tue, 02 Nov 2010 15:34:14 +0000
User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.0.14) Gecko/2009091010 Iceweasel/3.0.14 (Debian-3.0.14-1)

Follow-up Comment #2, bug #31471 (project octave):

The problem with trying to fix this issue is that the number of elements in
the range needs to be saved to the HDF5 file to ensure that range can be
restored with the correct number of elements. This of course introduces
forwards and backwards compatibility issues.

Conceptual the easiest way to fix this would be to add the number of elements
(or replace the limit value) in the HDF5 compound value that is saved for the
range. We could treat the backward compatibility issue by testing for both
compound types in the range code and use the right version, however this
technique can never ensure forward compatibility.

A method to ensure forward and backward compatibility would be to save the
number of elements of the range as an attribute of the range. With older
versions of Octave this attribute would be ignored and if it didn't exist in
newer versions of Octave we can just do without it, ensuring full
compatibility. However this means that one value needed to restore the range
has a different treatment from the others, so it might be considered as ugly.

I applied a changeset for this in

http://hg.savannah.gnu.org/hgweb/octave/rev/2271261f088a

If you are happy with this changeset then please close this bug. Otherwise,
feel free to back it out if you consider this idea too ugly.

Cheers
D.


    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?31471>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/




reply via email to

[Prev in Thread] Current Thread [Next in Thread]