[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Octave-bug-tracker] [bug #45225] HDF5 I/O broken for int32 variables
From: |
anonymous |
Subject: |
[Octave-bug-tracker] [bug #45225] HDF5 I/O broken for int32 variables |
Date: |
Mon, 01 Jun 2015 10:45:34 +0000 |
User-agent: |
Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/43.0.2357.81 Safari/537.36 |
URL:
<http://savannah.gnu.org/bugs/?45225>
Summary: HDF5 I/O broken for int32 variables
Project: GNU Octave
Submitted by: None
Submitted on: Mo 01 Jun 2015 10:45:32 UTC
Category: Interpreter
Severity: 3 - Normal
Priority: 5 - Normal
Item Group: Incorrect Result
Status: None
Assigned to: None
Originator Name: Markus Grabner
Originator Email: address@hidden
Open/Closed: Open
Discussion Lock: Any
Release: 4.0.0-rc4
Operating System: Microsoft Windows
_______________________________________________________
Details:
Writing/reading integer variables to/from HDF5 files worked in octave-3.8.2,
but is broken in octave-4.0.0 (tested with the final release and not rc4, but
4.0.0 is not yet in the bug report combo box). The issue appears both under
Linux and Windows.
This is demonstrated by the attached scripts. Running one in octave-3.8 and
the other one in octave-4.0 gives incorrect results:
[octave-3.8]
>> test_hdf5_save
x = 123456789
[octave-4.0]
>> test_hdf5_load
x = 255
The other way round is also incorrect:
[octave-4.0]
>> test_hdf5_save
x = 123456789
[octave-3.8]
>> test_hdf5_load
x = 21
(note that 123456789 == 0x75BCD15, and that its last two hex digits 0x15 ==
21, which is the value read by octave-3.8)
Only when both scripts are run in octave-3.8, we get correct results:
[octave-3.8]
>> test_hdf5_save
x = 123456789
>> test_hdf5_load
x = 123456789
The HDF viewer provided by the HDF group
(https://www.hdfgroup.org/products/java/hdfview) confirms that the values read
and written by octave-3.8 are correct in the above example.
_______________________________________________________
File Attachments:
-------------------------------------------------------
Date: Mo 01 Jun 2015 10:45:32 UTC Name: test_hdf5_save.m Size: 61B By:
None
scripts to demonstrate broken HDF5 I/O for integer variables in octave-4.0.0
<http://savannah.gnu.org/bugs/download.php?file_id=34137>
-------------------------------------------------------
Date: Mo 01 Jun 2015 10:45:32 UTC Name: test_hdf5_load.m Size: 36B By:
None
scripts to demonstrate broken HDF5 I/O for integer variables in octave-4.0.0
<http://savannah.gnu.org/bugs/download.php?file_id=34138>
_______________________________________________________
Reply to this item at:
<http://savannah.gnu.org/bugs/?45225>
_______________________________________________
Nachricht gesendet von/durch Savannah
http://savannah.gnu.org/
- [Octave-bug-tracker] [bug #45225] HDF5 I/O broken for int32 variables,
anonymous <=