|
Notes |
|
|
(0020643)
|
|
Utkarsh Ayachit
|
|
2010-05-07 13:29
|
|
|
I believe the reporter is referring to Ensight reader panel no longer having the endian-ness option. That's because the reader automatically figures out the endian-ness, without having the user to choose it. |
|
|
|
(0020664)
|
|
tzapuu
|
2010-05-08 10:45
(edited on: 2010-05-08 10:48) |
|
Hi,
Maybe I did not clarify the issue good enough. Sorry.
The endian detector you mention does very well the detection BUT I suppose it detect the endian writing based on the system - NOT the actual file. And here it comes the problem: If the files are produced on another Operating System/Machine/Cluster than the machine where paraview is running, one CAN NOT open them anymore.
So:
1 my desktop PC is a Linux machine (here i use Paraview)
2 I generate the "Ensight gold" files on a sun/solaris (different endian)
3 When I try to open the files the endianness is wrong taken from the Linux machine even if the files are written in different Endian order.
The files and so are corect as they work as a charm in Paraview 3.4
If more infos are needed please let me know.
Best regards,
Adrian
|
|
|
|
(0020709)
|
|
Ken Moreland
|
|
2010-05-11 11:51
|
|
|
It might be helpful if example files were uploaded/attached to this bug report. |
|
|
|
(0020827)
|
|
Utkarsh Ayachit
|
|
2010-05-26 15:05
|
|
|
Please attach a sample dataset. |
|
|
|
(0020835)
|
|
tzapuu
|
|
2010-05-27 09:19
|
|
Hi all,
As requested a small example has been uploaded.
Best regards,
Adrian |
|
|
|
(0020885)
|
|
Buddel
|
|
2010-06-01 13:09
|
|
Hey,
I have the same problem. If I try to open the "Ensight Gold" case-files on my Linux/Ubuntu mashine I get the same failure. I generate the files on sun solaris too. If I try to open the file unter Windows 32Bit, I get the same failure.
Here is the output message:
++++++++++++++++++++++++++++++++++++++++++++++
ERROR: In ..\..\..\src\Servers\Filters\vtkEnSightGoldBinaryReader2.cxx, line 150
vtkEnSightGoldBinaryReader2 (0543DB88): Error with subline extraction upon file initialization
ERROR: In ..\..\..\src\Servers\Filters\vtkEnSightReader2.cxx, line 307
vtkEnSightGoldBinaryReader2 (0543DB88): error reading geometry file
ERROR: In ..\..\..\src\VTK\Filtering\vtkExecutive.cxx, line 756
vtkCompositeDataPipeline (0C73C5D8): Algorithm vtkEnSightGoldBinaryReader2(0543DB88) returned failure for request: vtkInformation (05503CE0)
Debug: Off
Modified Time: 75761
Reference Count: 1
Registered Events: (none)
Request: REQUEST_DATA
FROM_OUTPUT_PORT: 0
FORWARD_DIRECTION: 0
ALGORITHM_AFTER_FORWARD: 1
++++++++++++++++++++++++++++++++++++++++++++++
In my opinion the problem is the endian-ness "autodetection" which don't get the correct endian-ness of the binary files (with ParaView 3.4 I get errors like this if I choose the wrong endian-ness).
Is there a switch or option to turn off the autodetection of endian-ness?
Regards |
|
|
|
(0020962)
|
|
Utkarsh Ayachit
|
|
2010-06-10 09:56
|
|
commit 860a8c5225743eb078b99a977ca064f5de4c824e
Author: Utkarsh Ayachit <utkarsh.ayachit@kitware.com>
Date: Thu Jun 10 09:53:17 2010 -0400
Fixed BUG 0010593 Bug in EnSight byte-ordering auto-detection.
vtkEnSightGoldBinaryReader2::ReadLine() was getting called before the code that
auto-detected the byte order. There was code in ReadLine() that relied on
byte-order being set!!! That was the bug (esp. with fortran written files).
Fixed ReadLine() to handle the case when byte order hasn't been set explicitly.
--------------------------------------------------------------------------------
TODO: Commit should be added to "release" branch. |
|
|
|
(0021052)
|
|
Alan Scott
|
|
2010-06-16 18:10
|
|
|
Tested using included data, trunk, Windows, local server. |
|
|
|
(0021324)
|
|
Utkarsh Ayachit
|
|
2010-07-08 10:39
|
|
|
Paraview.tar.gz (65,204) 2010-05-27 09:12