My bad. RE: [Insight-users] GIPL File Format

[Jurgen.Fripp at csiro.au]

Thanx Julien

   It was my bad (sorry, should have checked more thoroughly before emailing ;(), we use an INR image factory, which I disabled (commented out so it wasn't registered) when using it to test gipl reading/writing, but I was still linking the programs with it's library. Removing the library from the link list fixed the gipl reading problem in itk 2.8.1. I haven't worked out what in the inr factory causes the problem (or maybe what's changed in IO Factory since 2.2, as it's code hasn't been updated for a while) but that should be easy to track down.

Sorry, thanx for help

Cheers Jurgen

-----Original Message-----
From:	Julien Jomier [mailto:julien.jomier at kitware.com]
Sent:	Thu 7/13/2006 3:14 AM
To:	Fripp, Jurgen (ICT Centre, North Ryde)
Cc:	insight-users at itk.org
Subject:	Re: [Insight-users] GIPL File Format
Hi Jurgen,

This is strange...

Couple of questions to see where the error might be:

a) Can you try to run the itkGiplImageIOTest located in 
Insight/Testing/Code/IO?
b) Can your program read the Insight/Testing/Code/Data/Input/ramp.gipl file?
c) Are you using any other object factory in your programs? (I guess the 
answer is no...)

I've seen this kind of error before and most of the time this is due to 
a mix of versions between includes and libraries.

Let us know,

Julien

Jurgen.Fripp at csiro.au wrote:
> Hi
>    Hopefully this isn't a false bug report, as I haven't been able to determine it's cause (or a proper fix). But at least using itk 2.2 gipl works by default in the reader/write. However, using itk 2.8.1 all my programs abort when trying to read gipl unless I explicitly register the gipl factory 
> ie 
> itk::ObjectFactoryBase::RegisterFactory( itk::GiplImageIOFactory::New() );
> 
> Note: This was only tested on two ubuntu 5.10 pc's and both used gcc-3.3.
> 
> I checked itkImageIOFactory and gipl is still registered and the filename check in gipl appears valid (and obviously works when explicitly registered). I assume it's clashing with another factory but haven't worked out which (although haven't spent too much time looking into it).
> 
> Cheers Jurgen