[ome-users] Compile BF-ITK-pipe on Windows

[Mark Hiner]

Hi Joe,

 Just wanted to let you know that I opened this pull request(https://github.com/openmicroscopy/bioformats/pull/63) integrating your itkBioFormatsImageIO.cxx so hopefully it will make it to the main Bio-Formats repository soon.

 That branch also includes improved CMake support for Release/Debug builds and a few minor fixes.
 
 If any other issues come up, just let us know. Thank you again for your contribution!

- Mark

On 05/02/12, Joe Ping-Lin Hsiao 
 wrote:
> I am glad to know it's working on your side as well. Looking forward
> to seeing it integrate into the official release.
> 
> Thanks,
> Joe
> 
> On Tue, May 1, 2012 at 6:58 PM, Curtis Rueden <ctrueden at wisc.edu> wrote:
> > Hi Joe,
> >
> > Thank you very much for investigating, and sending a patch. We really
> > appreciate it. In the meantime, I did some digging of my own and concluded
> > that the previous code must never have worked correctly on Windows (unless
> > using gcc via Cygwin for the POSIX functions). So your work is definitely
> > valuable.
> >
> > Mark Hiner, one of the main BF-ITK developers, looked at your solution and
> > says it seems to be working well, so we will integrate it into the official
> > version shortly. However, I am traveling until next Monday, so it will be a
> > little while before you hear back from me again. But we'll keep you posted!
> >
> > Thanks again,
> > Curtis
> >
> >
> >
> > On Fri, Apr 27, 2012 at 5:47 PM, Joe Ping-Lin Hsiao <phsiao at cs.unc.edu>
> > wrote:
> >>
> >> Curtis,
> >>
> >> I modified bf-itk-pipe source so that it compiles and works on Windows
> >> 7 64bit. What I did was replacing Linux system calls to the ones that
> >> work on Windows and fixing the code which depends on new line
> >> characters (Windows uses '\r\n' instead of just '\n').
> >>
> >> All I changed is just one file: itkBioFormatsImageIO.cxx, and I have
> >> attached the modified version. Search for '#ifdef WIN32' to see what
> >> the modifications are.
> >>
> >> Thanks,
> >> Joe
> >>
> >> On Mon, Apr 16, 2012 at 1:25 PM, Joe Ping-Lin Hsiao <phsiao at cs.unc.edu>
> >> wrote:
> >> > Hi Curtis,
> >> >
> >> >> While we have not seen such compile errors per se, Windows support is
> >> >> definitely the least well tested. We have historically had a lot of
> >> >> difficulty getting the Bio-Formats ITK plugin (all three iterations of
> >> >> it,
> >> >> not just the pipes version) working in Windows at all. So it is not too
> >> >> surprising that it is not working for you now.
> >> >
> >> > May I ask what develop environment you used on Windows? I got the same
> >> > errors from both Visual Studio 2008 and Mingw with Cygwin. I also
> >> > noticed that Bio-Formats website has 32bit BF-ITK binary for download.
> >> > I need a 64bit version so I can't use it, but it's still good to know
> >> > how it was built.
> >> >
> >> >> If you do decide to investigate further and have any questions about
> >> >> it,
> >> >> please do not hesitate to ask!
> >> >
> >> > I fixed those errors by using _pipe(), _write(), and _close() from
> >> > <io.h>. The library itself compiled fine. But when load images at run
> >> > time, I got this error:
> >> >
> >> >
> >> > \src\bioformats\components\native\bf-itk-pipe\itkBioFormatsImageIO.cxx:229:
> >> > BioFormatsImageIO: ITKReadImageInformation error:
> >> > The handle is invalid
> >> >
> >> > The code before where this error happens is:
> >> >
> >> > int m_Pipe[2]; // Declared in
> >> > BioFormatsImageIO.h by me.
> >> > ...
> >> > _pipe( m_Pipe, 512, O_BINARY ) // This is my Windows version of pipe()
> >> > m_Process = itksysProcess_New();
> >> > itksysProcess_SetCommand( m_Process, m_Argv );
> >> > itksysProcess_SetPipeNative( m_Process, itksysProcess_Pipe_STDIN,
> >> > (itksysProcess_Pipe_Handle*)m_Pipe );
> >> > itksysProcess_Execute( m_Process );
> >> >
> >> > m_Pipe has '3' and '4' after the _pipe(), so I think they were setup
> >> > fine. Any idea?
> >> >
> >> > Thanks,
> >> > Joe
> >
> >
> _______________________________________________
> ome-users mailing list
> ome-users at lists.openmicroscopy.org.uk
> http://lists.openmicroscopy.org.uk/mailman/listinfo/ome-users
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openmicroscopy.org.uk/pipermail/ome-users/attachments/20120508/b9b9754f/attachment.html>