Bus errors on real files with MacBook Pro

massey

This is a complete mystery to me. Deidre has a relatively new Mac Pro, and whenever she tries to do anything with a "real" FITS file she gets a BUS error; i.e., even an imstat generates the message. On MY (older) Macbook Pro, looking at the identical files, and running the identical IRAF (I know it's identical because I just reinstalled it from my machine) I have NO problem with the same files.
Any ideas? I remember there was some issue involving this with the old PPC binaries once upon a time, but we are both running on modern INTEL Mac Pros with snow leopard and using version 2.14.1. Deidre has no problem looking at the same files on her Mac Pro desktop, again running the identical version of IRAF.

fitz · Posted: Fri Jun 24, 2011 4:55 pm Post subject:

Hi Phil,

There are no known problems or changes made in the system since that directly address this problem so I'm not sure what to suggest. If you could upload a sample file to ftp://iraf.noao.edu/pub I'll try to reproduce the problem.

I assume everything works fine for IMH images (dev$pix?) and FITS files in integer format? What about an integer FITS file with a BSCALE/BZERO value (which becomes 'real' internally)?? Any other clues??

massey · Posted: Sat Jun 25, 2011 2:08 pm Post subject: moved file where you can get it

Hi, Mike---

Thanks! I've put a file on the anon ftp area at Lowell:
ftp.lowell.edu pub/massey WLM_laboca.fits
But the file works fine on MY MacBook Pro, so I'm mystified! Yes,
Deidre's system has no problem with pix$dev.imh or any other interger
device.

She's using one of the latest MacBook Pro with 8Gb RAM, snow leopard, and IRAF 2.14.1

THANKS!

---phil

fitz · Posted: Sat Jun 25, 2011 6:27 pm Post subject:

Hi Phil,

Well, it seems to work for me as well. The image, however, has a garbage DATE-OBS value which might be causing havoc with the header parsing and leading to a pointer problem on some systems (just a guess). Try adding a valid value to see if the problem goes away.

-Mike

massey · Posted: Fri Jul 29, 2011 3:02 am Post subject: problem SOLVED

So, Mike----Deidre also has been having an obscure AIPS problem. Tonight
we tackled it and found that the "new" MacBook Pros all boot into the 64-bit kernel with Snow Leopard. Forcing it to 32-bit solved the AIPS problem, and on top of that she can now read the header in this FITS file. How's that for obscure?