msccmatch - floating point overflow

Mike Smutko wrote on Jun 26, 2013

I'm trying to fix up some WCS info in some images, but whenever the routine tries to update the coordinate information (update = yes) in the fits file, I get (for example) "ERROR: floating point overflow called as: `msccmatch (input=110613.0203.fits)' "

I admit it's been a while since I've used msccmatch, but I don't recall ever having had this problem before. If I set (update=no), the routine runs fine, except of course, it doesn't update the coord info in the fits file.

I'm using IRAF 2.16 on MacBookPro with 10.8.4. I'd be happy to send some example files if that would help.

---

Mike Fitzpatrick wrote on Jun 26, 2013

I've traced the problem to the CCMAP task which is called from MSCCMATCH when writing the solution. This appears to be a 64-bit issue that has already been fixed and the solution to your problem is to replace the $iraf/bin.macintel/x_images.e binary on your system with the patched version at

ftp://iraf.noao.edu/iraf/v216/support/macintel/x_images.e

Download in binary mode, make executable with the command "chmod 755 /path/x_images.e" and replace your existing $iraf/bin.macintel/x_images.e binary.

Post back if you still have problems.

---

Mike Smutko wrote on Jun 26, 2013

Fixed! Thanks for the quick reply.

(I thought that "make latest" would keep me up to date with these kinds of patches, but I guess that's not the case?)

---

Mike Fitzpatrick wrote on Jun 26, 2013

(I thought that "make latest" would keep me up to date with these kinds of patches, but I guess that's not the case?)

A 'make latest' will update full-system patch releases (e.g. v2.16.1) or updates to the distribution files. While this change will[\b] be part of the next release now in preparation, at the moment it is just a single-file update to fix a specific bug. I'll see whether it's easy to make updates like this possible with a 'make latest' as well .....[/quote] </div> --- *Last post on Jun 26, 2013*