clean install of V2.15.1a fails to solve splot prob.

[massey]

Hi, Mike et al.

I just recently upgraded my Macbook Pro to Lion. After a clean re-install of V2.15.1a using the "iraf.macx.x86_64" version on the downloads section of this site, I am getting pixels out of bounds error messages when I try to plot a spectrum using splot. Here are some details:
a) imhead on the image shows
295.J004056.72+410255.ms.fits[4543,1,3][real]: J004056.72+410255
i.e., it's one dimensional but had 3 bands.
b) Here's the exact error message:

splot 295.J004056.72+410255.ms.fits
Image band to plot (1 (1): [okay, it's funny that come out as a funny face. It's a parenthesis 1 colon end parenthesis obviously...]
ERROR: Pixel subscript out of bounds (295.J004056.72+410255.ms.fits)
I did an unlearn splot to make sure things were okay.

These spectra plotted fine before the upgrade to lion; i.e., when I was running
32-bit IRAF.

I know there's an error logged about this, and the suggestion is to reinstall.
But I did that and it hasn't helped. Could it be that the "iraf.macx.x86_64"
on the download site isn't updated?

Most other things work fine; i.e., I can do an implot dev$pix and display dev$pix and all works as expected.

thanks!
---phil

[massey]

Indeed, I think the file in the downloads area is bad. (It doesn't even have .tar at the end, although you can untar it like it was a tar file.) I found a version of the iraf.macx.x86_64.tar on another machine
of mine, installed that, and now splot works fine.

In summary, I think there's something screwy with the file labeled iraf.macx.x86_64 near the end of the list on the downloads page.

[fitz]

Hi Phil,

The .gz filenames are actually symlinks to the .tar.gz files, so even without the extension you get the same file either way. It may just be that the file you had locally that now works *is* the newer file and the original one was corrupted or didn't unpack correctly????

There've been a number of 64-bit issues fixed recently but it would help if you could upload a FITS file to ftp://iraf.noao.edu/pub so I can try to reproduce the problem. If the data are integers with bscale/bzero header values then I think this problem is already fixed. Another thing that might help is if you could post/send me an "ls -l" output of the x_onedspec.e binary so I can figure out exactly which version you might be using.

Cheers,
-Mike

[massey]

Hi, Mike---

I couldn't connect to the iraf ftp site, but I stuck an example
fits file at http://www.lowell.edu/users/massey/253.J004190.59+404856.ms.fits

It works fine with the 64-bit version I downloaded in July (2.15.1a), but not what I picked
up the other day at the bottom of http://iraf.net/downloads/ page, which
also says 2.15.1a.

You're right about the .gz extension in the download; apparently my downloader now un-g'zip things automatically and I'm left with a file with no extension,
which confused me.

I've just downloaded it from the website again, ran install, and confirmed it fails on this file.
kpnoslit> splot 253*
Image band to plot (1 (1):
ERROR: Pixel subscript out of bounds (253.J004109.59+404856.ms.fits)
kpnoslit>

With the version that dosn't work, if I go to bin.macintel and do an ls -l I see:
-rwxr-xr-x@ 1 312 112 1725872 Feb 21 2011 x_color.e*
-rwxr-xr-x@ 1 312 112 2778480 Feb 21 2011 x_dataio.e*
-rwxr-xr-x@ 1 312 112 6582856 Feb 21 2011 x_images.e*
-rwxr-xr-x@ 1 312 112 1890104 Feb 21 2011 x_imdkern.e*
-rwxr-xr-x@ 1 312 112 1785720 Feb 21 2011 x_lists.e*
-rwxr-xr-x@ 1 312 112 485576 Feb 21 2011 x_mkttydata.e*
-rwxr-xr-x@ 1 312 112 2036536 Feb 21 2011 x_ncar.e*
-rwxr-xr-x@ 1 312 112 4011528 Feb 21 2011 x_nttools.e*
-rwxr-xr-x@ 1 312 112 2607224 Feb 21 2011 x_obsolete.e*
-rwxr-xr-x@ 1 312 112 2351920 Feb 21 2011 x_plot.e*
-rwxr-xr-x@ 1 312 112 2843672 Feb 21 2011 x_proto.e*
-rwxr-xr-x@ 1 312 112 595912 Feb 21 2011 x_sgikern.e*
-rwxr-xr-x@ 1 312 112 490608 Feb 21 2011 x_softools.e*
-rwxr-xr-x@ 1 312 112 659448 Feb 21 2011 x_stdgraph.e*
-rwxr-xr-x@ 1 312 112 1057088 Feb 21 2011 x_system.e*
-rwxr-xr-x@ 1 312 112 3539928 Feb 21 2011 x_tv.e*
-rwxr-xr-x@ 1 312 112 2429256 Feb 21 2011 x_utilities.e*
-rwxr-xr-x@ 1 312 112 1926936 Feb 21 2011 x_vol.e*

Oddly, if I go to ondspec$ and do an ls -l I see:
-rw-r--r--@ 1 312 112 1000 Jun 21 2004 x_onedspec.x

which doesn't make sense to me; I must be looking in the wrong place?
I see exactly the same thing with the version that DOES work.

So, I'm looking at the wrong place. This is very repeatable (I've installed
each version multiple times to try this), so I repeat that I think if you go
to the downloads area and install the iraf.macx.x86_64.gz file rnear the
bottom, you'll see the same problem I'm seeing.

[fitz]

I get a 404 error trying to access the image URL.

The binary to look at is noao$bin.macintel/x_onedspec.e

[massey]

Sorry: my bad:
http://www.lowell.edu/users/massey/253.J004109.59+404856.ms.fits

[massey]

Mike, the version of x_onedspec.e that WORKS has a date of Mar 25 2011 and is derived from a tar file iraf.macx.x86_64.tar that I found somewhere.
The version that fails has a date of Feb 21, 2011. That is the one that you are linked to on the downloads page. So, I think all you have to do is change the downloads page link to the updated version and all is well.

[fitz]

I see what happened: The iraf.noao.edu archives were updated, but the download page points to a local iraf.net copy that uses the earlier files. I'm updating the files now and can confirm that your image works fine on macintel with the latest dev code.

Thanks for the heads-up,
-Mike