IRAF 2.15 and gkimos problem

Robert Janusz wrote on Nov 29, 2010

In the new 2.15 IRAF I can run

gkimos ... nx=1 ny=1

without problem, but for example

gkimos ... nx=3 ny=3

gives a message:

ERROR: floating point divide by zero
     called as: `gkimosaic...'

.
After the error is reported I can run

gkimos ... nx=1 ny=1
... nx=1 ny=2
... nx=2 ny=1
...
... nx=3 ny=3 # !!!

Sometimes IRAF does not report the error but hungs.

---

Mike Fitzpatrick wrote on Nov 29, 2010

Please tell me what platform you're on, I can use the standard dev$vdm.gki file or one created locally on a linux64 system with no problem. If you need to, you can also post the metacode file (in binary mode) to the anonftp on ftp://iraf.noao.edu/pub

---

Robert Janusz wrote on Nov 29, 2010

I put the "z.gki" small file. I run new 2.15 IRAF on 32-bit laptop. The problem is

unl gkimos
gkimos z.gki

with error

ERROR: floating point divide by zero

and on a starting xterm (which lunches the xgterm)

xgterm: warning, error event received:
X Error:  BadDrawable
  Request Major code 53 ()
  ResourceID 0x4400013
  Error Serial #1575
  Current Serial #1606

gkimos z.gki nx=1 ny=1

is running without any problem.

---

Mike Fitzpatrick wrote on Nov 29, 2010

I can reproduce the problem on a 32-bit linux system. It actually appears to be a bug in either the CL itself or the graphics system having to do with converting the normalized device coords to a plot WCS. In gkimos some areas of the screen don't have a plot displayed so something may be uninitialized, this would also explain why when nx=ny=1 it works.

I'll post back when I have a fix or workaround, thanks for the report.

---

Robert Janusz wrote on Nov 29, 2010

IRAF 2.15.1a continues to have this problem.

---

Mike Fitzpatrick wrote on Nov 29, 2010

It is still on the list, it just didn't make it into this patch.

---

Last post on Nov 29, 2010