[daip] Re: clean on DECs

[Eric Greisen]

Carlos De Breuck writes:

 > I did a few more tests on the IMAGR problem I just emailed you about.
 > It does appear to be a problem with clean on the DEC alpha's running OSF
 > 4.0 or 5.1. Or it has to be with the 31DEC00 version. The same dataset
 > does clean on a Linux machine running AIPS 31DEC99 (over in Leiden).

     These are probably the same version of AIPS although perhaps at
slightly different times (31DEC99 was a name we used for a long time,
then we switched to 31DEC01 and froze it).

 > Could you check if you run into the same problem with the DEC(s) over in
 > Socorro? If so, would there be an easy way to patch this, or should I just
 > install Linux AIPS on my laptop? 

     We have no DECs in Socorro - there is an old one in CV which I
will try to run a test with, but CfA uses DECs and they would report
any problems.  The 31DEC00 is pretty old by now.  If you install on
Linux, be very careful to use the correct compiler - see our web
site:

            http://www.aoc.nrao.edu/aips/

 > As an aside, I ran into exactly the same problem when installing MIRIAD on
 > our DECs here. It also had problems compiling the clean routines in f90.
 > But this could be solved by splitting some variable declarations in their
 > code. I wonder if AIPS and MIRIAD use the same clean routines?
 > 

There is no relationship between MIRIAD code and ours.  DEC compilers
stick things in different places from other compilers and so reveal
bugs not seen other ways.

Note that I cannot debug 31DEC00 and fix anything.  I will check
31DEC02.  Note also that I am away 5/17-6/12 on vacation.

Eric