19991123: Running McIDAS and McIDAS-XCD on OSF/1 (cont.)

>From: Erick Lorenz (address@hidden) <address@hidden>
>Organization: UC Davis
>Keywords: 199910272007.OAA03748 McIDAS XCD

Erick,

>As of this morning the situation has much improved.  I think that the
>protection on SYSKEY.TAB was the culprit.  

Sounds promising.

>I am now able to create plots from both the Fkeys and MCGUI with the
>following puzzling exceptions:
>
>       I can plot upper air values on a map in Fkeys but not MCGUI
>               MCGUI just puts up the map and the caption but no data
>
>       I can plot cross sections with MCGUI but not Fkeys
>               Fkeys just gives me a DATA NOT AVAILABLE message in the
>               graphics window
>
>It seems that the data is there because one or the other utility
>can access it.
>
>I am doing all this testing directly on ATM25 which is also the ldm host.

I logged onto ATM25 and verified the failure of upper air plotting from
the MCGUI.  I looked at the command output in the MCGUI Console window
(pop this up by clicking mouse button 1 on the keyboard icon on the
GUI) and saw:

--MDX-- INVALID PARAMETER T
MDX: Program called abort
--MDX-- INVALID PARAMETER T
MDX: Program called abort

I then looked back through the McIDAS support email archives to see if
INVALID PARAMETER T was reported before, and I found that it was;
the answer I saw reclued me into the fact that IF the string RAOB exists,
it will be used as a specifier of the MD file from which to read the MD
file number when UPC is run.  So, I deleted RAOB from the stringtable:

TD RAOB

and then tried the GUI command again; it failed again.  I then decided
to start debugging the UPC command, since it is the one run out of
MCGUI for upper air plotting.

I did the following:

o run the command from the Observations menu and let it fail
o rerun the command from the GUI console; this is easily done since the
  commands run from the GUI are saved in a command stack.  In doing this
  run, I turned debugging on and this is what I saw:

UPC T 500 USA LATEST DAY=1999327 GRA=1 BLANK=NO LSIZE=6 SF=YES DEV=GCC
UPC* CPRO = CONF
UPC* CPARM1 =T
UPC* CPARM2 =T
UPC* CDER =
UPC* DAY1 after day check is =        99327
UPC* DAY2 after day check is =        99327
UPC* MD1 is =           27
UPC* MD2 is =           27
UPC* DAY1 after MD check is =      1999327
UPC* DAY2 after MD check is =      1999327
UPC* TIME1 after time check is =           12
UPC* TIME2 after time check is =           12
--MDX-- INVALID PARAMETER T
MDX: Program called abort
UPC* UPC: GRID Increment =       1.4525
UPC* UPC: SMOOTHing parameter =           61
UPC* Standard Latitudes: 60.00 60.0
UPC* PREFIX = @
UPC* CPARM = T
UPC* CDER =
--MDX-- INVALID PARAMETER T
MDX: Program called abort
UPC* Color used:  -2139062144
UPC* Color used:  -2139062144
UPC* Color used:  3

The problem being encountered was that the MD file that was being read
from was not the one for mandatory level upper air data (17), but,
rather, the one for significant level upper air data (27).  Since this
made no sense to me, I decided to add some additional debug statements
in ~mcidas/mcidas7.6/src/upc.mac and see why this mistake was being
made.  As soon as I recompiled UPC, it started working correctly even
after I removed the debug statements I had put in!  I wish I could
explain this, but I can't!!  Now, when I plot upper air data from the
MCGUI, it works correctly every time (at least while I was logged on).
Problems like this are exasperating to say the least.

I next decided to try to verify the cross section failure.  When I
brought up the RAOB Cross Section widget in MCGUI, I saw that the
latitude extents were set to be '40 40' and the longitude extents were
set to be '-125 -75'.  This means that the last person to run the cross
section plotter was looking for a slice that was located over Asia
somewhere; not a problem, simply an observation.  This ran fine.

I then brought up the Fkey menu from the Misc menu on the MCGUI and
tried the cross section plotting from it.  It also ran fine giving
the exact same display as the plotting from the MCGUI.

So, the problem now is why you were seeing the failure with the Fkey
menu for RAOB cross sections and I did not.  The fact that the RAOB
plots were also failing and now are working has me worried and,
quite frankly, confused.

>I am going to check out my findings on a Linux client now.

>From address@hidden  Tue Nov 23 13:46:22 1999

>I have tested McIDAS on two Linux machines and the behavior with
>respect to upper air plots and cross sections is exactly the same
>with the MCGUI,  

>       Upper air plots display nothing but a map and a caption
>       Cross sections work well.

>I haven't been able to compare the Fkey menu because I didn't want
>to reconfigure the display.

This also has me really worried.  I think that we need to remake upc.k
and see if simply rebuilding it solves the upper air plotting problem
like it did on the DEC.  I could do this if I had a login, or you can
do the following:

<login to one of the Linux machines as 'mcidas'>
cd mcidas7.6/src
touch upc.mac
make ucp.k
rm ~/bin/upc.k
ln upc.k ~/bin

<start a MCGUI session>
<retry the upper air plotting AFTER you do the following from the
MCGUI Command window:

TD RAOB

Please let me know what happens.

Tom

>From address@hidden  Wed Nov 24 12:47:27 1999

>I did as you suggested on ATM50 and the upper air plots started working.
>
>I then fired up ATM52 and tried an upper air plot without first trying
>your fix and it worked as well!
>
>It seems that the problem shows up sporadically.
>
>I need to test the other three Linux boxes but the problem with dual
>boot systems in a student lab is that they are in use all the time
>and usually in WinNT mode so I need to wait for a quiet time to continue.
>
>Thanks

>Erick Lorenz, LAWR, UCDavis