19991122: quoting apostrophes for shell script files (cont.)

>From: Robert Mullenax <address@hidden>
>Organization: NMSU/NSBF
>Keywords: 199911192037.NAA23504 McIDAS shell scripts

Robert,

>I was able to get it to read the $WWW_tim by removing the quotes around
>EOF and was able to generate a thivkness chart for Texas.

OK, good.

>However,
>when I tried to do on for Sweden (overlaying a PIX map) I get a blank
>chart.  Looking at my CGI log I see a segmentation violation.   Running
>it from a MciDAS session I either it as well.  It happens when I
>specify a map that has any portion in the eastern hemisphere.  In other
>words:
>
>MAP OUTLHRES LAT=20 80 LON=0 110                                              
>MAP: completed frame  1                                                       
>                          
>GRDDISP RTGRIDS/AVN TIME=12 G1='PARAM Z;LEV 500;FHOUR 12' G2='LEV 1000' 
>MATH='G1-G2' NAV=C CINT=60     
>GRDDISP: Done with graphic frame 1                                            
>GRDDISP - done  

OK, I verified that this one works.

>But look at this:
>MAP OUTLHRES LAT=20 80 LON=-1 110                                            
>MAP: completed frame  1                                                       
>                          
>GRDDISP RTGRIDS/AVN TIME=12 G1='PARAM Z;LEV 500;FHOUR 12' G2='LEV 1000' 
>MATH='G1-G2' NAV=C CINT=60     
>Program terminated, segmentation violation                                    
>                          
>Program terminated, address out of range                                      

OK, I verified that this fails.

>I have been producing maps for Sweden and have not seen this problem
>except here where the math function is used.  Is this related to the
>old SC 5.0 Sparc problem..when I remade McIDAS after the last bugfix I
>did not go back and recompile mcgget.for without optimization.  Did I
>need to? I have seen no other problems with GRDDISP.

Yes, you do need to, but I ran a copy of GRDDISP linked against a mcgget.for
compiled with no optimization, so the problem is likely to be somewhere
else.

>>From address@hidden  Mon Nov 22 15:11:56 1999
>
>I did some testing.  I verified that I can display grids and compute
>them (vorticity) but I cannot do s grid subtraction east of Greenwich.
>If I try to produce thickness on a map that has no area west of
>Greenwich, I get "Failed to get data from server" instead of
>segmentation error.

This sounds like a problem that needs to be reported to SSEC; I will do
so.

>
>>From address@hidden  Mon Nov 22 15:32:07 1999
>
>I get the same failer to get data from server message on
>our x86 machine in New Mexico.

OK, thanks for adding this information.  I am creating an inquiry
at SSEC about this right now.

Sorry for the problems.

Tom