[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference iamok::dtrdig

Title:DATATRIEVE INTEREST GROUP
Notice:ADD KEYWORDS TO NOTES FOR EASY SEARCHES
Moderator:IMTDEV::KRATZER
Created:Fri Mar 21 1986
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:3011
Total number of notes:9337

2989.0. "MTH-F-FLOOVEMAT floating overflow in math library" by CSLALL::mko-ras-port-3.mko.dec.com::JKeenan (Jay Keenan (603) 883-7913) Tue Feb 18 1997 12:59

I have a set of procedures that has been running for about 8 years. 
The routines that create the data file are unchanged as are the 
reports that run off it. Out of nowhere I'm receiving a 

%MTH-F-FLOOVEMAT, floating overflow in math library
  user PC 00346E0F

error message. Does anyone have any idea what might cause this? This 
is on DTR 6.1, VMS 6.1

Thanks,

Jay Keenan
T.RTitleUserPersonal
Name
DateLines
2989.1CSC32::HOEPNERA closed mouth gathers no feetTue Feb 18 1997 21:2533
    
    
    
    From the 7.0 DTR release notes:
    

Publication Date:     March 1995


2.3 Error Converting D_FLOATING Values to Integer Values

           In Version 6.1 a problem occurred when converting a
           D_FLOATING value to an integer value. The following example
           illustrates the problem:

           DTR> DECLARE I LONG. ;
           DTR> DECLARE D D_FLOATING. ;
           DTR> D = 1E-30 ;
           DTR> I = D ;
           Variable "I" may contain an incorrect value due to error
           during assignment.
           %MTH-F-FLOOVEMAT, floating overflow in math library user PC 00321023

           In order to keep all the significant digits, DEC DATATRIEVE
           used a scale factor which created an integer value
           exceeding the D_ FLOATING range (in the example the scale
           factor is -47).

           In DEC DATATRIEVE Version 7.0 this problem has been fixed
           and the conversion of a D_FLOATING value to an integer
           value does not cause any problems.