Strange compile messages:

Posted by: neilrh

A/P

Jackson, MI

Strange compile messages:
has no ratings.

Published: 10 Aug 2012
Revised: 23 Jan 2013 - 4109 days ago
Last viewed on: 23 Apr 2024 (4867 views)

Main URL:

Strange compile messages: Published by: neilrh on 10 Aug 2012 view comments(3)

061900          if CusNr = 17570000 and Group = '291' or
  :
063900             CusNr = 13057710 and Group = '001';
063901             CusNr = 00077837 or                
063902             CusNr = 00042059 or                
063903             CusNr = 00071625 or                
  :
063908             CusNr = 00079357 and                  
063909               %scan(Group : '001 002 021') <> 0 or
063910             CusNr = 00077787 and                  
063911               %scan(Group : '087 329') <> 0;      


*RNF7416 30        063909  The types of the right and left hand side do not match in 
                           the EVAL operation.

Took me ages to figure out that I'd not changed the semi-colon at line 63900 to "or", so that the If statement was extended by the new conditions. But really? The compiler decided that the error was on the %scan line, that was part of the unintentional eval that began at 63901 and extends to 63911?? Ok, I finally understood what the error was when I paged back far enough to find the semi colon, but could we report either the first occurrence of bad logic (or last), just not one in the middle!!

Sort Ascend | Descend

COMMENTS

(Sign in to Post a Comment)

Posted by: bobcozzi

Site Admin ****

Chicagoland

www.cozTools.com/qryfile

Comment on: Strange compile messages:
Posted: 11 years 8 months 14 days 22 minutes ago

Neil, this too was discussed during our training at your location. I actually poke fun at the compile. I've got to get more interesting when I talk about this stuff--to keep you awake.

Posted by: neilrh

Premium member *

Jackson, MI

Comment on: Strange compile messages:
Posted: 11 years 8 months 14 days 8 minutes ago

Yeah, I got used to the compiler complaining the line below the bad statement, and I remember that part of the talk - sometimes you have an error and the compiler decides it's the next line that is really in error.

I'm just bemused that the compiler also seems to have a random number generator attached to the "pick which line to report the error for" routine. I looked back at the code and there is a similar "CusNr = AND %scan" on lines 63906/63907. Why pick 63909? It's more logical to pick 63902, the first occurrence of generating *on/*off to eval into a numeric field, or maybe 63911 the last line of the eval.

Posted by: bobcozzi

Site Admin ****

Chicagoland

www.cozTools.com/qryfile

Comment on: Strange compile messages:
Posted: 11 years 8 months 13 days 22 hours ago

Ah! But I do see the pattern and understand why they did it that way--I don't agree with their decision, but I see why.