Addressing a variable that does not exist causes confusion during a debug session

[brandy125]

When addressing any not existing variable during debugging of a program causes confusion. The only way to continue is to create that variable with any contents. Any way to resolve this?

pro var
    stop
    print,'ok'
end

GDL> var
% Compiled module: VAR.
% Stop encountered: VAR                  3 var.pro
GDL> print,variablethatdoesnotexist
% PRINT: Variable is undefined: VARIABLETHATDOESNOTEXIST
% Execution halted at: VAR                  3 var.pro
%                      $MAIN$          
GDL> .c
% PRINT: Variable is undefined: VARIABLETHATDOESNOTEXIST
% Execution halted at: VAR                  3 var.pro
%                      $MAIN$          
GDL> 

[ChunkyPanda03]

I just tested this on idl 6.3 and I can confirm that this is supposed to continue

IDL> var
% Stop encountered: VAR                 2  var.pro
IDL> print,variablethatdoesnotexist
% PRINT: Variable is undefined: VARIABLETHATDOESNOTEXIST.
% Execution halted at: VAR                 2  var.pro
%                      $MAIN$          
IDL> .c
ok

interesting idl states that it stopped on line 2 but GDL reports stopping on line 3 is this related? note .c = .continue I didn't know this before

[GillesDuvert]

Yes gdl is always "at the next line" . Dunno if this is unavoidable or if it could be patched. I think when trapped in an unvoluntary error loop as you report, one can do a stepover .so and then continue as usual.

[GillesDuvert]

Btw thanks for reporting, even if this issue will likely be not worked on soon.

[brandy125]

Today I saw that there is another problem related to this when the procedure name was misspelled like "prin" instead of "print":

This can be overcome by a ".skip" but it is boring. Imagine a long debug session where a "wrong" ".skip" could cause confusion in the debug session