>If your file is opened, then comparing it to an empty string is what will
>cause an invalid data type error.
>WHY!!!
>If the datatypes don't match, then the result of the comparison is
>FALSE, not INVALID.
>It is logically correct to do such a comparison. The result I am looking
>for is the logically correct response. It SHOULD NOT crash the program!
>And it worked fine on INFORMATION, why not UV?
It is a bad idea to allow it, it encourages bad programming practices and it
can hide a bug in your program for example if you do:
READ CAR FROM CARS ELSE CAR = ''
IF CARS # '' THEN
Here what was intended was to compare CAR to '' not CARS but having the S
there is a typo.
If we don't allow datatypes of different types being compared this bug has a
greater chance of getting caught during development because the program will
fail on comparing the file to a null string.
If you use FILEINFO it shows that you know that the variable you are testing is
a file.
I've seen a similar case where a subroutine was expecting:
SUBROUTINE ABC (MAT A, MAT B, MAT C)
A programmer only needed the contents of the matrix B and C to call it and did:
CALL ABC("",MAT B, MAT C)
Only thing Matrix ABC puts something in MAT A and since the programmer passed
an empty string what happens is that the content of Matrix B is over-written
causing data corruption.
It would probably have been a good idea to detect that since a Matrix was
expected, it shouldn't be legal to pass a dynamic array to the subroutine.
Most modern OOP programming languages don't allow for types of different kinds
to be compared to one another without some way of casting them.
_______________________________________________
U2-Users mailing list
U2-Users@listserver.u2ug.org
http://listserver.u2ug.org/mailman/listinfo/u2-users
