Variable declaration in extended types

[MrTAToad]

I've found that a program using a floating-point variable in an extended type will still compile even if the variable is access like an integer :

Code (glbasic) Select Expand

TYPE TTest
a=5.7

FUNCTION test%:
DEBUG self.a%+"\n"
ENDFUNCTION
ENDTYPE

LOCAL y AS TTest

y.test()

a is defined as a float, but is accessed like its an integer (and the result is a floating-point value).

Do it in a standard function

Code (glbasic) Select Expand

FUNCTION test%:
LOCAL a=5.7

DEBUG a%+"\n"
ENDFUNCTION

and you get a compiler error (which is correct).

[MrTAToad]

Thanks    Gernot's not going to be happy 

[Kitty Hello]

Oh, I'm so happy. Soo happy. ... if I wasn't to implement extended types at all :/

I'll have a close look, though.

[Moru]

Well you did say you would regret it, didn't you? :-)

[MrTAToad]

He did too

[bigsofty]

If it's any consolation, I initially disliked them but now I really enjoy them now!

[Moebius]

Not an extended types problem.  This compiles even with explicit decls. :

Code (glbasic) Select Expand

FOR i = 1 TO 10; NEXT;
I think variables in FOR/NEXT loops are defined temporarily.  This doesn't work:

Code (glbasic) Select Expand

FOR i = 1 TO 10; NEXT;
PRINT i, 10, 10

[MrTAToad]

Yes, I forgot about that - I'll delete my previous post 

Silly me...