Error with Constant comparison in SELECT CASE?

[daro]

I'm finding this strange error when using a constant inside of a select case, any ideas why?
I've tried many different ways but it seems to be the use of a CONST inside of the SELECT CASE giving the error.

Expected string expression on line 12
Caused by (or after) CASE "A", 1 , "a",1 , KEY_ARROW_LEFT

Code: QB64: [Select]

1. Const KEY_ARROW_LEFT = Chr$(0) + Chr$(72)
2. Const KEY_ARROW_UP = Chr$(0) + Chr$(72)
3. Const KEY_ARROW_RIGHT = Chr$(0) + Chr$(77)
4. Const KEY_ARROW_DOWN = Chr$(0) + Chr$(80)
5. Const KEY_PAGE_UP = Chr$(0) + Chr$(73)
6. Const KEY_PAGE_DOWN = Chr$(0) + Chr$(81)
7.  
8. Do
9.  
10.     keyp$ = InKey$
11.     Select Case keyp$
12.         Case "A", "a", KEY_ARROW_LEFT
13.             Print " Left ";
14.         Case "W", "w", KEY_ARROW_UP
15.             Print " Up   ";
16.         Case "D", "d" , KEY_ARROW_RIGHT
17.             Print " Right";
18.         Case "S", "s", KEY_ARROW_DOWN
19.             Print " Down ";
20.         Case "Z", "z", KEY_PAGE_DOWN
21.             Print " P.Up ";
22.         Case "X","x", KEY_PAGE_UP
23.             Print "P.Down ";
24.         Case KEY_ESCAPE
25.             Exit Do
26.     End Select
27.  
28. Loop
29. End

[FellippeHeitor]

@daro Looks indeed like a bug. We'll have a look.

In the meantime, make those DIM SHARED AS STRING instead of CONSTs, for the same expected effect:

Code: QB64: [Select]

1. DIM SHARED AS STRING KEY_ARROW_LEFT, KEY_ARROW_UP, KEY_ARROW_RIGHT, KEY_ARROW_DOWN, KEY_PAGE_UP, KEY_PAGE_DOWN
2. KEY_ARROW_LEFT = Chr$(0) + Chr$(72)
3. KEY_ARROW_UP = Chr$(0) + Chr$(72)
4. KEY_ARROW_RIGHT = Chr$(0) + Chr$(77)
5. KEY_ARROW_DOWN = Chr$(0) + Chr$(80)
6. KEY_PAGE_UP = Chr$(0) + Chr$(73)
7. KEY_PAGE_DOWN = Chr$(0) + Chr$(81)
8.  

Welcome aboard.

[Cobalt]

I would recommend using _KEYHIT or _KEYDOWN over INKEY$ and changing that SELECT CASE into IF THENs.


@FellippeHeitor
Doesn't CONST have issues when using variables as assignments?
Wasn't there something not that long ago about using a variable to assign constant?
something like:
A%=15-3
CONST VALUE = A%

I can't quite remember now. not having any luck with the search either, but I could swear we've had something like this before.

[bplus]

I would recommend using _KEYHIT or _KEYDOWN over INKEY$ and changing that SELECT CASE into IF THENs.


@FellippeHeitor
Doesn't CONST have issues when using variables as assignments?
Wasn't there something not that long ago about using a variable to assign constant?
something like:
A%=15-3
CONST VALUE = A%

I can't quite remember now. not having any luck with the search either, but I could swear we've had something like this before.

We've had trouble using Functions and doing math but I think it's been fixed. Maybe not all string stuff is, I think we are using Steve Math Evaluator?

[FellippeHeitor]

Turns out the CONST evaluator was never meant to be used with string functions, like Chr$, Space$, etc. - but Chr$ had been left there as valid - just a slip. The latest dev build generated today fixes that by not allowing string functions in CONSTs. This is a list of the functions that can be used in CONSTs. The wiki page for CONST is getting updated too.

Code: QB64: [Select]

1. _PI, _ACOS, _ASIN, _ARCSEC, _ARCCSC, _ARCCOT,
2. _SECH, _CSCH, _COTH, COS, SIN, TAN, LOG, EXP, ATN, _D2R,
3. _D2G, _R2D, _R2G, _G2D, _G2R, ABS, SGN, INT, _ROUND, _CEIL,
4. FIX, _SEC, _CSC, _COT, ASC, _RGB32, _RGBA32, _RGB, _RGBA,
5. _RED32, _GREEN32, _BLUE32, _ALPHA32, _RED, _GREEN, _BLUE,
6. _ALPHA

[FellippeHeitor]

I would recommend using _KEYHIT or _KEYDOWN over INKEY$ and changing that SELECT CASE into IF THENs.


@FellippeHeitor
Doesn't CONST have issues when using variables as assignments?
Wasn't there something not that long ago about using a variable to assign constant?
something like:
A%=15-3
CONST VALUE = A%

I can't quite remember now. not having any luck with the search either, but I could swear we've had something like this before.

It was never really allowed with variables, but we did have a problem recently (fixed in 1.5) that would cause issues when referencing existing CONSTs in a new CONST assignment.

[bplus]

edit: OP wanted strings not numbers