Select + Case + Default + EndSelect

[4RESTER]

What wrong in this code?

Code: Select all

Global BPB_Media.B = $F8

Select BPB_Media.B
    Case $F0,$F8 To $FF
        Debug "BPB_Media: $"+RSet(Hex(BPB_Media, #PB_Byte),2,"0") 
    Default
        Debug "Wrong BPB_Media: $"+RSet(Hex(BPB_Media, #PB_Byte),2,"0")
EndSelect

In fact, with such value
BPB_Media.B = $F8
the branch of Default shall not be executed.

[gnasen]

no, its correct, because in PB most variables are signed:
Byte: -128 to +127
So if you store $F8 = 248 in a byte, you get: -8. Use this instead:
Char (= unsigned Byte): 0 to 255

Code: Select all

Global BPB_Media.c = $F8

Select BPB_Media.c
  Case $F0,$F8 To $FF
    Debug "BPB_Media: $"+RSet(Hex(BPB_Media, #PB_Byte),2,"0")
  Default
    Debug "Wrong BPB_Media: $"+RSet(Hex(BPB_Media, #PB_Byte),2,"0")
EndSelect

[netmaestro]

.c is 16 bits in unicode mode, better use .a for unsigned byte.

[4RESTER]

no, its correct, because in PB most variables are signed:
Byte: -128 to +127
So if you store $F8 = 248 in a byte, you get: -8. Use this instead:

Code: Select all

Global BPB_Media.c = $F8

Select BPB_Media.c
  Case $F0,$F8 To $FF
    Debug "BPB_Media: $"+RSet(Hex(BPB_Media, #PB_Byte),2,"0")
  Default
    Debug "Wrong BPB_Media: $"+RSet(Hex(BPB_Media, #PB_Byte),2,"0")
EndSelect

Hmmm, very strange behavior of the compiler.
So literal $F8 must be also be as -8

[Demivec]

What wrong in this code?

Code: Select all

Global BPB_Media.B = $F8

Select BPB_Media.B
    Case $F0,$F8 To $FF
        Debug "BPB_Media: $"+RSet(Hex(BPB_Media, #PB_Byte),2,"0") 
    Default
        Debug "Wrong BPB_Media: $"+RSet(Hex(BPB_Media, #PB_Byte),2,"0")
EndSelect

In fact, with such value
BPB_Media.B = $F8
the branch of Default shall not be executed.

Same as the posting in your other thread. When the hexadecimal literals are converted to decimal it would read this way:

Code: Select all

Global BPB_Media.B = -8 ;$F8

Select BPB_Media.B
  Case 240, 248 To 255 ;$F0,$F8 To $FF
    Debug "BPB_Media: $"+RSet(Hex(BPB_Media, #PB_Byte),2,"0") 
  Default
    Debug "Wrong BPB_Media: $"+RSet(Hex(BPB_Media, #PB_Byte),2,"0")
EndSelect

Which clearly shows the default case will be executed.

Use this code to prove it to yourself:

Code: Select all

Global BPB_Media.B = $F8

Select BPB_Media.B
  Case $F0,$F8 To $FF
    Debug "BPB_Media: $"+RSet(Hex(BPB_Media, #PB_Byte),2,"0") 
  Case -8
    Debug "Right BPB_Media: " + Str(BPB_Media)
  Default
    Debug "Wrong BPB_Media: $"+RSet(Hex(BPB_Media, #PB_Byte),2,"0")
EndSelect

[gnasen]

.c is 16 bits in unicode mode, better use .a for unsigned byte.

Ah thanks, I missed that. I never used Unicode until now

[4RESTER]

What wrong in this code?

Code: Select all

Global BPB_Media.B = $F8

Select BPB_Media.B
    Case $F0,$F8 To $FF
        Debug "BPB_Media: $"+RSet(Hex(BPB_Media, #PB_Byte),2,"0") 
    Default
        Debug "Wrong BPB_Media: $"+RSet(Hex(BPB_Media, #PB_Byte),2,"0")
EndSelect

In fact, with such value
BPB_Media.B = $F8
the branch of Default shall not be executed.

Same as the posting in your other thread. When the hexadecimal literals are converted to decimal it would read this way:

Code: Select all

Global BPB_Media.B = -8 ;$F8

Select BPB_Media.B
  Case 240, 248 To 255 ;$F0,$F8 To $FF
    Debug "BPB_Media: $"+RSet(Hex(BPB_Media, #PB_Byte),2,"0") 
  Default
    Debug "Wrong BPB_Media: $"+RSet(Hex(BPB_Media, #PB_Byte),2,"0")
EndSelect

Which clearly shows the default case will be executed.

Use this code to prove it to yourself:

Code: Select all

Global BPB_Media.B = $F8

Select BPB_Media.B
  Case $F0,$F8 To $FF
    Debug "BPB_Media: $"+RSet(Hex(BPB_Media, #PB_Byte),2,"0") 
  Case -8
    Debug "Right BPB_Media: " + Str(BPB_Media)
  Default
    Debug "Wrong BPB_Media: $"+RSet(Hex(BPB_Media, #PB_Byte),2,"0")
EndSelect

Looks like this dialect of BASIC is a finding for CodeTalkers (NavajoBasic).

[4RESTER]

.c is 16 bits in unicode mode, better use .a for unsigned byte.

Okay.
8-bits unsigned as .A
16-bits unsigned as .U

What about 32-bits UNSIGNED values?

[luis]

They have been requested (see in the appropriate subforum) but the official answer for the team, if I recall correctly (please correct me if I'm wrong), is the .a and .u were added as a convenience for working with ascii/unicode, and no larger unsigned data type are planned.

Obviously is possible to work with signed integers just the same, but I would like a 32 byte unsigned type too, for what it's worth. Especially to work with c/c++ libraries in a simpler way.

edit: found the post I was mentioning
http://www.purebasic.fr/english/viewtop ... 80#p295680