PBWin 9 and Point

[Peter Lameijn]

A lot of my applications hold a variable called 'Point'.
For some reason PB9 seems to object to that, although Point isn't mentioned as a reserved word...
Did I miss something? (I get: "Duplicate name definition: POINT")

#Compile Exe
#Dim All

Function PBMain()
  Local Point As Long
End Function

[José Roca]

POINT is a built-in structure in the new compilers.

[Peter Lameijn]

I didn't find that anywhere in the helpfile...

[Michael Mattias]

> didn't find that anywhere in the helpfile...

Send those bug reports to [email protected]

[Rodney Hicks]

A list of reserved words here:
http://www.powerbasic.com/support/pb...light=reserved

[Walt Thompson]

Hi Peter;

FWIW when I run into this problem, my solution is to change the spelling of the word. changing Point to Pointe will, probably, solve the problem. How practical this suggestion depends on how many applications and how often Point is used.

I agree that you should report this to PB support. Hopefully, the documentation can be updated in time for the release of 9.02.

[Cliff Nichols]

I could not find POINT documented, so I started searching the Win32Api.inc (Thinking that I could SWEAR that I had seen it documented before, but maybe I was thinking of POINTAPI???)

Anyways while searching I found many references that I also can not find documented such as

aptl(0 TO 0) AS POINTL  ' array size may vary (fake type)

A variable array size within a TYPE??? huh

I found it in

TYPE EMRPOLYBEZIER16
  pEmr AS emr
  rclBounds AS RECTL      ' Inclusive-inclusive bounds in device units
  cpts AS DWORD
  aptl(0 TO 0) AS POINTL  ' array size may vary (fake type)
END TYPE

and many other "Fake Types"
Still can not find POINT but still looking as I KNOOOOWWWwwww I saw it somewhere????

[José Roca]

POINT

The POINT structure defines the x- and y- coordinates of a point.

TYPE POINT
   x AS LONG
   y AS LONG
END TYPE

Members

x Specifies the x-coordinate of the point.

y Specifies the y-coordinate of the point.

[José Roca]

Anyways while searching I found many references that I also can not find documented such as

aptl(0 TO 0) AS POINTL  ' array size may vary (fake type)

The help file doesn't document the Windows API. Search in the Microsoft documentation.

A variable array size within a TYPE??? huh :confused2:

Yes. PB doesn't support it, so you will need to make your own declaration with aptl dimensioned with as many elements you will need.

[Michael Mattias]

>A variable array size within a TYPE??? huh

The EMRPOLYLINE16, EMRPOLYBEZIER16, EMRPOLYGON16, EMRPOLYBEZIERTO16, and EMRPOLYLINETO16 structures contain members for the Polyline, PolyBezier, Polygon, PolyBezierTo, and PolylineTo enhanced metafile records.

typedef struct tagEMRPOLYLINE {
  EMR    emr; 
  RECTL  rclBounds; 
  DWORD  cpts; 
  POINTS apts[1]; 
} EMRPOLYLINE16,     *PEMRPOLYLINE16,
  EMRPOLYBEZIER16,   *PEMRPOLYBEZIER16,
  EMRPOLYGON16,      *PEMRPOLYGON16,
  EMRPOLYBEZIERTO16, *PEMRPOLYBEZIERTO16,
  EMRPOLYLINETO16

...
apts
Array of 16-bit points, in logical units.

[Cliff Nichols]

Well now that I know a variable array can be used in a type (Although Unsupported) I have yet another thing to test for in a different project I have been working on.

Just in the case it is used (Again Although NOT supported, and IMO not to be used until it is supported) it is a check I should perform.

[José Roca]

The use of DIMENSIONED arrays in types IS SUPPORTED, what is not supported is the use of dynamic arrays.

[Michael Mattias]

>Well now that I know a variable array can be used in a type

A [user-defined] TYPE (upper case, intentional) is simply a source-code level tool provided by the compiler. You can make anything into a "TYPE" of some kind.

Heck, you can make a long integer into a type:

TYPE MyLong
   LOWORD  AS WORD
   HIWORD   AS INTEGER
END TYPE

You could use this if you had, say, some kind of phobia about using LO(WORD, longVar&) or HI(INTEGER, longVar&)

MCM