Author Topic: AfxWin - a start  (Read 13337 times)

0 Members and 4 Guests are viewing this topic.

JRS

  • Guest
Re: AfxWin - a start
« Reply #45 on: October 12, 2018, 04:40:06 PM »
Thanks for the refresher course.  :)

JRS

  • Guest
Re: AfxWin - a start
« Reply #46 on: October 12, 2018, 05:52:11 PM »
I've always wondered what should come next, documentation or error checking?

The resent most online numbers for the O2 forum remind me of good old days on the PowerBASIC forum. I think José's effort can account for most of the increased activity.
« Last Edit: October 12, 2018, 06:55:08 PM by John »

Charles Pegge

  • Guest
Re: AfxWin - a start
« Reply #47 on: October 13, 2018, 01:06:32 AM »
I've fixed error trapping for undefined type parameters in C style headers.

Also L"..." string literals and UTF-8/UTF-16 source code are now supported :). Quite complex.

José Roca

  • Guest
Re: AfxWin - a start
« Reply #48 on: October 13, 2018, 01:39:15 AM »
These are great news.

Regarding COM, writing prototyped headers for the whole COM subsystem would be too much work and very difficult because they are very interwitned. For example the IDispatch interface references the VARIANT data type, and the VARIANT data type references the IDispatch interface.

Is there a way in O2 to solve it?

Charles Pegge

  • Guest
Re: AfxWin - a start
« Reply #49 on: October 13, 2018, 02:45:54 AM »
If they are only references, there is no problem with recursive definitions.

And you can use a dummy type to break circularity.

Code: [Select]
type yy{} 'dummy type

type xx
  yy*a
end type

type yy
  xx*b
end type


This solves the definitions problem. Gaining access to recursive members has to be done indirectly though.