>>>>> On 07 Nov 2000 13:55:47 -0600, Rob Browning <[EMAIL PROTECTED]> said:
Rob> Robert Graham Merkel <[EMAIL PROTECTED]> writes:
>> Yep, that makes sense.
>>
>> However, one minor concern, particularly if you want to wrap parts
>> of glib itself. A lot of those functions are access functions
>> that take a pointer to a glib container and return a void * (a
>> gpointer, actually) that is a pointer to an entry in the
>> container.
>>
>> Can we pass the returned gpointer to functions that take, say,
>> Account *?
Rob> After talking with Bill, I'm going to implement
Rob> gwp:pointer-token-coerce that'll take a pointer-token and
Rob> produce a new pointer-token that's of the requested type. i.e.:
Rob> (gwp:pointer-token-coerce pt 'Account*)
Rob> This will be used to implement things like
Rob> (gwp:glib-list->list foo 'Account*)
Rob> OK. It seems like we have agreement. I'll get on it.
Rob> Shouldn't take too long.
Rob> I still have to deal with your enum patch, but I'll do that
Rob> while working on this.
The only thing that concerns me about this is you are introducing into
scheme an easy way to cause a segfault. Maybe implement it this way
in the short term because it's the easiest and lets us get onto other
more important things quickly, but this is a bad long term solution
imo.
Jim
--
@James LewisMoss <[EMAIL PROTECTED]> | Blessed Be!
@ http://jimdres.home.mindspring.com | Linux is kewl!
@"Argue for your limitations and sure enough, they're yours." Bach
_______________________________________________
gnucash-devel mailing list
[EMAIL PROTECTED]
http://www.gnumatic.com/cgi-bin/mailman/listinfo/gnucash-devel
