From: "Ike" <drone AT ismi DOT net>
Newsgroups: comp.os.msdos.djgpp
References: <3 DOT 0 DOT 32 DOT 19990801181614 DOT 007b9130 AT mail DOT colba DOT net> <37a5e4a8 AT news DOT ismi DOT net> <Z3np3.154$z4 DOT 3489 AT news020 DOT image DOT dk>
Subject: Re: BITMAP pointers w/ Allegro.
Date: Mon, 2 Aug 1999 19:58:26 -0700
Lines: 29
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 5.00.2014.211
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2014.211
NNTP-Posting-Host: tc4-20.ismi.net
Message-ID: <37a6305b@news.ismi.net>
X-Trace: 2 Aug 1999 19:57:15 -0400, tc4-20.ismi.net
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Reply-To: djgpp AT delorie DOT com


> > I've got a question about that,  I was asking around on IRC about this
> > problem and they said that its improper programming practice (wow that's
> > hard to say) to have the return value of a function be a pointer.  They
> said
> > that it violates the law of encapsulation.  But I've noticed that the
> > Allegro library does it all the time with BITMAP structures,  is that
> bad?
>
> You (or should I say they) just can't generalize like that - it depends on
> the situation. In Allegro's case I can't remember any of such function
> right now, other than load_bmp and create_bitmap which can be thought of
as
> a constructor (thus it has to be like that, and it is good).
>
> And that it should violate the law of encapsulation - hmm... I think they
> were thinking about something slightly different, maybe.

Well maybe it was the example that I gave,  I asked if ...

int *function(void);

... was the correct declaration of a function that returns a pointer and
they said yes, but that it shouldn't be done, and they mumbled something
about functions "being responsible for creating and destroying all
dynamically allocated data" or something similar.  Maybe I just interpreted
it wrong.