Re: [AD] Documentation update

[Grzegorz Adam Hankiewicz <gradha@xxxxxxxxxx>]

On 2004-09-09, Chris <kcat@xxxxxxxxxx> wrote:
> I think it's a bad idea to encourage skipping return values, even on 
> functions that aren't likely to fail. The programmer can do so at his 
> own peril, but I don't think we should say "Don't worry about this 
> return value." It's there for a reason, after all.

The reason is a poorly designed API. See install_timer. You also replied
to the pending issue:

  http://sourceforge.net/mailarchive/message.php?msg_id=8457049

So even if we had a platform where a keyboard was physically not
attached, most likely we would provide a dummy keyboard driver which
could be implemented in the platform as something completely different
(think of PDA stylus input), transforming the function returning int
into a practical void one, since the function would always report
success. In fact, maybe we should make this kind of statement for these
type of functions.

But that note acknowledges the reason why no Allegro provided example
checks for the return value. I think it is fair, given that we usually
tell newbies to check the return value of create_bitmap and other.

On 2004-09-10, Peter Wang <tjaden@xxxxxxxxxx> wrote:
> I saw a couple of djgppisms:

Removed and fixed. New version attached.

Attachment: doc_update2.diff.gz
Description: Binary data