> On Mon, 13 Aug 2001, Jeroen van Wolffelaar wrote:
>
> >> Now you assume that you need to pass arguments to these functions.
> >> Which is not the case. Here's the proto:
> >>
> >> /* {{{ proto int rand([int min, int max])
> >> Returns a random number */
> >>
> >> So revert that patch. mt_rand() / rand() accecpt either 2 parameters
or
> >> none at all. It was at least LESS wrong before. Now it's totally wrong.
> >
> >_if_ I remember right (not sure), there are now 2 proto's, one with no
args,
> >and one with2 args.
>
> Both protos for both mt_rand() and rand() are the same and correct.
I never denied that the proto-in-the-source is correct. I've read the source
a time ago already, and I know how it works.
What I mean is, that in the manual, it says now:
rand function synopsis:
- rand()
- rand(int min, int max)
Indicating two possibilities. That's what I was referring to with TWO
proto's.
This is semantically identical with
- rand([int min, int max])
I prefer the latter.
I use the word proto in a general meaning: A prototype for that function, I
am not specifically referring to the {{{ proto ... in the sources.
In this case, I was referring to the protos (2 proto's for rand()) that are
currently in the manual.
> >That would mean it is correct now, though I agree with you that that
proto
> >is better.
>
> It's not the proto that matters here. It's the fact that the functions
> happen to WORK like this. Just check the sources yourself.
> Both of these functions take either no arguments at all OR 2 arguments.
I know, sorry that I was possibly unclear.
> >I prefer your patch though, although it is IMO a hack. But the result was
> >correct.
>
> I never claimed it to be the right way to do this but it seems to be the
> only way to get the correct result. Unless someone finds a better
> solution, of course. ( I dunno docbook that well )
I totally agree: it isn't "right" in the docbook-sense, but
<quote>it seems to be the only way to get the correct result</quote>
> --Jani
Jeroen
