Re: [Kawa-commonlisp-dev] [PATCH] DECLARE processing.

[Per Bothner]

On 08/14/2012 01:26 PM, Charles Turner wrote:
> On 14 August 2012 21:22, Jamison Hope <address@hidden> wrote:
> > I wasn't sure what he meant, either, but I think you got it right.
> > I don't see the FIXME you mentioned, though (at least not in so many
> > words).  Is it this comment?
> >
> > +        // if declIterator.next() is not an instance of Pair, fail.
>
> Sorry Jamison, don't know what's wrong with me tonight. That's what I
> was referring to.

If there is a syntax error, just make sure SourceMessages#error
is called - preferably by calling one of the Compilation#error
methods is call.  Then  try to recover is a sane way - at least
so we don't get an exception, and preferable to minimize the
chance of follow-on useless error messages.  For example:

  Object declNext =  declIterator.next();
  if (! (declNext instanceof Pair)) {
    Object save = tr.pushPositionOf(???);
    tr.error("malformed declare");
    tr.popPositionOf(save(;
    break;
  }

The pushPositionOf/popPositionOf pair is optional, to provide a
more precise error message.

You'll notice that I don't actually use SeqPosition/getIterator
much, and certainly not in scan/rewrite of syntax forms.
Not sure it buys you anything in this situation.
--
        --Per Bothner
address@hidden   http://per.bothner.com/