I agree that this seems like the right message -- testing with an
incomplete for, or select. shows that we get a different (more
detailed) message in that case.
That said, my memory is telling me that 3 :'' has always matched [: in
behavior. So I would be interested in learning about what had changed
there.
Thanks,
--
Raul
On Sun, Jul 9, 2023 at 2:04 PM Henry Rich <henryhr...@gmail.com> wrote:
>
> OK, I've changed the message to 'unmatched {{ or }}'. [This is in
> dev/eformat/eformat.ijs and I encourage you to make changes as you see fit.]
>
> It is true that 3 : '' is accepted; a questionable decision since the
> result is equivalent to [: (i. e. unexecutable). This has changed over
> the years.
>
> Henry Rich
>
> On 7/9/2023 1:23 PM, Raul Miller wrote:
> > Hmm... I can see that "unfinished {{ }} definition" would include
> > things like a missing end. statement.
> >
> > However, J has always supported empty definitions (in the sense of
> > allowing them to represent a definition with an empty domain). So that
> > part of the message is unclear to me.
> >
> > Thanks,
> >
>
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
