> -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf Of Raul Miller > Sent: Monday, May 21, 2007 8:28 PM > To: General forum > Subject: Re: [Jgeneral] is this syntax error a fixable bug? > > On 5/21/07, Pascal Jasmin <[EMAIL PROTECTED]> wrote: > > in J 6.02, v=verb, a=adverb, c=conjunction > > Perhaps you mean j 6.01? > > > v a c v is allowed and equivalent to: > > v a (c v) NB. best I can tell > > That depends on c and on a.
Isn't the statement always true? v a c v equals (v a) c v equals (v a) (c v) > > For something like +/ .* you are correct. > > > v a v c is allowed, but instead of > > > > v a (v c), it is equivalent to (v a v) c which doesn't obey parsing > > rules or my understanding. > > Again, this depends on c and on a. > > > +/+:@: > > (+/ +:)@: > > Yes. Wow. I didn't know that, but it does follow from the rules (not depending on c and a as far as I can tell) v a v c is (v a) v c which matches line 6 (bident) edge cavn cavn any so the two verbs become a hook. v v v c similarly. > > > Is there a reason why the full sentence: > > a c v or a v c should give a syntax error instead of mean > the adverb: > > a (c v) or a (v c) > > I believe this used to have a different meaning. I'm not sure what > current plans are. Roger got rid of those in 5.01 IIRC. The reasoning was that hardly anybody understood them, they led to code repeated between normal and debug cases, and they cost one extra line in the parsing table. Henry Rich > > -- > Raul > ---------------------------------------------------------------------- > For information about J forums see > http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
