Invisible wrote:

> This is basically what I did - and it worked exactly right 95% of the 
> time. But if you bracket your expressions in odd (but valid) ways, it 
> complains.

Ah-HAH!

I know what I did wrong. I was looking for brackets at *every* level of 
precidence. The correct behaviour is to look for them *only* at the 
lowest precidence level. Otherwise one of the higher levels gobbles them 
up, and then complains that the seperator is a lower-precidence 
operator. Oops.

OK, now I can bracket my expressions in all sorts of weird ways and the 
parser always seems to give me the correct answer. (But I've said that 
before!) Hmm, I might have to see about writing a formal test harness to 
check whether it "really is" working properly this time...

Post a reply to this message