[antlr-interest] ANTLR 3 dev status?
sk at z.pl
Sat Mar 19 04:44:09 PST 2005
On Friday 18 March 2005 20:51, Terence Parr wrote:
> Yes, i am very aware of the strange change in operator meaning and have
> struggled to figure out the right answer. I will probably go with the
> backward incompatibility because ^^ does really make more sense like ..
> to mean up a directory or whatever. Hmm...
Yup. That's why i tried to come with something among the lines. As single
character (here ^) can't be divided -- I tried to come some character which
looks like half of that ^, hence little ` -- it looks quite like the part
of ^ in many fonts, and ocupies similar place at the top of character row.
This is just my (silent) voice in the discussion.
I simply came with this now, as I rememeber the dicussion from some time
ago: ^ and ^^ breaks compatibility and might make some troubles in
migrating grammars from 2.x to 3.x, especially if someone left unchanged
single ^ somewhere in his .g file, esp in some cases this would express
itself only in (potentially rarely used) some single path of parsing. But
no one was able to come with anything even halfway reasonable, which would
not look misleading (esp for newcomers) and not terribly verbose (thus
hiding the token name inside noise of (,), %, $, @, whatever). Hence the
conslusion was then that nothing better than changing ^ to ^^ and
introducing new meaning to ^ had been found. I came with my little idea
only yesterday, so I posted it here. It well may get rejected, but I'll
sleep peacefully now, as I now I didn't keep the possibly "greatest idea to
save the world" ;) for myself only but put it for a judgement of others.
More information about the antlr-interest