FIRST_k, FOLLOW_k, k>1

From	Andy <borucki.andrzej@gmail.com>
Newsgroups	comp.compilers
Subject	FIRST_k, FOLLOW_k, k>1
Date	2020-02-06 10:43 -0800
Organization	Compilers Central
Message-ID	<20-02-004@comp.compilers> (permalink)

Show all headers | View raw

If LL(l) and LR(k) need sets FIRST_k, FOLLOW_k, k>1, for example LR(3) need this sets of degree 3?
How make it? How is the best structure for these sets? I think about pyramid:
- one bit - is epsilon
- bit table size n
- bit table size n^2
...
- bit table size n^k

because sets degree of k also have shorter strings.

This exponentially grows, in real grammars, number of tokens can be
quite big, ~80, It need then 0(80^k) bits. This sets will sparse and
is better organization of substrings in sets? For example, not bit set
but set of trees or DFA's ?

Back to comp.compilers | Previous | Next — Next in thread | Find similar

Thread

FIRST_k, FOLLOW_k, k>1 Andy <borucki.andrzej@gmail.com> - 2020-02-06 10:43 -0800
  Re: FIRST_k, FOLLOW_k, k>1 Andy <borucki.andrzej@gmail.com> - 2020-02-06 14:16 -0800
    Re: FIRST_k, FOLLOW_k, k>1 Hans-Peter Diettrich <DrDiettrich1@netscape.net> - 2020-02-08 11:00 +0100
      Re: FIRST_k, FOLLOW_k, k>1 Andy <borucki.andrzej@gmail.com> - 2020-02-08 11:54 -0800
        Re: FIRST_k, FOLLOW_k, k>1 Hans-Peter Diettrich <DrDiettrich1@netscape.net> - 2020-02-09 02:46 +0100
  FIRST_k, FOLLOW_k, k&gt;1 Christopher F Clark <christopher.f.clark@compiler-resources.com> - 2020-02-07 16:12 +0200
  Re: FIRST_k, FOLLOW_k, k>1 honey crisis <gaztoast@gmail.com> - 2020-02-08 18:18 -0800

csiph-web