1 ______________________________________________________________________________
4 - switch maximal to not-followed-by (~/~)
6 - should Union.add() be there?
7 - should Atom.top() be there?
9 - fix the location stuff, it's broken
10 - decent/better error messages
16 - RFC2822 (email message/headers)
19 ______________________________________________________________________________
22 - clean up the whole Walk situation
24 - cleaner solution to "maximal"?
27 - right now I can only lift the last child in a forest... begs
28 the question of what the right representation for Forests is
29 if we need to be able to do lift operations on it.
32 - "Regular Right Part" grammars (NP Chapman, etc)
33 - Attribute unification
35 - serialization of parse tables
36 - inference of rejections for literals
37 - "prefer whitespace higher up" (?)
38 - "ambiguity modulo dropped fragments"?
39 - can this be checked statically?
40 - eliminated statically?
42 ______________________________________________________________________________
45 - Implement a k-token peek buffer (for each state, see if it "dead
46 ends" during the next k Phases based solely on state -- ignoring
49 - Arrange for the SPPF corresponding to dropped subtrees to never be
50 generated (or merged, etc)
52 - Is there any way we can avoid creating a GSS.Node instance for
53 nodes which are transient in the sense that they have only one
56 - Implement "GLR syntactic predicates" -- the ability to do
57 arbitrary lookahead (ie "followed-by" and "not-followed-by" for
58 arbitrary patterns). This enables generalized longest-match and
59 lets us drop the Maximal hack.
61 - Re-read Rekers, particularly the stuff on optimal sharing
63 - Isolate the Element objects from Parse.Table/GSS so we can move
66 - consider allowing a Forest.Body to represent some other Tree whose
67 Body's should be [recursively] considered part of this Forest.
69 - perhaps not: right now we have a nice situation where
70 Forest.Ref instances become immutable once iterator()ed. This
71 also gives us a strong place to to culling with the certainty
72 that we won't throw out a Body which would later be salvaged
73 by some yet-to-be-added dependency.
75 - Figure out if there is a way to:
77 - allow unwrapping of children other than the very last one.
79 - fold repetitions into an array form in Forest, before
80 conversion to Tree. The major problem here is that multiple
81 tree-arrays are possible, all of different lengths. Worse,
82 even if they're all the same length, not all elements belong
83 in the same "possibility vector" as all others. You
84 essentially need a GSS to represent the array, which perhaps
85 is what the unfolded form was in the first place.
87 - Wikipedia grammar (needs to be both lexerless and boolean)
90 => Ordered Choice (";" operator)
92 - bring back in parse-table phase resolution of precedence (just
93 like associativity). This can be inferred from the use of ">"
94 when the rules are in one of these special forms:
105 where "_" is anything and "E" is the defining nonterminal.
106 Essentially what we're looking for is the situation where the
107 leftmost portion of one rule produces another rule, and the
108 rightmost portion of the latter produces the former.
110 I'm not 100% certain that this is as "strong" as the prefer/avoid
111 form (try to prove this, you probably can), but it's "what people
112 intend" most of the time.
114 - implement Johnstone's algorithm for "reduced, resolved LR
115 tables" to eliminate superfluous reductions on