1 _____________________________________________________________________________
4 - finalize metagrammar and rdp-op's
6 - Lay down the law on the different kinds of Sequence productions
12 - whitespace-in-braces?
13 - Deal with the problem of zero-rep productions and whitespace insertion
15 - switch maximal to not-followed-by (~/~)
17 - should Union.add() be there?
18 - should Atom.top() be there?
20 - fix the location stuff, it's broken
21 - decent/better error messages
27 - RFC2822 (email message/headers)
30 ______________________________________________________________________________
33 - clean up the whole Walk situation
35 - cleaner solution to "maximal"?
38 - right now I can only lift the last child in a forest... begs
39 the question of what the right representation for Forests is
40 if we need to be able to do lift operations on it.
43 - "Regular Right Part" grammars (NP Chapman, etc)
44 - Attribute unification
46 - serialization of parse tables
47 - inference of rejections for literals
48 - "prefer whitespace higher up" (?)
49 - "ambiguity modulo dropped fragments"?
50 - can this be checked statically?
51 - eliminated statically?
53 ______________________________________________________________________________
56 - Implement a k-token peek buffer (for each state, see if it "dead
57 ends" during the next k Phases based solely on state -- ignoring
60 - Arrange for the SPPF corresponding to dropped subtrees to never be
61 generated (or merged, etc)
63 - Is there any way we can avoid creating a GSS.Node instance for
64 nodes which are transient in the sense that they have only one
67 - Implement "GLR syntactic predicates" -- the ability to do
68 arbitrary lookahead (ie "followed-by" and "not-followed-by" for
69 arbitrary patterns). This enables generalized longest-match and
70 lets us drop the Maximal hack.
72 - Re-read Rekers, particularly the stuff on optimal sharing
74 - Isolate the Element objects from Parse.Table/GSS so we can move
77 - consider allowing a Forest.Body to represent some other Tree whose
78 Body's should be [recursively] considered part of this Forest.
80 - perhaps not: right now we have a nice situation where
81 Forest.Ref instances become immutable once iterator()ed. This
82 also gives us a strong place to to culling with the certainty
83 that we won't throw out a Body which would later be salvaged
84 by some yet-to-be-added dependency.
86 - Figure out if there is a way to:
88 - allow unwrapping of children other than the very last one.
90 - fold repetitions into an array form in Forest, before
91 conversion to Tree. The major problem here is that multiple
92 tree-arrays are possible, all of different lengths. Worse,
93 even if they're all the same length, not all elements belong
94 in the same "possibility vector" as all others. You
95 essentially need a GSS to represent the array, which perhaps
96 is what the unfolded form was in the first place.
98 - Wikipedia grammar (needs to be both lexerless and boolean)
101 => Ordered Choice (";" operator)
103 - bring back in parse-table phase resolution of precedence (just
104 like associativity). This can be inferred from the use of ">"
105 when the rules are in one of these special forms:
116 where "_" is anything and "E" is the defining nonterminal.
117 Essentially what we're looking for is the situation where the
118 leftmost portion of one rule produces another rule, and the
119 rightmost portion of the latter produces the former.
121 I'm not 100% certain that this is as "strong" as the prefer/avoid
122 form (try to prove this, you probably can), but it's "what people
123 intend" most of the time.
125 - implement Johnstone's algorithm for "reduced, resolved LR
126 tables" to eliminate superfluous reductions on
129 ______________________________________________________________________________
132 - Rekers & Koorn note that GLR Substring Parsing can be used to do
133 really elegant and generalized "autocompletion".