1 _____________________________________________________________________________
3 // use 'a'-'z' or 'a-z' instead of [a-z]?
5 // #include (with renaming?)
9 - better toString() methods all around...
11 - cleanup: Forest, ParseFailed, Parser, Union, (just a bit: GSS,Node)
17 foo.add(y.andnot(x)) ==> this is broken
20 ..................................................
22 - serializable parse tables?
23 - Treewalker code compiler?
25 ______________________________________________________________________________
28 - precedes restrictions ("<-")
29 - MUST HAVE BETTER ERROR MESSAGES
30 - use for developing java15.g
32 - once this is ready, do big announcement
33 - broader regression testing (for stuff like error messages, etc)
34 - More topology untangling [later]
35 - grammar highlighting?
36 - Forest needs a "manual access" API
37 - the unwrap bit in Forest makes it really hard to expose an API for forests
41 ______________________________________________________________________________
44 - finalize metagrammar and rdp-op's
49 - RFC2822 (email message/headers)
50 - clean up the whole Walk situation (?)
53 ______________________________________________________________________________
56 - serialization of parse tables
58 - "ambiguity modulo dropped fragments"?
59 - can this be checked statically?
60 - eliminated statically?
62 - substring parsing for better error messages
65 - "Regular Right Part" grammars (NP Chapman, etc)
66 - Attribute unification
68 - inference of rejections for literals
69 - "prefer whitespace higher up" (?)
71 - Labeled edges on trees (associate a label with each slot in the
72 child array in Forest.Body? might make equality tough) --
73 equivalent to Feature Structures. Colon-labeling.
75 ______________________________________________________________________________
78 - understand and implement the RNGLR "kernel state" optimization.
79 The _Practical Early Parsing_ paper may help.
81 - Partly-Linear-PATR? (O(n^6) unification grammar)
83 - Implement a k-token peek buffer (for each state, see if it "dead
84 ends" during the next k Phases based solely on state -- ignoring
87 - Arrange for the SPPF corresponding to dropped subtrees to never be
88 generated (or merged, etc)
90 - Is there any way we can avoid creating a GSS.Node instance for
91 nodes which are transient in the sense that they have only one
94 - Re-read Rekers, particularly the stuff on optimal sharing
96 - Isolate the Element objects from Parse.Table/GSS so we can move
99 - consider allowing a Forest.Body to represent some other Tree whose
100 Body's should be [recursively] considered part of this Forest.
102 - perhaps not: right now we have a nice situation where
103 Forest.Ref instances become immutable once iterator()ed. This
104 also gives us a strong place to to culling with the certainty
105 that we won't throw out a Body which would later be salvaged
106 by some yet-to-be-added dependency.
108 - Figure out if there is a way to:
110 - allow unwrapping of children other than the very last one.
112 - fold repetitions into an array form in Forest, before
113 conversion to Tree. The major problem here is that multiple
114 tree-arrays are possible, all of different lengths. Worse,
115 even if they're all the same length, not all elements belong
116 in the same "possibility vector" as all others. You
117 essentially need a GSS to represent the array, which perhaps
118 is what the unfolded form was in the first place.
120 - Wikipedia grammar (needs to be both lexerless and boolean)
123 => Ordered Choice (";" operator)
125 - bring back in parse-table phase resolution of precedence (just
126 like associativity). This can be inferred from the use of ">"
127 when the rules are in one of these special forms:
138 where "_" is anything and "E" is the defining nonterminal.
139 Essentially what we're looking for is the situation where the
140 leftmost portion of one rule produces another rule, and the
141 rightmost portion of the latter produces the former.
143 I'm not 100% certain that this is as "strong" as the prefer/avoid
144 form (try to prove this, you probably can), but it's "what people
145 intend" most of the time.
147 - implement Johnstone's algorithm for "reduced, resolved LR
148 tables" to eliminate superfluous reductions on
151 ______________________________________________________________________________
154 - Rekers & Koorn note that GLR Substring Parsing can be used to do
155 really elegant and generalized "autocompletion".
158 ______________________________________________________________________________
161 - Incremental parse table construction
162 - "lazy GLR" and "lazy trees" -> language with first-class CF matching
163 - perhaps linear boolean grammars instead? (linear time, quad space)
164 - Forest parsing => chained parsers
165 - unification parsing, attributes, etc
167 - Take another stab at maximal-match? Nonterminal not-followed-by is
169 - Error recovery based on substring parsing