ruco-cpp 0.2.1
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- checksums.yaml +7 -0
- data/Rakefile +17 -0
- data/bin/console +14 -0
- data/bin/ruco +30 -0
- data/bin/setup +7 -0
- data/data/ruco/Parser.frame +359 -0
- data/data/ruco/Scanner.frame +896 -0
- data/data/ruco/picojson/Changes +14 -0
- data/data/ruco/picojson/LICENSE +25 -0
- data/data/ruco/picojson/Makefile +8 -0
- data/data/ruco/picojson/README.mkdn +183 -0
- data/data/ruco/picojson/examples/github-issues.cc +110 -0
- data/data/ruco/picojson/examples/iostream.cc +70 -0
- data/data/ruco/picojson/examples/streaming.cc +76 -0
- data/data/ruco/picojson/picojson.h +1299 -0
- data/ext/cocor/Action.cpp +81 -0
- data/ext/cocor/Action.h +59 -0
- data/ext/cocor/ArrayList.cpp +79 -0
- data/ext/cocor/ArrayList.h +52 -0
- data/ext/cocor/BitArray.cpp +156 -0
- data/ext/cocor/BitArray.h +68 -0
- data/ext/cocor/CharClass.cpp +42 -0
- data/ext/cocor/CharClass.h +48 -0
- data/ext/cocor/CharSet.cpp +166 -0
- data/ext/cocor/CharSet.h +68 -0
- data/ext/cocor/Coco.atg +528 -0
- data/ext/cocor/Coco.cpp +173 -0
- data/ext/cocor/Comment.cpp +45 -0
- data/ext/cocor/Comment.h +51 -0
- data/ext/cocor/Copyright.frame +27 -0
- data/ext/cocor/DFA.cpp +865 -0
- data/ext/cocor/DFA.h +132 -0
- data/ext/cocor/Generator.cpp +182 -0
- data/ext/cocor/Generator.h +61 -0
- data/ext/cocor/Graph.h +59 -0
- data/ext/cocor/HashTable.cpp +115 -0
- data/ext/cocor/HashTable.h +84 -0
- data/ext/cocor/Makefile +11 -0
- data/ext/cocor/Melted.cpp +39 -0
- data/ext/cocor/Melted.h +51 -0
- data/ext/cocor/Node.cpp +69 -0
- data/ext/cocor/Node.h +86 -0
- data/ext/cocor/Parser.cpp +925 -0
- data/ext/cocor/Parser.frame +326 -0
- data/ext/cocor/Parser.h +153 -0
- data/ext/cocor/ParserGen.cpp +486 -0
- data/ext/cocor/ParserGen.h +99 -0
- data/ext/cocor/Position.cpp +37 -0
- data/ext/cocor/Position.h +46 -0
- data/ext/cocor/README.md +12 -0
- data/ext/cocor/Scanner.cpp +833 -0
- data/ext/cocor/Scanner.frame +897 -0
- data/ext/cocor/Scanner.h +291 -0
- data/ext/cocor/Sets.h +84 -0
- data/ext/cocor/SortedList.cpp +141 -0
- data/ext/cocor/SortedList.h +68 -0
- data/ext/cocor/State.cpp +77 -0
- data/ext/cocor/State.h +55 -0
- data/ext/cocor/StringBuilder.cpp +88 -0
- data/ext/cocor/StringBuilder.h +29 -0
- data/ext/cocor/Symbol.cpp +61 -0
- data/ext/cocor/Symbol.h +70 -0
- data/ext/cocor/Tab.cpp +1248 -0
- data/ext/cocor/Tab.h +245 -0
- data/ext/cocor/Target.cpp +41 -0
- data/ext/cocor/Target.h +48 -0
- data/ext/cocor/build.bat +3 -0
- data/ext/cocor/build.sh +4 -0
- data/ext/cocor/coc.bat +1 -0
- data/ext/cocor/coc.sh +2 -0
- data/ext/cocor/cocor_ruby_ext.cpp +124 -0
- data/ext/cocor/cygBuild.bat +1 -0
- data/ext/cocor/extconf.rb +5 -0
- data/ext/cocor/mingwbuild.bat +2 -0
- data/ext/cocor/mkmf.log +57 -0
- data/ext/cocor/zipsources.bat +1 -0
- data/lib/cocor.rb +14 -0
- data/lib/ruco/version.rb +3 -0
- data/lib/ruco.rb +728 -0
- metadata +195 -0
data/ext/cocor/Tab.cpp
ADDED
@@ -0,0 +1,1248 @@
|
|
1
|
+
/*-------------------------------------------------------------------------
|
2
|
+
Tab -- Symbol Table Management
|
3
|
+
Compiler Generator Coco/R,
|
4
|
+
Copyright (c) 1990, 2004 Hanspeter Moessenboeck, University of Linz
|
5
|
+
extended by M. Loeberbauer & A. Woess, Univ. of Linz
|
6
|
+
ported to C++ by Csaba Balazs, University of Szeged
|
7
|
+
with improvements by Pat Terry, Rhodes University
|
8
|
+
|
9
|
+
This program is free software; you can redistribute it and/or modify it
|
10
|
+
under the terms of the GNU General Public License as published by the
|
11
|
+
Free Software Foundation; either version 2, or (at your option) any
|
12
|
+
later version.
|
13
|
+
|
14
|
+
This program is distributed in the hope that it will be useful, but
|
15
|
+
WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
|
16
|
+
or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
|
17
|
+
for more details.
|
18
|
+
|
19
|
+
You should have received a copy of the GNU General Public License along
|
20
|
+
with this program; if not, write to the Free Software Foundation, Inc.,
|
21
|
+
59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
|
22
|
+
|
23
|
+
As an exception, it is allowed to write an extension of Coco/R that is
|
24
|
+
used as a plugin in non-free software.
|
25
|
+
|
26
|
+
If not otherwise stated, any source code generated by Coco/R (other than
|
27
|
+
Coco/R itself) does not fall under the GNU General Public License.
|
28
|
+
-------------------------------------------------------------------------*/
|
29
|
+
|
30
|
+
#include <wchar.h>
|
31
|
+
#include "Tab.h"
|
32
|
+
#include "Parser.h"
|
33
|
+
#include "BitArray.h"
|
34
|
+
#include "Scanner.h"
|
35
|
+
|
36
|
+
namespace Coco {
|
37
|
+
|
38
|
+
const char* Tab::nTyp[] =
|
39
|
+
{" ", "t ", "pr ", "nt ", "clas", "chr ", "wt ", "any ", "eps ",
|
40
|
+
"sync", "sem ", "alt ", "iter", "opt ", "rslv"};
|
41
|
+
|
42
|
+
const char* Tab::tKind[] = {"fixedToken", "classToken", "litToken", "classLitToken"};
|
43
|
+
|
44
|
+
Tab::Tab(Parser *parser) {
|
45
|
+
for (int i=0; i<10; i++) ddt[i] = false;
|
46
|
+
|
47
|
+
terminals = new ArrayList();
|
48
|
+
pragmas = new ArrayList();
|
49
|
+
nonterminals = new ArrayList();
|
50
|
+
nodes = new ArrayList();
|
51
|
+
dummyNode = NULL;
|
52
|
+
classes= new ArrayList();
|
53
|
+
dummyName = 'A';
|
54
|
+
|
55
|
+
this->parser = parser;
|
56
|
+
trace = parser->trace;
|
57
|
+
errors = parser->errors;
|
58
|
+
eofSy = NewSym(Node::t, L"EOF", 0);
|
59
|
+
dummyNode = NewNode(Node::eps, (Symbol*)NULL, 0);
|
60
|
+
literals = new HashTable();
|
61
|
+
checkEOF = true;
|
62
|
+
}
|
63
|
+
|
64
|
+
|
65
|
+
Symbol* Tab::NewSym(int typ, const wchar_t* name, int line) {
|
66
|
+
if (coco_string_length(name) == 2 && name[0] == '"') {
|
67
|
+
parser->SemErr(L"empty token not allowed");
|
68
|
+
name = coco_string_create(L"???");
|
69
|
+
}
|
70
|
+
Symbol *sym = new Symbol(typ, name, line);
|
71
|
+
|
72
|
+
if (typ == Node::t) {
|
73
|
+
sym->n = terminals->Count; terminals->Add(sym);
|
74
|
+
} else if (typ == Node::pr) {
|
75
|
+
pragmas->Add(sym);
|
76
|
+
} else if (typ == Node::nt) {
|
77
|
+
sym->n = nonterminals->Count; nonterminals->Add(sym);
|
78
|
+
}
|
79
|
+
|
80
|
+
return sym;
|
81
|
+
}
|
82
|
+
|
83
|
+
Symbol* Tab::FindSym(const wchar_t* name) {
|
84
|
+
Symbol *s;
|
85
|
+
int i;
|
86
|
+
for (i=0; i<terminals->Count; i++) {
|
87
|
+
s = (Symbol*)((*terminals)[i]);
|
88
|
+
if (coco_string_equal(s->name, name)) return s;
|
89
|
+
}
|
90
|
+
for (i=0; i<nonterminals->Count; i++) {
|
91
|
+
s = (Symbol*)((*nonterminals)[i]);
|
92
|
+
if (coco_string_equal(s->name, name)) return s;
|
93
|
+
}
|
94
|
+
return NULL;
|
95
|
+
}
|
96
|
+
|
97
|
+
int Tab::Num(Node *p) {
|
98
|
+
if (p == NULL) return 0; else return p->n;
|
99
|
+
}
|
100
|
+
|
101
|
+
void Tab::PrintSym(Symbol *sym) {
|
102
|
+
wchar_t *paddedName = Name(sym->name);
|
103
|
+
fwprintf(trace, L"%3d %14s %ls", sym->n, paddedName, nTyp[sym->typ]);
|
104
|
+
coco_string_delete(paddedName);
|
105
|
+
|
106
|
+
if (sym->attrPos==NULL) fwprintf(trace, L" false "); else fwprintf(trace, L" true ");
|
107
|
+
if (sym->typ == Node::nt) {
|
108
|
+
fwprintf(trace, L"%5d", Num(sym->graph));
|
109
|
+
if (sym->deletable) fwprintf(trace, L" true "); else fwprintf(trace, L" false ");
|
110
|
+
} else
|
111
|
+
fwprintf(trace, L" ");
|
112
|
+
|
113
|
+
fwprintf(trace, L"%5d %ls\n", sym->line, tKind[sym->tokenKind]);
|
114
|
+
}
|
115
|
+
|
116
|
+
void Tab::PrintSymbolTable() {
|
117
|
+
fwprintf(trace, L"Symbol Table:\n");
|
118
|
+
fwprintf(trace, L"------------\n\n");
|
119
|
+
fwprintf(trace, L" nr name typ hasAt graph del line tokenKind\n");
|
120
|
+
|
121
|
+
Symbol *sym;
|
122
|
+
int i;
|
123
|
+
for (i=0; i<terminals->Count; i++) {
|
124
|
+
sym = (Symbol*)((*terminals)[i]);
|
125
|
+
PrintSym(sym);
|
126
|
+
}
|
127
|
+
for (i=0; i<pragmas->Count; i++) {
|
128
|
+
sym = (Symbol*)((*pragmas)[i]);
|
129
|
+
PrintSym(sym);
|
130
|
+
}
|
131
|
+
for (i=0; i<nonterminals->Count; i++) {
|
132
|
+
sym = (Symbol*)((*nonterminals)[i]);
|
133
|
+
PrintSym(sym);
|
134
|
+
}
|
135
|
+
|
136
|
+
|
137
|
+
fwprintf(trace, L"\nLiteral Tokens:\n");
|
138
|
+
fwprintf(trace, L"--------------\n");
|
139
|
+
|
140
|
+
Iterator *iter = literals->GetIterator();
|
141
|
+
while (iter->HasNext()) {
|
142
|
+
DictionaryEntry *e = iter->Next();
|
143
|
+
fwprintf(trace, L"_%ls = %ls.\n", ((Symbol*) (e->val))->name, e->key);
|
144
|
+
}
|
145
|
+
fwprintf(trace, L"\n");
|
146
|
+
}
|
147
|
+
|
148
|
+
void Tab::PrintSet(BitArray *s, int indent) {
|
149
|
+
int col, len;
|
150
|
+
col = indent;
|
151
|
+
Symbol *sym;
|
152
|
+
for (int i=0; i<terminals->Count; i++) {
|
153
|
+
sym = (Symbol*)((*terminals)[i]);
|
154
|
+
if ((*s)[sym->n]) {
|
155
|
+
len = coco_string_length(sym->name);
|
156
|
+
if (col + len >= 80) {
|
157
|
+
fwprintf(trace, L"\n");
|
158
|
+
for (col = 1; col < indent; col++) fwprintf(trace, L" ");
|
159
|
+
}
|
160
|
+
fwprintf(trace, L"%ls ", sym->name);
|
161
|
+
col += len + 1;
|
162
|
+
}
|
163
|
+
}
|
164
|
+
if (col == indent) fwprintf(trace, L"-- empty set --");
|
165
|
+
fwprintf(trace, L"\n");
|
166
|
+
}
|
167
|
+
|
168
|
+
//---------------------------------------------------------------------
|
169
|
+
// Syntax graph management
|
170
|
+
//---------------------------------------------------------------------
|
171
|
+
|
172
|
+
Node* Tab::NewNode(int typ, Symbol *sym, int line) {
|
173
|
+
Node* node = new Node(typ, sym, line);
|
174
|
+
node->n = nodes->Count;
|
175
|
+
nodes->Add(node);
|
176
|
+
return node;
|
177
|
+
}
|
178
|
+
|
179
|
+
|
180
|
+
Node* Tab::NewNode(int typ, Node* sub) {
|
181
|
+
Node* node = NewNode(typ, (Symbol*)NULL, 0);
|
182
|
+
node->sub = sub;
|
183
|
+
return node;
|
184
|
+
}
|
185
|
+
|
186
|
+
Node* Tab::NewNode(int typ, int val, int line) {
|
187
|
+
Node* node = NewNode(typ, (Symbol*)NULL, line);
|
188
|
+
node->val = val;
|
189
|
+
return node;
|
190
|
+
}
|
191
|
+
|
192
|
+
|
193
|
+
void Tab::MakeFirstAlt(Graph *g) {
|
194
|
+
g->l = NewNode(Node::alt, g->l); g->l->line = g->l->sub->line;
|
195
|
+
g->r->up = true;
|
196
|
+
g->l->next = g->r;
|
197
|
+
g->r = g->l;
|
198
|
+
}
|
199
|
+
|
200
|
+
// The result will be in g1
|
201
|
+
void Tab::MakeAlternative(Graph *g1, Graph *g2) {
|
202
|
+
g2->l = NewNode(Node::alt, g2->l); g2->l->line = g2->l->sub->line;
|
203
|
+
g2->l->up = true;
|
204
|
+
g2->r->up = true;
|
205
|
+
Node *p = g1->l; while (p->down != NULL) p = p->down;
|
206
|
+
p->down = g2->l;
|
207
|
+
p = g1->r; while (p->next != NULL) p = p->next;
|
208
|
+
// append alternative to g1 end list
|
209
|
+
p->next = g2->l;
|
210
|
+
// append g2 end list to g1 end list
|
211
|
+
g2->l->next = g2->r;
|
212
|
+
}
|
213
|
+
|
214
|
+
// The result will be in g1
|
215
|
+
void Tab::MakeSequence(Graph *g1, Graph *g2) {
|
216
|
+
Node *p = g1->r->next; g1->r->next = g2->l; // link head node
|
217
|
+
while (p != NULL) { // link substructure
|
218
|
+
Node *q = p->next; p->next = g2->l;
|
219
|
+
p = q;
|
220
|
+
}
|
221
|
+
g1->r = g2->r;
|
222
|
+
}
|
223
|
+
|
224
|
+
void Tab::MakeIteration(Graph *g) {
|
225
|
+
g->l = NewNode(Node::iter, g->l);
|
226
|
+
g->r->up = true;
|
227
|
+
Node *p = g->r;
|
228
|
+
g->r = g->l;
|
229
|
+
while (p != NULL) {
|
230
|
+
Node *q = p->next; p->next = g->l;
|
231
|
+
p = q;
|
232
|
+
}
|
233
|
+
}
|
234
|
+
|
235
|
+
void Tab::MakeOption(Graph *g) {
|
236
|
+
g->l = NewNode(Node::opt, g->l);
|
237
|
+
g->r->up = true;
|
238
|
+
g->l->next = g->r;
|
239
|
+
g->r = g->l;
|
240
|
+
}
|
241
|
+
|
242
|
+
void Tab::Finish(Graph *g) {
|
243
|
+
Node *p = g->r;
|
244
|
+
while (p != NULL) {
|
245
|
+
Node *q = p->next; p->next = NULL;
|
246
|
+
p = q;
|
247
|
+
}
|
248
|
+
}
|
249
|
+
|
250
|
+
void Tab::DeleteNodes() {
|
251
|
+
nodes = new ArrayList();
|
252
|
+
dummyNode = NewNode(Node::eps, (Symbol*)NULL, 0);
|
253
|
+
}
|
254
|
+
|
255
|
+
Graph* Tab::StrToGraph(const wchar_t* str) {
|
256
|
+
wchar_t *subStr = coco_string_create(str, 1, coco_string_length(str)-2);
|
257
|
+
wchar_t *s = Unescape(subStr);
|
258
|
+
coco_string_delete(subStr);
|
259
|
+
if (coco_string_length(s) == 0) parser->SemErr(L"empty token not allowed");
|
260
|
+
Graph *g = new Graph();
|
261
|
+
g->r = dummyNode;
|
262
|
+
for (int i = 0; i < coco_string_length(s); i++) {
|
263
|
+
Node *p = NewNode(Node::chr, (int)s[i], 0);
|
264
|
+
g->r->next = p; g->r = p;
|
265
|
+
}
|
266
|
+
g->l = dummyNode->next; dummyNode->next = NULL;
|
267
|
+
coco_string_delete(s);
|
268
|
+
return g;
|
269
|
+
}
|
270
|
+
|
271
|
+
|
272
|
+
void Tab::SetContextTrans(Node *p) { // set transition code in the graph rooted at p
|
273
|
+
while (p != NULL) {
|
274
|
+
if (p->typ == Node::chr || p->typ == Node::clas) {
|
275
|
+
p->code = Node::contextTrans;
|
276
|
+
} else if (p->typ == Node::opt || p->typ == Node::iter) {
|
277
|
+
SetContextTrans(p->sub);
|
278
|
+
} else if (p->typ == Node::alt) {
|
279
|
+
SetContextTrans(p->sub); SetContextTrans(p->down);
|
280
|
+
}
|
281
|
+
if (p->up) break;
|
282
|
+
p = p->next;
|
283
|
+
}
|
284
|
+
}
|
285
|
+
|
286
|
+
//------------ graph deletability check -----------------
|
287
|
+
|
288
|
+
bool Tab::DelGraph(Node* p) {
|
289
|
+
return p == NULL || (DelNode(p) && DelGraph(p->next));
|
290
|
+
}
|
291
|
+
|
292
|
+
bool Tab::DelSubGraph(Node* p) {
|
293
|
+
return p == NULL || (DelNode(p) && (p->up || DelSubGraph(p->next)));
|
294
|
+
}
|
295
|
+
|
296
|
+
bool Tab::DelNode(Node* p) {
|
297
|
+
if (p->typ == Node::nt) {
|
298
|
+
return p->sym->deletable;
|
299
|
+
}
|
300
|
+
else if (p->typ == Node::alt) {
|
301
|
+
return DelSubGraph(p->sub) || (p->down != NULL && DelSubGraph(p->down));
|
302
|
+
}
|
303
|
+
else {
|
304
|
+
return p->typ == Node::iter || p->typ == Node::opt || p->typ == Node::sem
|
305
|
+
|| p->typ == Node::eps || p->typ == Node::rslv || p->typ == Node::sync;
|
306
|
+
}
|
307
|
+
}
|
308
|
+
|
309
|
+
//----------------- graph printing ----------------------
|
310
|
+
|
311
|
+
int Tab::Ptr(Node *p, bool up) {
|
312
|
+
if (p == NULL) return 0;
|
313
|
+
else if (up) return -(p->n);
|
314
|
+
else return p->n;
|
315
|
+
}
|
316
|
+
|
317
|
+
wchar_t* Tab::Pos(Position *pos) {
|
318
|
+
wchar_t* format = new wchar_t[10];
|
319
|
+
if (pos == NULL) {
|
320
|
+
coco_swprintf(format, 10, L" ");
|
321
|
+
} else {
|
322
|
+
coco_swprintf(format, 10, L"%5d", pos->beg);
|
323
|
+
}
|
324
|
+
return format;
|
325
|
+
}
|
326
|
+
|
327
|
+
wchar_t* Tab::Name(const wchar_t *name) {
|
328
|
+
wchar_t *name2 = coco_string_create_append(name, L" ");
|
329
|
+
wchar_t *subName2 = coco_string_create(name2, 0, 12);
|
330
|
+
coco_string_delete(name2);
|
331
|
+
return subName2;
|
332
|
+
// found no simpler way to get the first 12 characters of the name
|
333
|
+
// padded with blanks on the right
|
334
|
+
}
|
335
|
+
|
336
|
+
void Tab::PrintNodes() {
|
337
|
+
fwprintf(trace, L"Graph nodes:\n");
|
338
|
+
fwprintf(trace, L"----------------------------------------------------\n");
|
339
|
+
fwprintf(trace, L" n type name next down sub pos line\n");
|
340
|
+
fwprintf(trace, L" val code\n");
|
341
|
+
fwprintf(trace, L"----------------------------------------------------\n");
|
342
|
+
|
343
|
+
Node *p;
|
344
|
+
for (int i=0; i<nodes->Count; i++) {
|
345
|
+
p = (Node*)((*nodes)[i]);
|
346
|
+
fwprintf(trace, L"%4d %ls ", p->n, (nTyp[p->typ]));
|
347
|
+
if (p->sym != NULL) {
|
348
|
+
wchar_t *paddedName = Name(p->sym->name);
|
349
|
+
fwprintf(trace, L"%12s ", paddedName);
|
350
|
+
coco_string_delete(paddedName);
|
351
|
+
} else if (p->typ == Node::clas) {
|
352
|
+
CharClass *c = (CharClass*)(*classes)[p->val];
|
353
|
+
wchar_t *paddedName = Name(c->name);
|
354
|
+
fwprintf(trace, L"%12s ", paddedName);
|
355
|
+
coco_string_delete(paddedName);
|
356
|
+
} else fwprintf(trace, L" ");
|
357
|
+
fwprintf(trace, L"%5d ", Ptr(p->next, p->up));
|
358
|
+
|
359
|
+
if (p->typ == Node::t || p->typ == Node::nt || p->typ == Node::wt) {
|
360
|
+
fwprintf(trace, L" %5s", Pos(p->pos));
|
361
|
+
} if (p->typ == Node::chr) {
|
362
|
+
fwprintf(trace, L"%5d %5d ", p->val, p->code);
|
363
|
+
} if (p->typ == Node::clas) {
|
364
|
+
fwprintf(trace, L" %5d ", p->code);
|
365
|
+
} if (p->typ == Node::alt || p->typ == Node::iter || p->typ == Node::opt) {
|
366
|
+
fwprintf(trace, L"%5d %5d ", Ptr(p->down, false), Ptr(p->sub, false));
|
367
|
+
} if (p->typ == Node::sem) {
|
368
|
+
fwprintf(trace, L" %5s", Pos(p->pos));
|
369
|
+
} if (p->typ == Node::eps || p->typ == Node::any || p->typ == Node::sync) {
|
370
|
+
fwprintf(trace, L" ");
|
371
|
+
}
|
372
|
+
fwprintf(trace, L"%5d\n", p->line);
|
373
|
+
}
|
374
|
+
fwprintf(trace, L"\n");
|
375
|
+
}
|
376
|
+
|
377
|
+
//---------------------------------------------------------------------
|
378
|
+
// Character class management
|
379
|
+
//---------------------------------------------------------------------
|
380
|
+
|
381
|
+
|
382
|
+
CharClass* Tab::NewCharClass(const wchar_t* name, CharSet *s) {
|
383
|
+
CharClass *c;
|
384
|
+
if (coco_string_equal(name, L"#")) {
|
385
|
+
wchar_t* temp = coco_string_create_append(name, (wchar_t) dummyName++);
|
386
|
+
c = new CharClass(temp, s);
|
387
|
+
coco_string_delete(temp);
|
388
|
+
} else {
|
389
|
+
c = new CharClass(name, s);
|
390
|
+
}
|
391
|
+
c->n = classes->Count;
|
392
|
+
classes->Add(c);
|
393
|
+
return c;
|
394
|
+
}
|
395
|
+
|
396
|
+
CharClass* Tab::FindCharClass(const wchar_t* name) {
|
397
|
+
CharClass *c;
|
398
|
+
for (int i=0; i<classes->Count; i++) {
|
399
|
+
c = (CharClass*)((*classes)[i]);
|
400
|
+
if (coco_string_equal(c->name, name)) return c;
|
401
|
+
}
|
402
|
+
return NULL;
|
403
|
+
}
|
404
|
+
|
405
|
+
CharClass* Tab::FindCharClass(CharSet *s) {
|
406
|
+
CharClass *c;
|
407
|
+
for (int i=0; i<classes->Count; i++) {
|
408
|
+
c = (CharClass*)((*classes)[i]);
|
409
|
+
if (s->Equals(c->set)) return c;
|
410
|
+
}
|
411
|
+
return NULL;
|
412
|
+
}
|
413
|
+
|
414
|
+
CharSet* Tab::CharClassSet(int i) {
|
415
|
+
return ((CharClass*)((*classes)[i]))->set;
|
416
|
+
}
|
417
|
+
|
418
|
+
//----------- character class printing
|
419
|
+
|
420
|
+
wchar_t* Tab::Ch(const wchar_t ch) {
|
421
|
+
wchar_t* format = new wchar_t[10];
|
422
|
+
if (ch < L' ' || ch >= 127 || ch == L'\'' || ch == L'\\') {
|
423
|
+
coco_swprintf(format, 10, L"%d", ch);
|
424
|
+
return format;
|
425
|
+
} else {
|
426
|
+
coco_swprintf(format, 10, L"'%lc'", ch);
|
427
|
+
return format;
|
428
|
+
}
|
429
|
+
}
|
430
|
+
|
431
|
+
void Tab::WriteCharSet(CharSet *s) {
|
432
|
+
for (CharSet::Range *r = s->head; r != NULL; r = r->next) {
|
433
|
+
if (r->from < r->to) {
|
434
|
+
wchar_t *from = Ch(r->from);
|
435
|
+
wchar_t *to = Ch(r->to);
|
436
|
+
fwprintf(trace, L"%ls .. %ls ", from, to);
|
437
|
+
delete [] from;
|
438
|
+
delete [] to;
|
439
|
+
}
|
440
|
+
else {
|
441
|
+
wchar_t *from = Ch(r->from);
|
442
|
+
fwprintf(trace, L"%ls ", from);
|
443
|
+
delete [] from;
|
444
|
+
}
|
445
|
+
}
|
446
|
+
}
|
447
|
+
|
448
|
+
void Tab::WriteCharClasses () {
|
449
|
+
CharClass *c;
|
450
|
+
for (int i=0; i<classes->Count; i++) {
|
451
|
+
c = (CharClass*)((*classes)[i]);
|
452
|
+
|
453
|
+
wchar_t* format2 = coco_string_create_append(c->name, L" ");
|
454
|
+
wchar_t* format = coco_string_create(format2, 0, 10);
|
455
|
+
coco_string_merge(format, L": ");
|
456
|
+
fwprintf(trace, format);
|
457
|
+
|
458
|
+
WriteCharSet(c->set);
|
459
|
+
fwprintf(trace, L"\n");
|
460
|
+
coco_string_delete(format);
|
461
|
+
coco_string_delete(format2);
|
462
|
+
}
|
463
|
+
fwprintf(trace, L"\n");
|
464
|
+
}
|
465
|
+
|
466
|
+
//---------------------------------------------------------------------
|
467
|
+
// Symbol set computations
|
468
|
+
//---------------------------------------------------------------------
|
469
|
+
|
470
|
+
/* Computes the first set for the given Node. */
|
471
|
+
BitArray* Tab::First0(Node *p, BitArray *mark) {
|
472
|
+
BitArray *fs = new BitArray(terminals->Count);
|
473
|
+
while (p != NULL && !((*mark)[p->n])) {
|
474
|
+
mark->Set(p->n, true);
|
475
|
+
if (p->typ == Node::nt) {
|
476
|
+
if (p->sym->firstReady) {
|
477
|
+
fs->Or(p->sym->first);
|
478
|
+
} else {
|
479
|
+
BitArray *fs0 = First0(p->sym->graph, mark);
|
480
|
+
fs->Or(fs0);
|
481
|
+
delete fs0;
|
482
|
+
}
|
483
|
+
}
|
484
|
+
else if (p->typ == Node::t || p->typ == Node::wt) {
|
485
|
+
fs->Set(p->sym->n, true);
|
486
|
+
}
|
487
|
+
else if (p->typ == Node::any) {
|
488
|
+
fs->Or(p->set);
|
489
|
+
}
|
490
|
+
else if (p->typ == Node::alt) {
|
491
|
+
BitArray *fs0 = First0(p->sub, mark);
|
492
|
+
fs->Or(fs0);
|
493
|
+
delete fs0;
|
494
|
+
fs0 = First0(p->down, mark);
|
495
|
+
fs->Or(fs0);
|
496
|
+
delete fs0;
|
497
|
+
}
|
498
|
+
else if (p->typ == Node::iter || p->typ == Node::opt) {
|
499
|
+
BitArray *fs0 = First0(p->sub, mark);
|
500
|
+
fs->Or(fs0);
|
501
|
+
delete fs0;
|
502
|
+
}
|
503
|
+
|
504
|
+
if (!DelNode(p)) break;
|
505
|
+
p = p->next;
|
506
|
+
}
|
507
|
+
return fs;
|
508
|
+
}
|
509
|
+
|
510
|
+
BitArray* Tab::First(Node *p) {
|
511
|
+
BitArray *mark = new BitArray(nodes->Count);
|
512
|
+
BitArray *fs = First0(p, mark);
|
513
|
+
delete mark;
|
514
|
+
if (ddt[3]) {
|
515
|
+
fwprintf(trace, L"\n");
|
516
|
+
if (p != NULL) fwprintf(trace, L"First: node = %d\n", p->n );
|
517
|
+
else fwprintf(trace, L"First: node = null\n");
|
518
|
+
PrintSet(fs, 0);
|
519
|
+
}
|
520
|
+
return fs;
|
521
|
+
}
|
522
|
+
|
523
|
+
|
524
|
+
void Tab::CompFirstSets() {
|
525
|
+
Symbol *sym;
|
526
|
+
int i;
|
527
|
+
for (i=0; i<nonterminals->Count; i++) {
|
528
|
+
sym = (Symbol*)((*nonterminals)[i]);
|
529
|
+
sym->first = new BitArray(terminals->Count);
|
530
|
+
sym->firstReady = false;
|
531
|
+
}
|
532
|
+
for (i=0; i<nonterminals->Count; i++) {
|
533
|
+
sym = (Symbol*)((*nonterminals)[i]);
|
534
|
+
sym->first = First(sym->graph);
|
535
|
+
sym->firstReady = true;
|
536
|
+
}
|
537
|
+
}
|
538
|
+
|
539
|
+
void Tab::CompFollow(Node *p) {
|
540
|
+
while (p != NULL && !((*visited)[p->n])) {
|
541
|
+
visited->Set(p->n, true);
|
542
|
+
if (p->typ == Node::nt) {
|
543
|
+
BitArray *s = First(p->next);
|
544
|
+
p->sym->follow->Or(s);
|
545
|
+
if (DelGraph(p->next))
|
546
|
+
p->sym->nts->Set(curSy->n, true);
|
547
|
+
} else if (p->typ == Node::opt || p->typ == Node::iter) {
|
548
|
+
CompFollow(p->sub);
|
549
|
+
} else if (p->typ == Node::alt) {
|
550
|
+
CompFollow(p->sub); CompFollow(p->down);
|
551
|
+
}
|
552
|
+
p = p->next;
|
553
|
+
}
|
554
|
+
}
|
555
|
+
|
556
|
+
void Tab::Complete(Symbol *sym) {
|
557
|
+
if (!((*visited)[sym->n])) {
|
558
|
+
visited->Set(sym->n, true);
|
559
|
+
Symbol *s;
|
560
|
+
for (int i=0; i<nonterminals->Count; i++) {
|
561
|
+
s = (Symbol*)((*nonterminals)[i]);
|
562
|
+
if ((*(sym->nts))[s->n]) {
|
563
|
+
Complete(s);
|
564
|
+
sym->follow->Or(s->follow);
|
565
|
+
if (sym == curSy) sym->nts->Set(s->n, false);
|
566
|
+
}
|
567
|
+
}
|
568
|
+
}
|
569
|
+
}
|
570
|
+
|
571
|
+
void Tab::CompFollowSets() {
|
572
|
+
Symbol *sym;
|
573
|
+
int i;
|
574
|
+
for (i=0; i<nonterminals->Count; i++) {
|
575
|
+
sym = (Symbol*)((*nonterminals)[i]);
|
576
|
+
sym->follow = new BitArray(terminals->Count);
|
577
|
+
sym->nts = new BitArray(nonterminals->Count);
|
578
|
+
}
|
579
|
+
gramSy->follow->Set(eofSy->n, true);
|
580
|
+
visited = new BitArray(nodes->Count);
|
581
|
+
for (i=0; i<nonterminals->Count; i++) { // get direct successors of nonterminals
|
582
|
+
sym = (Symbol*)((*nonterminals)[i]);
|
583
|
+
curSy = sym;
|
584
|
+
CompFollow(sym->graph);
|
585
|
+
}
|
586
|
+
|
587
|
+
for (i=0; i<nonterminals->Count; i++) { // add indirect successors to followers
|
588
|
+
sym = (Symbol*)((*nonterminals)[i]);
|
589
|
+
visited = new BitArray(nonterminals->Count);
|
590
|
+
curSy = sym;
|
591
|
+
Complete(sym);
|
592
|
+
}
|
593
|
+
}
|
594
|
+
|
595
|
+
Node* Tab::LeadingAny(Node *p) {
|
596
|
+
if (p == NULL) return NULL;
|
597
|
+
Node *a = NULL;
|
598
|
+
if (p->typ == Node::any) a = p;
|
599
|
+
else if (p->typ == Node::alt) {
|
600
|
+
a = LeadingAny(p->sub);
|
601
|
+
if (a == NULL) a = LeadingAny(p->down);
|
602
|
+
}
|
603
|
+
else if (p->typ == Node::opt || p->typ == Node::iter) a = LeadingAny(p->sub);
|
604
|
+
if (a == NULL && DelNode(p) && !p->up) a = LeadingAny(p->next);
|
605
|
+
return a;
|
606
|
+
}
|
607
|
+
|
608
|
+
void Tab::FindAS(Node *p) { // find ANY sets
|
609
|
+
Node *a;
|
610
|
+
while (p != NULL) {
|
611
|
+
if (p->typ == Node::opt || p->typ == Node::iter) {
|
612
|
+
FindAS(p->sub);
|
613
|
+
a = LeadingAny(p->sub);
|
614
|
+
if (a != NULL) Sets::Subtract(a->set, First(p->next));
|
615
|
+
} else if (p->typ == Node::alt) {
|
616
|
+
BitArray *s1 = new BitArray(terminals->Count);
|
617
|
+
Node *q = p;
|
618
|
+
while (q != NULL) {
|
619
|
+
FindAS(q->sub);
|
620
|
+
a = LeadingAny(q->sub);
|
621
|
+
if (a != NULL) {
|
622
|
+
BitArray *tmp = First(q->down);
|
623
|
+
tmp->Or(s1);
|
624
|
+
Sets::Subtract(a->set, tmp);
|
625
|
+
} else {
|
626
|
+
BitArray *f = First(q->sub);
|
627
|
+
s1->Or(f);
|
628
|
+
delete f;
|
629
|
+
}
|
630
|
+
q = q->down;
|
631
|
+
}
|
632
|
+
}
|
633
|
+
|
634
|
+
// Remove alternative terminals before ANY, in the following
|
635
|
+
// examples a and b must be removed from the ANY set:
|
636
|
+
// [a] ANY, or {a|b} ANY, or [a][b] ANY, or (a|) ANY, or
|
637
|
+
// A = [a]. A ANY
|
638
|
+
if (DelNode(p)) {
|
639
|
+
a = LeadingAny(p->next);
|
640
|
+
if (a != NULL) {
|
641
|
+
Node *q = (p->typ == Node::nt) ? p->sym->graph : p->sub;
|
642
|
+
Sets::Subtract(a->set, First(q));
|
643
|
+
}
|
644
|
+
}
|
645
|
+
|
646
|
+
if (p->up) break;
|
647
|
+
p = p->next;
|
648
|
+
}
|
649
|
+
}
|
650
|
+
|
651
|
+
void Tab::CompAnySets() {
|
652
|
+
Symbol *sym;
|
653
|
+
for (int i=0; i<nonterminals->Count; i++) {
|
654
|
+
sym = (Symbol*)((*nonterminals)[i]);
|
655
|
+
FindAS(sym->graph);
|
656
|
+
}
|
657
|
+
}
|
658
|
+
|
659
|
+
BitArray* Tab::Expected(Node *p, Symbol *curSy) {
|
660
|
+
BitArray *s = First(p);
|
661
|
+
if (DelGraph(p))
|
662
|
+
s->Or(curSy->follow);
|
663
|
+
return s;
|
664
|
+
}
|
665
|
+
|
666
|
+
// does not look behind resolvers; only called during LL(1) test and in CheckRes
|
667
|
+
BitArray* Tab::Expected0(Node *p, Symbol *curSy) {
|
668
|
+
if (p->typ == Node::rslv) return new BitArray(terminals->Count);
|
669
|
+
else return Expected(p, curSy);
|
670
|
+
}
|
671
|
+
|
672
|
+
void Tab::CompSync(Node *p) {
|
673
|
+
while (p != NULL && !(visited->Get(p->n))) {
|
674
|
+
visited->Set(p->n, true);
|
675
|
+
if (p->typ == Node::sync) {
|
676
|
+
BitArray *s = Expected(p->next, curSy);
|
677
|
+
s->Set(eofSy->n, true);
|
678
|
+
allSyncSets->Or(s);
|
679
|
+
p->set = s;
|
680
|
+
} else if (p->typ == Node::alt) {
|
681
|
+
CompSync(p->sub); CompSync(p->down);
|
682
|
+
} else if (p->typ == Node::opt || p->typ == Node::iter)
|
683
|
+
CompSync(p->sub);
|
684
|
+
p = p->next;
|
685
|
+
}
|
686
|
+
}
|
687
|
+
|
688
|
+
void Tab::CompSyncSets() {
|
689
|
+
allSyncSets = new BitArray(terminals->Count);
|
690
|
+
allSyncSets->Set(eofSy->n, true);
|
691
|
+
visited = new BitArray(nodes->Count);
|
692
|
+
|
693
|
+
Symbol *sym;
|
694
|
+
for (int i=0; i<nonterminals->Count; i++) {
|
695
|
+
sym = (Symbol*)((*nonterminals)[i]);
|
696
|
+
curSy = sym;
|
697
|
+
CompSync(curSy->graph);
|
698
|
+
}
|
699
|
+
}
|
700
|
+
|
701
|
+
void Tab::SetupAnys() {
|
702
|
+
Node *p;
|
703
|
+
for (int i=0; i<nodes->Count; i++) {
|
704
|
+
p = (Node*)((*nodes)[i]);
|
705
|
+
if (p->typ == Node::any) {
|
706
|
+
p->set = new BitArray(terminals->Count, true);
|
707
|
+
p->set->Set(eofSy->n, false);
|
708
|
+
}
|
709
|
+
}
|
710
|
+
}
|
711
|
+
|
712
|
+
void Tab::CompDeletableSymbols() {
|
713
|
+
bool changed;
|
714
|
+
Symbol *sym;
|
715
|
+
int i;
|
716
|
+
do {
|
717
|
+
changed = false;
|
718
|
+
for (i=0; i<nonterminals->Count; i++) {
|
719
|
+
sym = (Symbol*)((*nonterminals)[i]);
|
720
|
+
if (!sym->deletable && sym->graph != NULL && DelGraph(sym->graph)) {
|
721
|
+
sym->deletable = true; changed = true;
|
722
|
+
}
|
723
|
+
}
|
724
|
+
} while (changed);
|
725
|
+
|
726
|
+
for (i=0; i<nonterminals->Count; i++) {
|
727
|
+
sym = (Symbol*)((*nonterminals)[i]);
|
728
|
+
if (sym->deletable)
|
729
|
+
wprintf(L" %ls deletable\n", sym->name);
|
730
|
+
}
|
731
|
+
}
|
732
|
+
|
733
|
+
void Tab::RenumberPragmas() {
|
734
|
+
int n = terminals->Count;
|
735
|
+
Symbol *sym;
|
736
|
+
for (int i=0; i<pragmas->Count; i++) {
|
737
|
+
sym = (Symbol*)((*pragmas)[i]);
|
738
|
+
sym->n = n++;
|
739
|
+
}
|
740
|
+
}
|
741
|
+
|
742
|
+
void Tab::CompSymbolSets() {
|
743
|
+
CompDeletableSymbols();
|
744
|
+
CompFirstSets();
|
745
|
+
CompAnySets();
|
746
|
+
CompFollowSets();
|
747
|
+
CompSyncSets();
|
748
|
+
if (ddt[1]) {
|
749
|
+
fwprintf(trace, L"\n");
|
750
|
+
fwprintf(trace, L"First & follow symbols:\n");
|
751
|
+
fwprintf(trace, L"----------------------\n\n");
|
752
|
+
|
753
|
+
Symbol *sym;
|
754
|
+
for (int i=0; i<nonterminals->Count; i++) {
|
755
|
+
sym = (Symbol*)((*nonterminals)[i]);
|
756
|
+
fwprintf(trace, L"%ls\n", sym->name);
|
757
|
+
fwprintf(trace, L"first: "); PrintSet(sym->first, 10);
|
758
|
+
fwprintf(trace, L"follow: "); PrintSet(sym->follow, 10);
|
759
|
+
fwprintf(trace, L"\n");
|
760
|
+
}
|
761
|
+
}
|
762
|
+
if (ddt[4]) {
|
763
|
+
fwprintf(trace, L"\n");
|
764
|
+
fwprintf(trace, L"ANY and SYNC sets:\n");
|
765
|
+
fwprintf(trace, L"-----------------\n");
|
766
|
+
|
767
|
+
Node *p;
|
768
|
+
for (int i=0; i<nodes->Count; i++) {
|
769
|
+
p = (Node*)((*nodes)[i]);
|
770
|
+
if (p->typ == Node::any || p->typ == Node::sync) {
|
771
|
+
fwprintf(trace, L"%4d %4s ", p->n, nTyp[p->typ]);
|
772
|
+
PrintSet(p->set, 11);
|
773
|
+
}
|
774
|
+
}
|
775
|
+
}
|
776
|
+
}
|
777
|
+
|
778
|
+
//---------------------------------------------------------------------
|
779
|
+
// String handling
|
780
|
+
//---------------------------------------------------------------------
|
781
|
+
|
782
|
+
wchar_t Tab::Hex2Char(const wchar_t* s) {
|
783
|
+
int val = 0;
|
784
|
+
int len = coco_string_length(s);
|
785
|
+
for (int i = 0; i < len; i++) {
|
786
|
+
wchar_t ch = s[i];
|
787
|
+
if ('0' <= ch && ch <= '9') val = 16 * val + (ch - '0');
|
788
|
+
else if ('a' <= ch && ch <= 'f') val = 16 * val + (10 + ch - 'a');
|
789
|
+
else if ('A' <= ch && ch <= 'F') val = 16 * val + (10 + ch - 'A');
|
790
|
+
else parser->SemErr(L"bad escape sequence in string or character");
|
791
|
+
}
|
792
|
+
if (val >= COCO_WCHAR_MAX) {/* pdt */
|
793
|
+
parser->SemErr(L"bad escape sequence in string or character");
|
794
|
+
}
|
795
|
+
return (wchar_t) val;
|
796
|
+
}
|
797
|
+
|
798
|
+
wchar_t* Tab::Char2Hex(const wchar_t ch) {
|
799
|
+
wchar_t* format = new wchar_t[10];
|
800
|
+
coco_swprintf(format, 10, L"\\0x%04x", ch);
|
801
|
+
return format;
|
802
|
+
}
|
803
|
+
|
804
|
+
wchar_t* Tab::Unescape (const wchar_t* s) {
|
805
|
+
/* replaces escape sequences in s by their Unicode values. */
|
806
|
+
StringBuilder buf = StringBuilder();
|
807
|
+
int i = 0;
|
808
|
+
int len = coco_string_length(s);
|
809
|
+
while (i < len) {
|
810
|
+
if (s[i] == '\\') {
|
811
|
+
switch (s[i+1]) {
|
812
|
+
case L'\\': buf.Append(L'\\'); i += 2; break;
|
813
|
+
case L'\'': buf.Append(L'\''); i += 2; break;
|
814
|
+
case L'\"': buf.Append(L'\"'); i += 2; break;
|
815
|
+
case L'r': buf.Append(L'\r'); i += 2; break;
|
816
|
+
case L'n': buf.Append(L'\n'); i += 2; break;
|
817
|
+
case L't': buf.Append(L'\t'); i += 2; break;
|
818
|
+
case L'0': buf.Append(L'\0'); i += 2; break;
|
819
|
+
case L'a': buf.Append(L'\a'); i += 2; break;
|
820
|
+
case L'b': buf.Append(L'\b'); i += 2; break;
|
821
|
+
case L'f': buf.Append(L'\f'); i += 2; break;
|
822
|
+
case L'v': buf.Append(L'\v'); i += 2; break;
|
823
|
+
case L'u': case L'x':
|
824
|
+
if (i + 6 <= coco_string_length(s)) {
|
825
|
+
wchar_t *subS = coco_string_create(s, i+2, 4);
|
826
|
+
buf.Append(Hex2Char(subS)); i += 6; break;
|
827
|
+
coco_string_delete(subS);
|
828
|
+
} else {
|
829
|
+
parser->SemErr(L"bad escape sequence in string or character");
|
830
|
+
i = coco_string_length(s); break;
|
831
|
+
}
|
832
|
+
default:
|
833
|
+
parser->SemErr(L"bad escape sequence in string or character");
|
834
|
+
i += 2; break;
|
835
|
+
}
|
836
|
+
} else {
|
837
|
+
buf.Append(s[i]);
|
838
|
+
i++;
|
839
|
+
}
|
840
|
+
}
|
841
|
+
|
842
|
+
return buf.ToString();
|
843
|
+
}
|
844
|
+
|
845
|
+
|
846
|
+
wchar_t* Tab::Escape (const wchar_t* s) {
|
847
|
+
StringBuilder buf = StringBuilder();
|
848
|
+
wchar_t ch;
|
849
|
+
int len = coco_string_length(s);
|
850
|
+
for (int i=0; i < len; i++) {
|
851
|
+
ch = s[i];
|
852
|
+
switch(ch) {
|
853
|
+
case L'\\': buf.Append(L"\\\\"); break;
|
854
|
+
case L'\'': buf.Append(L"\\'"); break;
|
855
|
+
case L'\"': buf.Append(L"\\\""); break;
|
856
|
+
case L'\t': buf.Append(L"\\t"); break;
|
857
|
+
case L'\r': buf.Append(L"\\r"); break;
|
858
|
+
case L'\n': buf.Append(L"\\n"); break;
|
859
|
+
default:
|
860
|
+
if ((ch < L' ') || (ch > 0x7f)) {
|
861
|
+
wchar_t* res = Char2Hex(ch);
|
862
|
+
buf.Append(res);
|
863
|
+
delete [] res;
|
864
|
+
} else
|
865
|
+
buf.Append(ch);
|
866
|
+
break;
|
867
|
+
}
|
868
|
+
}
|
869
|
+
return buf.ToString();
|
870
|
+
}
|
871
|
+
|
872
|
+
|
873
|
+
//---------------------------------------------------------------------
|
874
|
+
// Grammar checks
|
875
|
+
//---------------------------------------------------------------------
|
876
|
+
|
877
|
+
bool Tab::GrammarOk() {
|
878
|
+
bool ok = NtsComplete()
|
879
|
+
&& AllNtReached()
|
880
|
+
&& NoCircularProductions()
|
881
|
+
&& AllNtToTerm();
|
882
|
+
if (ok) { CheckResolvers(); CheckLL1(); }
|
883
|
+
return ok;
|
884
|
+
}
|
885
|
+
|
886
|
+
|
887
|
+
//--------------- check for circular productions ----------------------
|
888
|
+
|
889
|
+
void Tab::GetSingles(Node *p, ArrayList *singles) {
|
890
|
+
if (p == NULL) return; // end of graph
|
891
|
+
if (p->typ == Node::nt) {
|
892
|
+
if (p->up || DelGraph(p->next)) singles->Add(p->sym);
|
893
|
+
} else if (p->typ == Node::alt || p->typ == Node::iter || p->typ == Node::opt) {
|
894
|
+
if (p->up || DelGraph(p->next)) {
|
895
|
+
GetSingles(p->sub, singles);
|
896
|
+
if (p->typ == Node::alt) GetSingles(p->down, singles);
|
897
|
+
}
|
898
|
+
}
|
899
|
+
if (!p->up && DelNode(p)) GetSingles(p->next, singles);
|
900
|
+
}
|
901
|
+
|
902
|
+
bool Tab::NoCircularProductions() {
|
903
|
+
bool ok, changed, onLeftSide, onRightSide;
|
904
|
+
ArrayList *list = new ArrayList();
|
905
|
+
|
906
|
+
|
907
|
+
Symbol *sym;
|
908
|
+
int i;
|
909
|
+
for (i=0; i<nonterminals->Count; i++) {
|
910
|
+
sym = (Symbol*)((*nonterminals)[i]);
|
911
|
+
ArrayList *singles = new ArrayList();
|
912
|
+
GetSingles(sym->graph, singles); // get nonterminals s such that sym-->s
|
913
|
+
Symbol *s;
|
914
|
+
for (int j=0; j<singles->Count; j++) {
|
915
|
+
s = (Symbol*)((*singles)[j]);
|
916
|
+
list->Add(new CNode(sym, s));
|
917
|
+
}
|
918
|
+
}
|
919
|
+
|
920
|
+
CNode *n;
|
921
|
+
do {
|
922
|
+
changed = false;
|
923
|
+
for (i = 0; i < list->Count; i++) {
|
924
|
+
n = (CNode*)(*list)[i];
|
925
|
+
onLeftSide = false; onRightSide = false;
|
926
|
+
CNode *m;
|
927
|
+
for (int j=0; j<list->Count; j++) {
|
928
|
+
m = (CNode*)((*list)[j]);
|
929
|
+
if (n->left == m->right) onRightSide = true;
|
930
|
+
if (n->right == m->left) onLeftSide = true;
|
931
|
+
}
|
932
|
+
if (!onLeftSide || !onRightSide) {
|
933
|
+
list->Remove(n); i--; changed = true;
|
934
|
+
}
|
935
|
+
}
|
936
|
+
} while(changed);
|
937
|
+
ok = true;
|
938
|
+
|
939
|
+
for (i=0; i<list->Count; i++) {
|
940
|
+
n = (CNode*)((*list)[i]);
|
941
|
+
ok = false; errors->count++;
|
942
|
+
wprintf(L" %ls --> %ls", n->left->name, n->right->name);
|
943
|
+
}
|
944
|
+
return ok;
|
945
|
+
}
|
946
|
+
|
947
|
+
|
948
|
+
//--------------- check for LL(1) errors ----------------------
|
949
|
+
|
950
|
+
void Tab::LL1Error(int cond, Symbol *sym) {
|
951
|
+
wprintf(L" LL1 warning in %ls: ", curSy->name);
|
952
|
+
if (sym != NULL) wprintf(L"%ls is ", sym->name);
|
953
|
+
switch (cond) {
|
954
|
+
case 1: wprintf(L"start of several alternatives\n"); break;
|
955
|
+
case 2: wprintf(L"start & successor of deletable structure\n"); break;
|
956
|
+
case 3: wprintf(L"an ANY node that matches no symbol\n"); break;
|
957
|
+
case 4: wprintf(L"contents of [...] or {...} must not be deletable\n"); break;
|
958
|
+
}
|
959
|
+
}
|
960
|
+
|
961
|
+
|
962
|
+
void Tab::CheckOverlap(BitArray *s1, BitArray *s2, int cond) {
|
963
|
+
Symbol *sym;
|
964
|
+
for (int i=0; i<terminals->Count; i++) {
|
965
|
+
sym = (Symbol*)((*terminals)[i]);
|
966
|
+
if ((*s1)[sym->n] && (*s2)[sym->n]) {
|
967
|
+
LL1Error(cond, sym);
|
968
|
+
}
|
969
|
+
}
|
970
|
+
}
|
971
|
+
|
972
|
+
void Tab::CheckAlts(Node *p) {
|
973
|
+
BitArray *s1, *s2;
|
974
|
+
while (p != NULL) {
|
975
|
+
if (p->typ == Node::alt) {
|
976
|
+
Node *q = p;
|
977
|
+
s1 = new BitArray(terminals->Count);
|
978
|
+
while (q != NULL) { // for all alternatives
|
979
|
+
s2 = Expected0(q->sub, curSy);
|
980
|
+
CheckOverlap(s1, s2, 1);
|
981
|
+
s1->Or(s2);
|
982
|
+
CheckAlts(q->sub);
|
983
|
+
q = q->down;
|
984
|
+
}
|
985
|
+
} else if (p->typ == Node::opt || p->typ == Node::iter) {
|
986
|
+
if (DelSubGraph(p->sub)) LL1Error(4, NULL); // e.g. [[...]]
|
987
|
+
else {
|
988
|
+
s1 = Expected0(p->sub, curSy);
|
989
|
+
s2 = Expected(p->next, curSy);
|
990
|
+
CheckOverlap(s1, s2, 2);
|
991
|
+
}
|
992
|
+
CheckAlts(p->sub);
|
993
|
+
} else if (p->typ == Node::any) {
|
994
|
+
if (Sets::Elements(p->set) == 0) LL1Error(3, NULL);
|
995
|
+
// e.g. {ANY} ANY or [ANY] ANY or ( ANY | ANY )
|
996
|
+
}
|
997
|
+
if (p->up) break;
|
998
|
+
p = p->next;
|
999
|
+
}
|
1000
|
+
}
|
1001
|
+
|
1002
|
+
void Tab::CheckLL1() {
|
1003
|
+
Symbol *sym;
|
1004
|
+
for (int i=0; i<nonterminals->Count; i++) {
|
1005
|
+
sym = (Symbol*)((*nonterminals)[i]);
|
1006
|
+
curSy = sym;
|
1007
|
+
CheckAlts(curSy->graph);
|
1008
|
+
}
|
1009
|
+
}
|
1010
|
+
|
1011
|
+
//------------- check if resolvers are legal --------------------
|
1012
|
+
|
1013
|
+
void Tab::ResErr(Node *p, const wchar_t* msg) {
|
1014
|
+
errors->Warning(p->line, p->pos->col, msg);
|
1015
|
+
}
|
1016
|
+
|
1017
|
+
void Tab::CheckRes(Node *p, bool rslvAllowed) {
|
1018
|
+
while (p != NULL) {
|
1019
|
+
|
1020
|
+
Node *q;
|
1021
|
+
if (p->typ == Node::alt) {
|
1022
|
+
BitArray *expected = new BitArray(terminals->Count);
|
1023
|
+
for (q = p; q != NULL; q = q->down)
|
1024
|
+
expected->Or(Expected0(q->sub, curSy));
|
1025
|
+
BitArray *soFar = new BitArray(terminals->Count);
|
1026
|
+
for (q = p; q != NULL; q = q->down) {
|
1027
|
+
if (q->sub->typ == Node::rslv) {
|
1028
|
+
BitArray *fs = Expected(q->sub->next, curSy);
|
1029
|
+
if (Sets::Intersect(fs, soFar))
|
1030
|
+
ResErr(q->sub, L"Warning: Resolver will never be evaluated. Place it at previous conflicting alternative.");
|
1031
|
+
if (!Sets::Intersect(fs, expected))
|
1032
|
+
ResErr(q->sub, L"Warning: Misplaced resolver: no LL(1) conflict.");
|
1033
|
+
} else soFar->Or(Expected(q->sub, curSy));
|
1034
|
+
CheckRes(q->sub, true);
|
1035
|
+
}
|
1036
|
+
} else if (p->typ == Node::iter || p->typ == Node::opt) {
|
1037
|
+
if (p->sub->typ == Node::rslv) {
|
1038
|
+
BitArray *fs = First(p->sub->next);
|
1039
|
+
BitArray *fsNext = Expected(p->next, curSy);
|
1040
|
+
if (!Sets::Intersect(fs, fsNext))
|
1041
|
+
ResErr(p->sub, L"Warning: Misplaced resolver: no LL(1) conflict.");
|
1042
|
+
}
|
1043
|
+
CheckRes(p->sub, true);
|
1044
|
+
} else if (p->typ == Node::rslv) {
|
1045
|
+
if (!rslvAllowed)
|
1046
|
+
ResErr(p, L"Warning: Misplaced resolver: no alternative.");
|
1047
|
+
}
|
1048
|
+
|
1049
|
+
if (p->up) break;
|
1050
|
+
p = p->next;
|
1051
|
+
rslvAllowed = false;
|
1052
|
+
}
|
1053
|
+
}
|
1054
|
+
|
1055
|
+
void Tab::CheckResolvers() {
|
1056
|
+
for (int i=0; i<nonterminals->Count; i++) {
|
1057
|
+
curSy = (Symbol*)((*nonterminals)[i]);
|
1058
|
+
CheckRes(curSy->graph, false);
|
1059
|
+
}
|
1060
|
+
}
|
1061
|
+
|
1062
|
+
|
1063
|
+
//------------- check if every nts has a production --------------------
|
1064
|
+
|
1065
|
+
bool Tab::NtsComplete() {
|
1066
|
+
bool complete = true;
|
1067
|
+
Symbol *sym;
|
1068
|
+
for (int i=0; i<nonterminals->Count; i++) {
|
1069
|
+
sym = (Symbol*)((*nonterminals)[i]);
|
1070
|
+
if (sym->graph == NULL) {
|
1071
|
+
complete = false; errors->count++;
|
1072
|
+
wprintf(L" No production for %ls\n", sym->name);
|
1073
|
+
}
|
1074
|
+
}
|
1075
|
+
return complete;
|
1076
|
+
}
|
1077
|
+
|
1078
|
+
//-------------- check if every nts can be reached -----------------
|
1079
|
+
|
1080
|
+
void Tab::MarkReachedNts(Node *p) {
|
1081
|
+
while (p != NULL) {
|
1082
|
+
if (p->typ == Node::nt && !((*visited)[p->sym->n])) { // new nt reached
|
1083
|
+
visited->Set(p->sym->n, true);
|
1084
|
+
MarkReachedNts(p->sym->graph);
|
1085
|
+
} else if (p->typ == Node::alt || p->typ == Node::iter || p->typ == Node::opt) {
|
1086
|
+
MarkReachedNts(p->sub);
|
1087
|
+
if (p->typ == Node::alt) MarkReachedNts(p->down);
|
1088
|
+
}
|
1089
|
+
if (p->up) break;
|
1090
|
+
p = p->next;
|
1091
|
+
}
|
1092
|
+
}
|
1093
|
+
|
1094
|
+
bool Tab::AllNtReached() {
|
1095
|
+
bool ok = true;
|
1096
|
+
visited = new BitArray(nonterminals->Count);
|
1097
|
+
visited->Set(gramSy->n, true);
|
1098
|
+
MarkReachedNts(gramSy->graph);
|
1099
|
+
Symbol *sym;
|
1100
|
+
for (int i=0; i<nonterminals->Count; i++) {
|
1101
|
+
sym = (Symbol*)((*nonterminals)[i]);
|
1102
|
+
if (!((*visited)[sym->n])) {
|
1103
|
+
ok = false; errors->count++;
|
1104
|
+
wprintf(L" %ls cannot be reached\n", sym->name);
|
1105
|
+
}
|
1106
|
+
}
|
1107
|
+
return ok;
|
1108
|
+
}
|
1109
|
+
|
1110
|
+
//--------- check if every nts can be derived to terminals ------------
|
1111
|
+
|
1112
|
+
bool Tab::IsTerm(Node *p, BitArray *mark) { // true if graph can be derived to terminals
|
1113
|
+
while (p != NULL) {
|
1114
|
+
if (p->typ == Node::nt && !((*mark)[p->sym->n])) return false;
|
1115
|
+
if (p->typ == Node::alt && !IsTerm(p->sub, mark)
|
1116
|
+
&& (p->down == NULL || !IsTerm(p->down, mark))) return false;
|
1117
|
+
if (p->up) break;
|
1118
|
+
p = p->next;
|
1119
|
+
}
|
1120
|
+
return true;
|
1121
|
+
}
|
1122
|
+
|
1123
|
+
|
1124
|
+
bool Tab::AllNtToTerm() {
|
1125
|
+
bool changed, ok = true;
|
1126
|
+
BitArray *mark = new BitArray(nonterminals->Count);
|
1127
|
+
// a nonterminal is marked if it can be derived to terminal symbols
|
1128
|
+
Symbol *sym;
|
1129
|
+
int i;
|
1130
|
+
do {
|
1131
|
+
changed = false;
|
1132
|
+
|
1133
|
+
for (i=0; i<nonterminals->Count; i++) {
|
1134
|
+
sym = (Symbol*)((*nonterminals)[i]);
|
1135
|
+
if (!((*mark)[sym->n]) && IsTerm(sym->graph, mark)) {
|
1136
|
+
mark->Set(sym->n, true); changed = true;
|
1137
|
+
}
|
1138
|
+
}
|
1139
|
+
} while (changed);
|
1140
|
+
for (i=0; i<nonterminals->Count; i++) {
|
1141
|
+
sym = (Symbol*)((*nonterminals)[i]);
|
1142
|
+
if (!((*mark)[sym->n])) {
|
1143
|
+
ok = false; errors->count++;
|
1144
|
+
wprintf(L" %ls cannot be derived to terminals\n", sym->name);
|
1145
|
+
}
|
1146
|
+
}
|
1147
|
+
return ok;
|
1148
|
+
}
|
1149
|
+
|
1150
|
+
//---------------------------------------------------------------------
|
1151
|
+
// Cross reference list
|
1152
|
+
//---------------------------------------------------------------------
|
1153
|
+
|
1154
|
+
void Tab::XRef() {
|
1155
|
+
SortedList *xref = new SortedList();
|
1156
|
+
// collect lines where symbols have been defined
|
1157
|
+
Symbol *sym;
|
1158
|
+
int i, j;
|
1159
|
+
for (i=0; i<nonterminals->Count; i++) {
|
1160
|
+
sym = (Symbol*)((*nonterminals)[i]);
|
1161
|
+
ArrayList *list = (ArrayList*)(xref->Get(sym));
|
1162
|
+
if (list == NULL) {list = new ArrayList(); xref->Set(sym, list);}
|
1163
|
+
int *intg = new int(- sym->line);
|
1164
|
+
list->Add(intg);
|
1165
|
+
}
|
1166
|
+
// collect lines where symbols have been referenced
|
1167
|
+
Node *n;
|
1168
|
+
for (i=0; i<nodes->Count; i++) {
|
1169
|
+
n = (Node*)((*nodes)[i]);
|
1170
|
+
if (n->typ == Node::t || n->typ == Node::wt || n->typ == Node::nt) {
|
1171
|
+
ArrayList *list = (ArrayList*)(xref->Get(n->sym));
|
1172
|
+
if (list == NULL) {list = new ArrayList(); xref->Set(n->sym, list);}
|
1173
|
+
int *intg = new int(n->line);
|
1174
|
+
list->Add(intg);
|
1175
|
+
}
|
1176
|
+
}
|
1177
|
+
// print cross reference list
|
1178
|
+
fwprintf(trace, L"\n");
|
1179
|
+
fwprintf(trace, L"Cross reference list:\n");
|
1180
|
+
fwprintf(trace, L"--------------------\n\n");
|
1181
|
+
|
1182
|
+
for (i=0; i<xref->Count; i++) {
|
1183
|
+
sym = (Symbol*)(xref->GetKey(i));
|
1184
|
+
wchar_t *paddedName = Name(sym->name);
|
1185
|
+
fwprintf(trace, L" %12ls", paddedName);
|
1186
|
+
coco_string_delete(paddedName);
|
1187
|
+
ArrayList *list = (ArrayList*)(xref->Get(sym));
|
1188
|
+
int col = 14;
|
1189
|
+
int line;
|
1190
|
+
for (j=0; j<list->Count; j++) {
|
1191
|
+
line = *(int*)((*list)[j]);
|
1192
|
+
if (col + 5 > 80) {
|
1193
|
+
fwprintf(trace, L"\n");
|
1194
|
+
for (col = 1; col <= 14; col++) fwprintf(trace, L" ");
|
1195
|
+
}
|
1196
|
+
fwprintf(trace, L"%5d", line); col += 5;
|
1197
|
+
}
|
1198
|
+
fwprintf(trace, L"\n");
|
1199
|
+
}
|
1200
|
+
fwprintf(trace, L"\n\n");
|
1201
|
+
}
|
1202
|
+
|
1203
|
+
void Tab::SetDDT(const wchar_t* s) {
|
1204
|
+
wchar_t* st = coco_string_create_upper(s);
|
1205
|
+
wchar_t ch;
|
1206
|
+
int len = coco_string_length(st);
|
1207
|
+
for (int i = 0; i < len; i++) {
|
1208
|
+
ch = st[i];
|
1209
|
+
if (L'0' <= ch && ch <= L'9') ddt[ch - L'0'] = true;
|
1210
|
+
else switch (ch) {
|
1211
|
+
case L'A' : ddt[0] = true; break; // trace automaton
|
1212
|
+
case L'F' : ddt[1] = true; break; // list first/follow sets
|
1213
|
+
case L'G' : ddt[2] = true; break; // print syntax graph
|
1214
|
+
case L'I' : ddt[3] = true; break; // trace computation of first sets
|
1215
|
+
case L'J' : ddt[4] = true; break; // print ANY and SYNC sets
|
1216
|
+
case L'P' : ddt[8] = true; break; // print statistics
|
1217
|
+
case L'S' : ddt[6] = true; break; // list symbol table
|
1218
|
+
case L'X' : ddt[7] = true; break; // list cross reference table
|
1219
|
+
default : break;
|
1220
|
+
}
|
1221
|
+
}
|
1222
|
+
coco_string_delete(st);
|
1223
|
+
}
|
1224
|
+
|
1225
|
+
|
1226
|
+
void Tab::SetOption(const wchar_t* s) {
|
1227
|
+
// example: $namespace=xxx
|
1228
|
+
// index of '=' is 10 => nameLenght = 10
|
1229
|
+
// start index of xxx = 11
|
1230
|
+
|
1231
|
+
int nameLenght = coco_string_indexof(s, '=');
|
1232
|
+
int valueIndex = nameLenght + 1;
|
1233
|
+
|
1234
|
+
wchar_t *name = coco_string_create(s, 0, nameLenght);
|
1235
|
+
wchar_t *value = coco_string_create(s, valueIndex);
|
1236
|
+
|
1237
|
+
if (coco_string_equal(L"$namespace", name)) {
|
1238
|
+
if (nsName == NULL) nsName = coco_string_create(value);
|
1239
|
+
} else if (coco_string_equal(L"$checkEOF", name)) {
|
1240
|
+
checkEOF = coco_string_equal(L"true", value);
|
1241
|
+
}
|
1242
|
+
|
1243
|
+
delete [] name;
|
1244
|
+
delete [] value;
|
1245
|
+
}
|
1246
|
+
|
1247
|
+
|
1248
|
+
}; // namespace
|