summaryrefslogtreecommitdiff
path: root/grammars
diff options
context:
space:
mode:
Diffstat (limited to 'grammars')
-rw-r--r--grammars/c/disambiguate.c49
-rw-r--r--grammars/c/grammar.txt17
-rw-r--r--grammars/expression/disambiguate.c49
-rw-r--r--grammars/expression/grammar.txt18
4 files changed, 133 insertions, 0 deletions
diff --git a/grammars/c/disambiguate.c b/grammars/c/disambiguate.c
new file mode 100644
index 0000000..9a8bf08
--- /dev/null
+++ b/grammars/c/disambiguate.c
@@ -0,0 +1,49 @@
+int disambiguator(struct state *old, struct state *new) {
+ // printf("Old tree: ");
+ // print_parse_tree2(old);
+ // printf("New tree: ");
+ // print_parse_tree2(new);
+
+ if (old->start_idx != new->start_idx) {
+ // printf("\t\tIGNORING "); print_parse_tree2(old);
+ // printf("\t\tVS: "); print_parse_tree2(new);
+ return 2;
+ }
+
+ // Prefer the earlier parsings in the grammar when two entirely different
+ // productions are taken.
+ if (old->production_id != new->production_id)
+ return old->production_id < new->production_id
+ ? 0 : 1;
+
+ // If they're the same production ...
+ prod_id_t prod = old->production_id;
+ if (PRODUCTION_ID_TO_SYMBOL[prod] == SYMBOL_EXPR) {
+ if (PRODUCTION_ID_TO_PRODUCTION[prod][1] == SYMBOL_OP) {
+ struct token *old_tok = find_token(old, 1),
+ *new_tok = find_token(new, 1);
+ char *old_s = old_tok->string, *new_s = new_tok->string;
+ const char *precedence[] = {".", "->", "*", "/", "%", "+", "-",
+ "<<", ">>", "<", "<=", ">", ">=", "==", "!=", "&", "|", "&&",
+ "||", "=", "+=", "-=", "*=", "/=", "%=", "<<=", ">>=", "&=",
+ "^=", "|=", ",", 0};
+ if (strcmp(old_s, new_s)) {
+ for (const char **p = precedence; *p; p++) {
+ if (!strcmp(old_s, *p)) {
+ return 1;
+ } else if (!strcmp(new_s, *p)) {
+ return 0;
+ }
+ }
+ // BAD!
+ return 2;
+ } else {
+ // Associate RIGHT
+ if (old_tok < new_tok) return 1;
+ else if (old_tok > new_tok) return 0;
+ }
+ }
+ }
+ printf("TOTALLY UNKNOWN!\n");
+ return 2;
+}
diff --git a/grammars/c/grammar.txt b/grammars/c/grammar.txt
new file mode 100644
index 0000000..486f319
--- /dev/null
+++ b/grammars/c/grammar.txt
@@ -0,0 +1,17 @@
+KEYWORDS list
+ switch volatile case while do else const for if
+
+IDENT regex
+ [a-zA-Z_][0-9a-zA-Z_]*
+
+INT regex
+ [0-9]+
+
+OPS list
+ ( ) { } [ ]
+ ; ,
+ - + ! % * & / << >> ^ |
+ -= += != %= *= &= /= <<= == >>= ^= |=
+ && || ++ --
+ < <= > >= =
+ . -> ? :
diff --git a/grammars/expression/disambiguate.c b/grammars/expression/disambiguate.c
new file mode 100644
index 0000000..9a8bf08
--- /dev/null
+++ b/grammars/expression/disambiguate.c
@@ -0,0 +1,49 @@
+int disambiguator(struct state *old, struct state *new) {
+ // printf("Old tree: ");
+ // print_parse_tree2(old);
+ // printf("New tree: ");
+ // print_parse_tree2(new);
+
+ if (old->start_idx != new->start_idx) {
+ // printf("\t\tIGNORING "); print_parse_tree2(old);
+ // printf("\t\tVS: "); print_parse_tree2(new);
+ return 2;
+ }
+
+ // Prefer the earlier parsings in the grammar when two entirely different
+ // productions are taken.
+ if (old->production_id != new->production_id)
+ return old->production_id < new->production_id
+ ? 0 : 1;
+
+ // If they're the same production ...
+ prod_id_t prod = old->production_id;
+ if (PRODUCTION_ID_TO_SYMBOL[prod] == SYMBOL_EXPR) {
+ if (PRODUCTION_ID_TO_PRODUCTION[prod][1] == SYMBOL_OP) {
+ struct token *old_tok = find_token(old, 1),
+ *new_tok = find_token(new, 1);
+ char *old_s = old_tok->string, *new_s = new_tok->string;
+ const char *precedence[] = {".", "->", "*", "/", "%", "+", "-",
+ "<<", ">>", "<", "<=", ">", ">=", "==", "!=", "&", "|", "&&",
+ "||", "=", "+=", "-=", "*=", "/=", "%=", "<<=", ">>=", "&=",
+ "^=", "|=", ",", 0};
+ if (strcmp(old_s, new_s)) {
+ for (const char **p = precedence; *p; p++) {
+ if (!strcmp(old_s, *p)) {
+ return 1;
+ } else if (!strcmp(new_s, *p)) {
+ return 0;
+ }
+ }
+ // BAD!
+ return 2;
+ } else {
+ // Associate RIGHT
+ if (old_tok < new_tok) return 1;
+ else if (old_tok > new_tok) return 0;
+ }
+ }
+ }
+ printf("TOTALLY UNKNOWN!\n");
+ return 2;
+}
diff --git a/grammars/expression/grammar.txt b/grammars/expression/grammar.txt
new file mode 100644
index 0000000..a1590fa
--- /dev/null
+++ b/grammars/expression/grammar.txt
@@ -0,0 +1,18 @@
+INT regex
+ [0-9]+
+
+OP list
+ || && -- ++ = == !
+ / * - + << >> % ^
+ /= *= -= += <<= >>= &= |= %= ^=
+ -> .
+ < > <= >= !=
+ ? :
+ ( ) { } [ ]
+ , ;
+
+EXPR nonterm .start
+ EXPR OP EXPR
+ INT
+ ( EXPR )
+ EXPR ? EXPR : EXPR
generated by cgit on debian on lair
contact matthew@masot.net with questions or feedback