parsergen: flip ordering of precedence declarations.

[ocean] / csrc / parsergen.mdc

diff --git a/csrc/parsergen.mdc b/csrc/parsergen.mdc
index 478482e0f6b0512b9fc09946b4b1b5ead04fe7fd..d12a4000d3712f6a44eaf486a284793073e1abf8 100644 (file)
--- a/csrc/parsergen.mdc
+++ b/csrc/parsergen.mdc
@@ -277,6 +277,9 @@ declares how it associates.  This level is stored in each symbol
 listed and may be inherited by any production which uses the symbol.  A
 production inherits from the last symbol which has a precedence.
 
+The symbols on the first precedence line have the lowest precedence.
+Subsequent lines introduce symbols with higher precedence.
+
 ###### grammar fields
        struct text current_type;
        int type_isref;
@@ -1287,12 +1290,11 @@ When itemsets are created we assign a precedence to the itemset from
 the complete item, if there is one.  We ignore the possibility of
 there being two and don't (currently) handle precedence in such
 grammars.  When completing a grammar we ignore any item where DOT is
-followed by a terminal with a precedence lower (numerically higher)
-than that for the itemset.  Unless the terminal has right
-associativity, we also ignore items where the terminal has the same
-precedence.  The result is that unwanted items are still in the
-itemset, but the terminal doesn't get into the go to set, so the item
-is ineffective.
+followed by a terminal with a precedence lower than that for the
+itemset.  Unless the terminal has right associativity, we also ignore
+items where the terminal has the same precedence.  The result is that
+unwanted items are still in the itemset, but the terminal doesn't get
+into the go to set, so the item is ineffective.
 
 ###### complete itemset
        for (i = 0; i < is->items.cnt; i++) {
@@ -1311,7 +1313,7 @@ is ineffective.
                        continue;
                s = pr->body[bs];
                if (s->precedence && is->precedence &&
-                   is->precedence < s->precedence)
+                   is->precedence > s->precedence)
                        /* This terminal has a low precedence and
                         * shouldn't be shifted
                         */
@@ -1404,8 +1406,7 @@ with a pre-existing itemset).
                        pos = symset_find(&newitemset, pr->head->num);
                        if (bp + 1 == pr->body_size &&
                            pr->precedence > 0 &&
-                           (precedence == 0 ||
-                            pr->precedence < precedence)) {
+                           pr->precedence > precedence) {
                                // new itemset is reducible and has a precedence.
                                precedence = pr->precedence;
                                assoc = pr->assoc;
@@ -1900,7 +1901,9 @@ pieces of code provided in the grammar file, so they are generated first.
 ### Known words table
 
 The known words table is simply an array of terminal symbols.
-The table of nonterminals used for tracing is a similar array.
+The table of nonterminals used for tracing is a similar array.  We
+include virtual symbols in the table of non_terminals to keep the
+numbers right.
 
 ###### functions
 
@@ -1925,7 +1928,7 @@ The table of nonterminals used for tracing is a similar array.
                for (i = TK_reserved;
                     i < g->num_syms;
                     i++)
-                       if (g->symtab[i]->type == Nonterminal)
+                       if (g->symtab[i]->type != Terminal)
                                fprintf(f, "\t\"%.*s\",\n", g->symtab[i]->name.len,
                                        g->symtab[i]->name.txt);
                fprintf(f, "};\n\n");
@@ -3087,8 +3090,8 @@ an error.
 
 # calc: grammar
 
-       $LEFT * /
        $LEFT + -
+       $LEFT * /
 
        Session -> Session Line
                | Line