* src/reduce.c (nonterminals_reduce): Don't loop over RITEM: loop
over the RHS of each rule.
* src/gram.h, src/gram.c (nritems): Is `unsigned int', not int.
* src/state.h (state_t): Member `nitems' is unsigned short.
* src/LR0.c (get_state): Adjust.
* src/reader.c (packgram): Likewise.
* src/output.c (GENERATE_MUSCLE_INSERT_TABLE): `max' is of type
`Type'.
(muscle_insert_int_table): Remove, unused.
(prepare_rules): Remove `max'.
RITEM. Note that it must be able to code symbol numbers as
positive number, and the negation of rule numbers as negative
numbers.
Adjust all dependencies (pretty many).
* src/reduce.c (rule): Remove this `short *' pointer: use
item_number_t.
* src/system.h (MINSHORT, MAXSHORT): Remove.
Include `limits.h'.
Adjust dependencies to using SHRT_MAX and SHRT_MIN.
(shortcpy): Remove.
(MAXTABLE): Move to...
* src/output.c (MAXTABLE): here.
(prepare_rules): Use output_int_table to output rhs.
* data/bison.simple, data/bison.c++: Adjust.
* tests/torture.at (Big triangle): Move the limit from 254 to
500.
* tests/regression.at (Web2c Actions): Ajust.
Trying with bigger grammars shows various phenomena: at 3000 (28Mb
of grammar file) bison is killed by my system, at 2000 (12Mb) bison
passes, but produces negative #line number, once fixed, GCC is
killed while compiling 14Mb, at 1500 (6.7 Mb of grammar, 8.2Mb of
C), it passes.
* src/state.h (state_h): Code input lines on ints, not shorts.
* src/gram.c (ntokens): Initialize to 1, to reserve a slot for
EOF.
* src/lex.c (lex): Set the `number' member of new terminals.
* src/reader.c (bucket_check_defined, bucket_make_alias)
(bucket_check_alias_consistence, bucket_translation): New.
(reader, grammar_free, readgram, token_translations_init)
(packsymbols): Adjust.
(reader): Number the predefined tokens.
* src/reduce.c (inaccessable_symbols): Just use hard coded numbers
for predefined tokens.
* src/symtab.h (bucket): Remove all the hash table related
members.
* src/symtab.c (symtab): Replace by...
(bucket_table): this.
(bucket_new, bucket_free, hash_compare_bucket, hash_bucket)
(buckets_new, buckets_do): New.
bucket.
Adjust all dependencies.
* src/reduce.c (nonterminals_reduce): Don't forget to renumber the
`number' of the buckets too.
* src/gram.h: Include `symtab.h'.
(associativity): Move to...
* src/symtab.h: here.
No longer include `gram.h'.
(ritem_longest_rhs): Use it.
* src/gram.h (rule_t): `number' is a new member.
* src/reader.c (packgram): Set it.
* src/reduce.c (reduce_grammar_tables): Move the useless rules at
the end of `rules', and count them out of `nrules'.
(reduce_output, dump_grammar): Adjust.
* src/print.c (print_grammar): It is no longer needed to check for
the usefulness of a rule, as useless rules are beyond `nrules + 1'.
* tests/reduce.at (Reduced Automaton): New test.
* src/nullable.c, src/output.c, src/print.c, src/print_graph.c,
* src/reader.c, src/reduce.c: Let rule_t.rhs point directly to the
RHS, instead of being an index in RITEMS.
and rule_table.
* src/reader.c (packsymbols): Fill this table.
Drop sprec.
* src/conflicts.c (resolve_sr_conflict): Adjust.
* src/reduce.c (reduce_grammar): Adjust: just sort symbols, a
single table.
Use symbols[i]->tag instead of tags[i].
when walking through ritem, even via rule->rhs.
* src/reduce.c (dump_grammar, useful_production, reduce_output)
(useful_production, useless_nonterminals): Likewise.
(reduce_grammar_tables): Likewise, plus update nritems.
* src/nullable.c (set_nullable): Likewise.
* src/lalr.c (build_relations): Likewise.
* tests/sets.at (Nullable): Adjust.
Fortunately, now, the $axiom is no longer nullable.
is reached by eoftoken.
(insert_start_shifting_state, insert_eof_shifting_state)
(insert_accepting_state, augment_automaton): Remove, since now
these states are automatically computed from the initial state.
(generate_states): Adjust.
* src/print.c: When reporting a rule number to the user, substract
1, so that the axiom rule is rule 0, and the first user rule is 1.
* src/reduce.c: Likewise.
* src/print_graph.c (print_core): For the time being, just as for
the report, depend upon --trace-flags to dump the full set of
items.
* src/reader.c (readgram): Once the grammar read, insert the rule
0: `$axiom: START-SYMBOL $'.
* tests/set.at: Adjust: rule 0 is now displayed, and since the
number of the states has changed (the final state is no longer
necessarily the last), catch up.
removal of useless rules via CPP but via `if (0)', so that the
compiler still check the code is valid.
For instance, it should have noticed `rline' no longer exists: use
the `line' member of rule_t.
* src/gram.c (dummy, rline): Remove, unused.
`destroyed'), we now sometimes examine them, and store information
related to them. Hence we need to know their number, and adjust
memory allocations.
* src/reduce.c, src/reduce.h (nuseless_nonterminals): No longer
static.
* src/LR0.c (allocate_itemsets): The memory allocated to
`symbol_count' was used for two different purpose: once to count
the number of occurrences of each symbol, and later reassigned to
`shift_symbol', containing the symbol that can be shifted from a
given state.
Deobfuscate, i.e., allocate, use and free `symbol_count' here
only, and...
(new_itemsets): Allocate `shift_symbol' here.
(allocate_itemsets): symbol_count includes useless nonterminals.
Make room for them.
(free_storage): Use `free', not `XFREE', for pointers that cannot
be null.
* src/gram.c (dummy): Remove, now there is actual code in gram.c.
(This useless function was defined only to work around VMS linkers
that can't handle compilation units with variables only).
* src/reduce.c (dump_grammar): Use it to trace the construction of
ritem.
* src/closure.c: Include `reader.h' to get `tags', needed by the
trace code.
Rename the conditional DEBUG as TRACE.
Output consistently TRACEs to stderr, not stdout.
* src/derives.c: Likewise.
* src/reduce.c: (inaccessable_symbols): Using if is better style
than goto.
Use `#if TRACE' instead of `#if 0' for tracing code.
(reduce_output): This.
Output to OUT, passed as argument, instead of output_obstack.
(dump_grammar): Likewise.
(reduce_free): New.
Also free V1.
(reduce_grammar): No longer call reduce_output, since...
* src/print.c (print_results): do it.
* src/main.c (main): Call reduce_free;
* files.c (foutput): Remove.
(output_obstack): New.
Adjust all dependencies.
* src/conflicts.c: Return a string.
* src/system.h (obstack_grow_string): Rename as...
(obstack_sgrow): this. Be ready to work with non literals.
(obstack_fgrow4): New.
messages. Nevertheless, it is enabled only when `@n' is used in
the grammar, which is a different issue (you can use it in error
message, but not in the grammar per se). Therefore, there should
be another means to enable it.
* src/getargs.c (getargs): Support `--locations'.
(usage): Report it.
* src/getargs.h (locationsflag): Export it.
* src/lex.c (percent_table): Support `%locations'.
* src/reader.c (yylsp_needed): Remove this variable, now replaced
with `locationsflag'.
* doc/bison.texinfo: Document `--locations' and `%locations'.
Sort the options.
* tests/calc.m4: Test it.
For regularity of the names, replace each
(nolineflag, toknumflag, rawtokenumflag, noparserflag): with...
(no_lineflag, token_tableflag, rawflag, no_parserflag): this.
In addition replace each `flag' with `_flag'.
Bison's `allocate.c' and `alloc.h'. This patch was surprisingly
difficult and introduced a lot of core dump. It turns out that
Bison used an implementation of `xmalloc' based on `calloc', and
at various places it does depend upon the initialization to 0. I
have not tried to isolate the pertinent places, and all the former
calls to Bison's `xmalloc' are now using `XCALLOC'. Someday,
someone should address this issue.
* src/allocate.c, src/alloc.h, m4/bison-decl.m4: Remove.
* lib/xmalloc.c, lib/xalloc.h, m4/malloc.m4, m4/realloc.m4: New
files.
Adjust dependencies.
* src/warshall.h: New file.
Propagate.
