After all, why not?
* src/reader.c (switching_token): Use symbol_id_get.
(check_start_symbols): Require that the start symbol is a token only
if it's the only one.
* examples/c/lexcalc/parse.y: Let NUM be a start symbol.
For each start symbol, generate a parsing function with a richer
return value than the usual of yyparse. Reserve a place for the
returned semantic value, in order to avoid having to pass a pointer as
argument to "return" that value. This also makes the call to the
parsing function independent of whether a given start-symbol is typed.
For instance, if the grammar file contains:
%type <int> expression
%start input expression
(so "input" is valueless) we get
typedef struct
{
int yystatus;
} yyparse_input_t;
yyparse_input_t yyparse_input (void);
typedef struct
{
int yyvalue;
int yystatus;
} yyparse_expression_t;
yyparse_expression_t yyparse_expression (void);
This commit also changes the implementation of the parser termination:
when there are multiple start symbols, it is the initial rules that
explicitly YYACCEPT. They do that after having exported the
start-symbol's value (if it is typed):
switch (yyn)
{
case 1: /* $accept: YY_EXPRESSION expression $end */
{ ((*yyvalue).TOK_expression) = (yyvsp[-1].TOK_expression); YYACCEPT; }
break;
case 2: /* $accept: YY_INPUT input $end */
{ YYACCEPT; }
break;
I have tried several ways to deal with termination, and this is the
one that appears the best one to me. It is also the most natural.
* src/scan-code.h, src/scan-code.l (obstack_for_actions): New.
* src/reader.c (grammar_rule_check_and_complete): Generate the actions
of the rules for each start symbol.
* data/skeletons/bison.m4 (b4_symbol_slot): New, with safer semantics
than type and type_tag.
* data/skeletons/yacc.c (b4_accept): New.
Generates the body of the action of the start rules.
(_b4_declare_sub_yyparse): For each start symbol define a dedicated
return type for its parsing function.
Adjust the declaration of its parsing function.
(_b4_define_sub_yyparse): Adjust the definition of the function.
* examples/c/lexcalc/parse.y: Check the case of valueless symbols.
* examples/c/lexcalc/lexcalc.test: Check start symbols.
So far we were not checking the generated rule 0 at all. Now there
can be several of them. Instead of not checking at all, let's be more
selective on the check to run on them.
* src/reader.c (grammar_rule_check_and_complete): Don't check for
value usage for generated rules, it is ok to have a valued start
symbol, in which case it is ok for the generated rule ("accept: start
$end {}") to not use $1.
(packgram): Call grammar_rule_check_and_complete for all the rules.
Currently the core of the initial state is limited to the single rule
on $accept.
* src/lr0.c (generate_states): There may now be several rules on
$accept.
* src/graphviz.c (conclude_red): Recognize "final" transitions by the
fact that we reduce to "$accept".
* src/print.c (print_reduction): Likewise.
* src/print-xml.c (print_reduction): Likewise.
Now that the parser can read several start symbols, let's process
them, and create the corresponding rules.
* src/parse-gram.y (grammar_declaration): Accept a list of start symbols.
* src/reader.h, src/reader.c (grammar_start_symbol_set): Rename as...
(grammar_start_symbols_set): this.
* src/reader.h, src/reader.c (start_flag): Replace with...
(start_symbols): this.
* src/reader.c (grammar_start_symbols_set): Build a list of start
symbols.
(switching_token, create_start_rules): New.
(check_and_convert_grammar): Use them to turn the list of start
symbols into a set of rules.
* src/reduce.c (nonterminals_reduce): Don't complain about $accept,
it's an internal detail.
(reduce_grammar): Complain about all the start symbols that don't
derive sentences.
* src/symtab.c (startsymbol, startsymbol_loc): Remove, replaced by
start_symbols.
symbols_pack): Move the check about the start symbols
to...
* src/symlist.c (check_start_symbols): here.
Adjust to multiple start symbols.
* tests/reduce.at (Empty Language): Generalize into...
(Bad start symbols): this.
This is consistent with --defines being deprecated in favor of
--header. The directive %defines is also too similar to %define.
And %header matches nicely with api.header.name.
* src/scan-gram.l (%defines): Deprecate to %header.
(%header): Scan it.
* src/parse-gram.y (PERCENT_DEFINES): Replace with...
(PERCENT_HEADER): this.
* data/skeletons/lalr1.java
* doc/bison.texi
* tests/actions.at, tests/c++.at, tests/calc.at, tests/conflicts.at,
* tests/input.at, tests/java.at, tests/local.at, tests/output.at,
* tests/synclines.at, tests/types.at:
Convert most tests to check %header instead of %defines.
The name "defines" is incorrect, the generated file contains far more
than just #defines.
* src/getargs.h, src/getargs.c (-H, --header): New option.
With optional argument, just like --defines, --xml, etc.
(defines_flag): Rename as...
(header_flag): this.
Adjust dependencies.
* data/skeletons/bison.m4, data/skeletons/c.m4, data/skeletons/glr.c,
* data/skeletons/glr.cc, data/skeletons/glr2.cc, data/skeletons/lalr1.cc,
* data/skeletons/yacc.c:
Adjust.
* examples, doc/bison.texi: Adjust.
* tests/headers.at, tests/local.at, tests/output.at: Convert most
tests from using --defines to using --header.
On a case such as
%%
exp
: empty "a"
| "a" empty
empty
: %empty
we used to display
warning: shift/reduce conflict on token "a" [-Wcounterexamples]
Example: • "a"
Shift derivation
exp
↳ 2: • "a" empty
↳ 2: ε
Example: • "a"
Reduce derivation
exp
↳ 1: empty "a"
↳ 3: •
where the shift derivation shows an item "2: empty → ε", with an
explicit "ε", but the reduce derivation shows "3: empty → •", without
"ε".
For consistency, let's always show ε/%empty in rules with an empty
rhs:
Reduce derivation
exp
↳ 1: empty "a"
↳ 3: ε •
* src/derivation.c (derivation_width, derivation_print_tree_impl):
Always show ε/%empty in counterexamples.
* tests/diagnostics.at: Check that case.
* tests/conflicts.at, tests/counterexample.at: Adjust.
This is a fork of glr.cc to be c++-first instead of a wrapper around
glr.c.
* data/skeletons/glr2.cc: New.
* data/skeletons/bison.m4, data/skeletons/c++.m4: Adjust.
* data/skeletons/c.m4 (b4_user_args_no_comma): New.
* src/reader.c (grammar_rule_check_and_complete): glr2.cc is C++.
* tests/actions.at, tests/c++.at, tests/calc.at, tests/conflicts.at,
* tests/input.at, tests/local.at, tests/regression.at, tests/scanner.at,
* tests/synclines.at, tests/types.at: Also check glr2.cc.
An assertion failed when the last character is a '\' and we're in a
character or a string.
Reported by Agency for Defense Development.
https://lists.gnu.org/r/bug-bison/2020-08/msg00009.html
* src/scan-gram.l: Catch unterminated escapes.
* tests/input.at (Unexpected end of file): New.
Reported by Agency for Defense Development.
https://lists.gnu.org/r/bug-bison/2020-08/msg00008.html
On an empty such as
%token FOO
BAR
FOO 0
%%
input: %empty
we crash because when we find FOO 0, we decrement ntokens (since FOO
was discovered to be EOF, which is already known to be a token, so we
increment ntokens for it, and need to cancel this). This "works well"
when EOF is properly defined in one go, but here it is first defined
and later only assign token code 0. In the meanwhile BAR was given
the token number that we just decremented.
To fix this, assign symbol numbers after parsing, not during parsing,
so that we also saw all the explicit token codes. To maintain the
current numbers (I'd like to keep no difference in the output, not
just equivalence), we need to make sure the symbols are numbered in
the same order: that of appearance in the source file. So we need the
locations to be correct, which was almost the case, except for nterms
that appeared several times as LHS (i.e., several times as "foo:
..."). Fixing the use of location_of_lhs sufficed (it appears it was
intended for this use, but its implementation was unfinished: it was
always set to "false" only).
* src/symtab.c (symbol_location_as_lhs_set): Update location_of_lhs.
(symbol_code_set): Remove broken hack that decremented ntokens.
(symbol_class_set, dummy_symbol_get): Don't set number, ntokens and
nnterms.
(symbol_check_defined): Do it.
(symbols): Don't count nsyms here.
Actually, don't count nsyms at all: let it be done in...
* src/reader.c (check_and_convert_grammar): here. Define nsyms from
ntokens and nnterms after parsing.
* tests/input.at (EOF redeclared): New.
* examples/c/bistromathic/bistromathic.test: Adjust the traces: in
"%nterm <double> exp %% input: ...", exp used to be numbered before
input.
* cfg.mk (_space_before_paren_exempt): Be less laxist.
* src/output.c, src/reader.c: Fix space before paren issues.
Pacify the warnings where applicable.
Older versions of GCC (4.1.2 here) don't like repeated typedefs.
CC src/bison-parse-simulation.o
src/parse-simulation.c:61: error: redefinition of typedef 'parse_state'
src/parse-simulation.h:74: error: previous declaration of 'parse_state' was here
make: *** [Makefile:7876: src/bison-parse-simulation.o] Error 1
Reported by Nelson H. F. Beebe.
* src/parse-simulation.c (parse_state): Don't typedef,
parse-simulation.h did it already.
This reverts commit d0bec3175f (which
should have read "We have a clash...", not "With have a clash...").
Now that `max()` was renamed `max_int()`, we can use `max` again, as
elsewhere in the code.
* src/counterexample.c (visited_hasher): Alpha reconversion.
Reported by Maarten De Braekeleer.
https://lists.gnu.org/r/bison-patches/2020-07/msg00080.html
We don't want to use gnulib's min and max macros, since we use
function calls in min/max arguments.
* src/location.c (max_int, min_int): Move to...
* src/system.h: here.
* src/counterexample.c, src/derivation.c: Use max_int instead of max.
From
foo.y:1.7-11: error: %type redeclaration for bar
1 | %type <foo> bar bar
| ^~~~~
foo.y:1.7-11: note: previous declaration
1 | %type <foo> bar bar
| ^~~~~
to
foo.y:1.17-19: error: %type redeclaration for bar
1 | %type <foo> bar bar
| ^~~
foo.y:1.13-15: note: previous declaration
1 | %type <foo> bar bar
| ^~~
* src/symlist.h, src/symlist.c (symbol_list_type_set): There's no need
for the tag's location, use that of the symbol.
* src/parse-gram.y: Adjust.
* tests/input.at: Adjust.
We crash if the input contains a string containing a NUL byte.
Reported by Suhwan Song.
https://lists.gnu.org/r/bug-bison/2020-07/msg00051.html
* src/flex-scanner.h (STRING_FREE): Avoid accidental use of
last_string.
* src/scan-gram.l: Don't call STRING_FREE without calling
STRING_FINISH first.
* tests/input.at (Invalid inputs): Check that case.
With have a clash with the "max" function.
src/counterexample.c: In function 'visited_hasher':
src/counterexample.c:720:48: error: declaration of 'max' shadows a global declaration [-Werror=shadow]
src/counterexample.c:116:12: error: shadowed declaration is here [-Werror=shadow]
* src/counterexample.c (visited_hasher): Alpha conversion.
Currently the suggestion to rerun is a -Wother warning:
warning: 2 shift/reduce conflicts [-Wconflicts-sr]
warning: rerun with option '-Wcounterexamples' to generate conflict counterexamples [-Wother]
Instead, let's attach it as a subnote of the diagnostic (in the
current case, -Wconflicts-sr):
warning: 2 shift/reduce conflicts [-Wconflicts-sr]
note: rerun with option '-Wcounterexamples' to generate conflict counterexamples
* src/conflicts.c (conflicts_print): Do that.
Adjust the test suite.
From
input.y: warning: reduce/reduce conflict on token $end [-Wcounterexamples]
Example: A b .
First derivation
a
`-> A b .
Second derivation
a
`-> A b
`-> b .
to
input.y: warning: reduce/reduce conflict on token $end [-Wcounterexamples]
Example: A b .
First reduce derivation
a
`-> A b .
Second reduce derivation
a
`-> A b
`-> b .
* src/counterexample.c (print_counterexample): here.
Compute the width of the labels to properly align the values.
* tests/conflicts.at, tests/counterexample.at, tests/diagnostics.at,
* tests/report.at: Adjust.
Now that the derivation is no longer printed on one line, aligning the
example and the derivation is no longer useful. It can actually be
harmful, as it makes the overall structure less clear.
* src/derivation.h, src/derivation.c (derivation_print_leaves): Remove
the `prefix` argument.
* src/counterexample.c (print_counterexample): Put the example next to
its label.
* tests/conflicts.at, tests/counterexample.at, tests/diagnostics.at,
* tests/report.at: Adjust.
Now that we use complain, the "sections" are clearer.
* src/counterexample.c (print_counterexample): Use the empty line only
in reports.
* tests/counterexample.at, tests/diagnostics.at, tests/report.at: Adjust.
This is more consistent, and brings benefits: users know that these
diagnostics are attached to -Wcounterexamples, and they can also click
on the hyperlink if permitted by their terminal.
We go from
warning: 1 reduce/reduce conflict [-Wconflicts-rr]
Reduce/reduce conflict on token $end:
Example A b .
First derivation a -> [ A b . ]
Second derivation a -> [ A b -> [ b . ] ]
to
warning: 1 reduce/reduce conflict [-Wconflicts-rr]
input.y: warning: reduce/reduce conflict on token $end [-Wcounterexamples]
Example A b .
First derivation a -> [ A b . ]
Second derivation a -> [ A b -> [ b . ] ]
with an hyperlink on -Wcounterexamples.
* src/counterexample.c (counterexample_report_reduce_reduce):
Use complain.
* tests/counterexample.at, tests/diagnostics.at, tests/report.at:
Adjust.
