Prompted on Piotr Marcińczyk's message:
http://lists.gnu.org/archive/html/bug-bison/2017-06/msg00000.html.
See also http://lists.gnu.org/archive/html/bug-bison/2018-06/msg00001.html.
Because their type is unknown to Bison, the values of midrule actions are
not treated like the others: they don't have %printer and %destructor
support. In addition, in C++, (Bison) variants cannot work properly.
Typed midrule actions address these issues. Instead of:
exp: { $<ival>$ = 1; } { $<ival>$ = 2; } { $$ = $<ival>1 + $<ival>2; }
write:
exp: <ival>{ $$ = 1; } <ival>{ $$ = 2; } { $$ = $1 + $2; }
* src/scan-code.h, src/scan-code.l (code_props): Add a `type` field to
record the declared type of an action.
(code_props_rule_action_init): Add a type argument.
* src/parse-gram.y: Accept an optional type tag for actions.
* src/reader.h, src/reader.c (grammar_current_rule_action_append): Add
a type argument.
(grammar_midrule_action): When a mid-rule is typed, pass its type to
the defined dummy non terminal symbol.
grammar_current_rule_action_append was used in two different places:
for actual action (`{...}`), and for predicates (`%?{...}`). Let's
split this in two different functions.
* src/reader.h, src/reader.c (grammar_current_rule_predicate_append): New.
Extracted from...
(grammar_current_rule_action_append): here.
Remove arguments that don't apply.
Adjust dependencies.
Currently on the following grammar:
%type <foo> foo
%%
start: foo | bar | "baz"
foo: foo
bar: bar
bison reports:
warning: 2 nonterminals useless in grammar [-Wother]
warning: 4 rules useless in grammar [-Wother]
1.13-15: warning: nonterminal useless in grammar: foo [-Wother]
%type <foo> foo
^^^
3.14-16: warning: nonterminal useless in grammar: bar [-Wother]
start: foo | bar | "baz"
^^^
[...]
i.e., the location of the first occurrence of a symbol is taken as its
definition point. In the case of nonterminals, the first occurrence
as a left-hand side of a rule makes more sense:
warning: 2 nonterminals useless in grammar [-Wother]
warning: 4 rules useless in grammar [-Wother]
4.1-3: warning: nonterminal useless in grammar: foo [-Wother]
foo: foo
^^^
5.1-3: warning: nonterminal useless in grammar: bar [-Wother]
bar: bar
^^^
[...]
* src/symtab.h, src/symtab.c (symbol::location_of_lhs): New.
(symbol_location_as_lhs_set): New.
* src/parse-gram.y (current_lhs): Use it.
* tests/reduce.at: Update locations.
Bison supports a union tag, for obscure reasons. But it does a poor
job at it, especially since Bison 3.0.
Reported by Stephen Cameron and Tobias Frost.
It did not ensure that the name was not given several times. An easy
way to do this is to make the %union tag be handled as a %define
variable, as they cannot be defined several times.
Since Bison 3.0, the synclines were wrongly placed, resulting in
invalid code. Addressing this issue, because of the way the union tag
was stored (as a code muscle), would have been tedious. Unless we
rather define the %union tag as a %percent variable, whose synclines
are easier to manipulate.
So replace the b4_union_name muscle by the api.value.union.name
%define variable, document, and check.
* data/bison.m4: Make sure that api.value.union.name has a keyword value.
* data/c++.m4: Make sure that api.value.union.name is not defined.
* data/c.m4 (b4_union_name): No longer use it, use api.value.union.name.
* doc/bison.texi (%define Summary): Document it.
* src/parse-gram.y (union_name): No longer define b4_uion_name, but
api.value.union.name.
* tests/input.at (Redefined %union name): New.
* tests/synclines.at (%union name syncline): New.
* tests/types.at: Check named %unions.
Using %code for config.h is wrong, as some headers will already have
been included by Bison. In some cases, e.g., glibc's string.h, this
results in some declaration not being made for lack of definition of
_GNU_SOURCE, which is performed by config.h.
* src/parse-gram.y: here.
The purpose of this symbol was only to factor function calls. As a
result the actions were indeed simpler, but the grammar was somewhat
uselessly obfuscated. Get rid of this symbol, but introduce functions
to simplify dependencies.
There is no (intended) changes of behavior here.
* src/parse-gram.y (strip_braces, translate_code(
(translate_code_braceless): New.
(braceless): Remove, use "{...}" instead, and one of the previous
functions depending on the context.
(STRING, "%{...%}", EPILOGUE): Declare as <code>, instead of <chars>,
the difference between both is useless (well, I couldn't make sense of
it, even after having read the initial commit that introduced them).
(%union): Remove the now useless "chars" type.
Adjust the printers.
* src/scan-gram.l: Adjust.
* src/parse-gram.y (version_check, gram_error, char_name, lloc_default):
Move their prototypes from %{...%} to %code.
(YYLLOC_DEFAULT, YY_LOCATION_PRINT): Move from %{...%} to %code.
(current_lhs): Move its implementation to the epilogue.
Now that we use "braceless" (which is {...} blocks of code with
initial and final braces stripped) to denote "short" values (such as
api.namespaces), the added end-of-line is a nuisance. As a matter of
fact, this extra-safety was useless, as every expansion of "braceless"
(aka, "user code") is followed by an end of line.
* src/parse-gram.y, src/parse-gram.c (braceless): Instead of replacing
the final brace by \n, just delete the brace.
Provide a means to tell the difference between "keyword" values (e.g.,
%define api.pull both), "string" values (e.g., %define file.name
"foo"), and "code" values (e.g., %define api.namespace {calc}).
Suggested by Joel E. Denny.
http://lists.gnu.org/archive/html/bison-patches/2013-03/msg00016.html
* src/muscle-tab.h, src/muscle-tab.c (muscle_kind, muscle_kind_new)
(muscle_kind_string): New.
(muscle_percent_define_insert): Take the kind as new argument.
Insert it in the muscle table.
Adjust callers.
* src/getargs.c: Adjust callers.
* src/parse-gram.y: Ditto.
(content.opt): Remove, replaced by...
(value): this new non-terminal, whose semantics value is stored
in the new "value" union member.
Provide a printer.
Support values in braces in additions to keyword and string values.
fuse me
* src/parse-gram.y (PARAM_TYPE): Remove useless typedef guard.
There's a header guard.
Use 'yyo' with %printer.
Use a consistent style for %union one-liners.
* src/parse-gram.y (INT): Fuse the %type and %token declaration.
Move its %union right before its introduction.
(%union): Split in several %unions, right before their use.
In a precedence declaration, when tokens are declared with a litteral
character (e.g., 'a') or with a identifier (e.g., B), Bison behaved
differently: the litteral tokens would be numbered first, and then the
other ones, leading to the following grammar:
%right A B 'c' 'd'
being numbered as such: 'c' 'd' A B.
* src/parse-gram.y (symbol.prec): Set the symbol number when reading the
symbols.
* tests/conflicts.at (Token declaration order: literals vs. identifiers):
New.
Signed-off-by: Akim Demaille <akim@lrde.epita.fr>
Provide a means to explicitly denote empty right-hand sides of rules:
instead of
exp: { ... }
allow
exp: %empty { ... }
Make sure that %empty is properly used.
With help from Joel E. Denny and Gabriel Rassoul.
http://lists.gnu.org/archive/html/bison-patches/2013-01/msg00142.html
* src/reader.h, src/reader.c (grammar_current_rule_empty_set): New.
* src/parse-gram.y (%empty): New token.
Use it.
* src/scan-gram.l (%empty): Scan it.
* src/reader.c (grammar_rule_check): Check that %empty is properly used.
* tests/actions.at (Invalid uses of %empty, Valid uses of %empty): New.
"stype" is quite unclear, and it also collides with the former %define
variable that had the same name (replaced by api.value.type).
* src/parse-gram.y (stype): Rename as...
(union_members): this.
* data/bison.m4: Adjust.
(b4_user_stype): Rename as...
(b4_user_union_members): this.
* data/c++.m4, data/c.m4: Adjust.
* src/parse-gram.c: regen.
* src/location.h, src/location.c (location_print): For consistency
with other data structures and other location_* routines, pass the
location argument first.
* src/complain.c: Adjust.
(location_caret): Likewise.
* src/parse-gram.y: Adjust.
In a declaration %token A B, the token A is declared before B, but in %left
A B (or with %precedence or %nonassoc or %right), the token B was declared
before A (tokens were declared in reverse order).
* src/symlist.h, src/symlist.c (symbol_list_append): New.
* src/parse-gram.y: Use it instead of symbol_list_prepend.
* tests/input.at: Adjust expectations.
* origin/maint: (46 commits)
doc: minor style change
maint: use gendocs's new -I option
regen
yacc.c: do not define location support when not using locations
maint: be compilable with GCC 4.0
tests: address a warning from GCC 4.4
tests: don't use options that Clang does not support
tests: restore the tests on -Werror
regen
parse-gram: update the Bison interface
fix comment
maint: post-release administrivia
version 2.6.4
regen
2.6.4: botched 2.6.3
maint: post-release administrivia
version 2.6.3
gnulib: update
tests: check %no-lines
NEWS: warnings with clang
...
Conflicts:
NEWS
TODO
data/c.m4
data/java.m4
doc/Makefile.am
src/getargs.c
src/getargs.h
src/output.c
src/parse-gram.c
src/parse-gram.h
src/parse-gram.y
src/reader.h
The "shadows a global declaration" warning in GCC 4.0 was a bit
annoying. It does not like that a type name be used in a prototype of
a function (not the implementation, just the declaration):
In file included from src/LR0.c:38:
src/reader.h:56: warning: declaration of 'named_ref' shadows a
global declaration
src/named-ref.h:35: warning: shadowed declaration is here
It does not like either when a global variable name is used in a
prototype. Flex 2.5.37 generates this prototype:
void gram_set_debug (int debug_flag );
* src/getargs.h, src/getargs.c (debug_flag): Rename as...
(debug): this.
Adjust dependencies.
* src/reader.h: Don't use "named_ref" as a formal argument name.
* origin/maint:
maint: remove useless file
update files to ignore
remove useless include
use locale-indep. c_is* functions for parsing, not isspace, isprint etc
gnulib: update
--help: include a place to report translation issues
tests: style changes
tests: fix push-pull test
yacc.c: style changes
Conflicts:
src/system.h
* src/parse-gram.y: Include "c-ctype.h".
(add_param): Parse with c_isspace, not isspace.
* src/parse-gram.c: Likewise.
* src/scan-gram.l: Include c-ctype.h, not ctype.h.
(SC_ESCAPED_STRING,SC_ESCAPED_CHARACTER): Use c_isspace and c_isprint,
not ctype.h's locale-dependent functions.
* origin/maint:
maint: post-release administrivia
version 2.6.2
NEWS: update.
yacc: remove trailing end of line at end of file
thanks: fix a contributor name
gnulib: update
tests: synch line -> syncline, for consistency
tests: synclines: style changes
tests: synclines: fix perl invocation
regen
c++: trailing end-of-lines in %parse-param
tests: simplify
Conflicts:
cfg.mk
data/glr.cc
data/yacc.c
src/parse-gram.c
src/parse-gram.h
src/parse-gram.y
* src/parse-gram.y (add_param): No only skip ' ' and '\t', skip all
leading and trailing spaces.
* tests/regression.at (Lex and parse params): Check it.
* NEWS: Document it.
* origin/maint: (29 commits)
regen
synclines: remove spurious empty line
also support $<foo>$ in the %initial-action
skeletons: b4_dollar_pushdef and popdef to simpify complex definitions
regen
printer/destructor: translate only once
factor the handling of m4 escaping
news: schedule the removal of the ";" hack
style changes in the scanners
regen
support $<tag>$ in printers and destructors
scan-code: factor the handling of the type in $<TYPE>$
muscles: fix another occurrence of unescaped type name
glr.cc: fix the handling of yydebug
gnulib: update
formatting changes
tests: fix an assertion
tests: adjust to GCC 4.8, which displays caret errors
be sure to properly escape type names
obstack_quote: escape and quote for M4
muscles: shuffle responsabilities
muscles: make private functions static
muscles: rename private functions/macros
obstack_escape: escape M4 characters
remove dead macro
maint: style changes
doc: avoid problems with case insensitive file systems
configure: fix botched quoting
news: fix typo.
Conflicts:
NEWS
data/c.m4
data/glr.cc
data/lalr1.cc
examples/rpcalc/local.mk
src/muscle-tab.h
src/output.c
src/parse-gram.c
src/parse-gram.h
src/parse-gram.y
src/scan-code.l
src/symlist.c
src/symlist.h
src/symtab.h
tests/calc.at
Currently "%printer {...} a b c d e f" translates the {...} six times.
Not only is this bad for time and space, it also issues six times the
same warnings.
* src/symlist.h, src/symlist.c (symbol_list_destructor_set)
(symbol_list_printer_set): Take the action as code_props instead of
const char *.
* src/parse-gram.y: Translate these actions here.
* src/scan-code.h: Comment change.
* tests/input.at: Check that warnings are issued only once.
Currently they are treated in separated variables, contrary to other
<TYPE> code_props. This duplicates code (and messages for translators)
uselessly, as demonstrated by the fact that thanks to this patch, now
useless <*> and <> code_props are reported like the others.
* src/parse-gram.y (generic_symlist_item): Treat "<*>" and "<>" as regular
type tags.
* src/symlist.h, src/symlist.c (symbol_list_default_tagged_new)
(symbol_list_default_tagless_new,SYMLIST_DEFAULT_TAGGED)
(SYMLIST_DEFAULT_TAGLESS): Remove.
* src/symtab.h, src/symtab.c (default_tagged_code_props)
(default_tagless_code_props, default_tagged_code_props_set)
(default_tagless_code_props_set): Remove.
(symbol_code_props_get): Default to <*> or <>'s code_props.
* tests/actions.at: Complete expected errors: there are new warnings.
* tests/input.at: Likewise.
(Useless printers or destructors): Extend.
