r60655 changed the usage of bison by defining a customized YYLTYPE,
which seemed to cause a random stall at rb_thread_terminate_all.
Kazuki Tsujimoto investigated the issue and pointed out the commit that
caused it. Thanks!
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60660 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
rb_code_range_t has two t_code_location_t, i.e., the first and last
locations.
This is used for YYLTYPE, tracked locations of bison, and will be also
used for representing the "range", the first and the last locations of
each NODE. Currently, each NODE keeps only the first location, though.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60655 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* parse.y (rb_strterm_mark): do not define in ripper to get rid of
duplicate symbol with static-linked-ext.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60654 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This magic number has two meanings depending upon the context:
* "required keyword argument (no name)" on NODE_LASGN (`def foo(x:)`)
* "rest argument (no name)" on NODE_MASGN and NODE_POSTARG
('a, b, * = ary` or `a, b, *, z = ary`)
To show this intention explicitly, two macros are introduced:
NODE_SPECIAL_REQUIRED_KEYWORD and NODE_SPECIAL_NO_NAME_REST.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60650 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
`new_args_tail_gen` returned imemo, but the value was later accessed as
`NODE*`. This prevented change of NODE structure.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60647 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* parse.y (rb_discard_node_gen): rb_discard_node() is not used in
ripper right now.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60645 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* parse.y (parser_set_line): removed no longer used function.
* parse.y (rb_strterm_heredoc_t): adjust type of sourceline to
ruby_sourceline.
* parse.y (rb_strterm_t): get rid of redefinition.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60644 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This is a follow-up of r60488. Not only the Ruby parser but also Ripper
now use NODE buffer instead of NODE objects managed by GC.
Now we can change the struct of NODEs so that it can keep detailed
location information, perhaps (not tried yet). Note that, however, the
first word of each NODE must be still compatible with RBasic structure.
This is because Ripper handles NODEs and other objects uniformly;
especially, it still uses `RB_TYPE_P(obj, T_NODE)` for distinguishing
between NODEs and other objects.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60639 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Just refactoring.
NODE_STRTERM and NODE_HEREDOC are not an internal node of AST, but a
temporary storage for managing termination of string literals and
heredocs. Instead of NODE abuse, I want to use imemo for the storage
in order to avoid (my) confusion.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60634 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
When NODE_WHEN is compiled by iseq_compile_each0,
the node passed to compile_when is NODE_WHEN (not NODE_CASE).
So we can not handle the location of NODE_CASE of
case statements which don't have case expressions.
e.g. :
```
case; when 1; foo; when 2; bar; else baz; end
```
This commit adds NODE_CASE2, and compiles it by
iseq_compile_each0.
* compile.c (compile_case): Does not call COMPILE_ when
NODE_CASE does not have case expressions.
* compile.c (compile_case2): Compile NODE_CASE2 by compile_case2.
* compile.c (compile_when): Delete an obsoleted function.
* compile.c (iseq_compile_each0): Compile NODE_CASE2.
* ext/objspace/objspace.c (count_nodes): Add NODE_CASE2 case.
* node.c (dump_node, rb_gc_mark_node): Add NODE_CASE2 case.
* node.h (node_type): Add NODE_CASE2.
* node.h (NEW_CASE2): Add a macro which generates NODE_CASE2.
* parse.y: Generate NODE_CASE2 if case expressions don't exist.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60585 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* node.c (A_NODE_HEADER): Print lineno.
* node.h (nd_column, nd_set_column): Store column
number with 16-bit.
* node.h (nd_lineno, nd_set_lineno): Define
getter/setter macros for lineno of RNode.
* parse.y : Set first lineno of tokens.
Thanks to takeshinoda for review.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60580 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* parse.y : Column of resq is set by new_resbody, so
this line is needless.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60576 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This re-introduces r60485.
This reverts commit 5a176b75b1.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60488 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
NODEs in AST are no longer objects managed by GC. This change will
remove the restriction imposed by the GC. For example, a NODE can use
more than five words (this is my primary purpose; we want to store the
position data for each NODE, for coverage library), or even a NODE can
have variable length (some kinds of NODEs have unused fields).
To do this, however, we need more work, since Ripper still uses T_NODE
objects managed by the GC.
The life time of NODEs is more obvious than other kinds of objects; they
are created at parsing, and they become disused immediately after
compilation. This change releases all NODEs by a few `xfree`s after
compilation, so performance will be improved a bit. In extreme example,
`eval("x=1;" * 10000000)` runs much faster (40 sec. -> 7.8 sec. on my
machine).
The most important part of this change is `ast_t` struct, which has
three contents: (1) NODE buffer (malloc'ed memory), (2) a reference to
the root NODE, and (3) an array that contains objects that must be
marked during parsing (such as literal objects). Some functions that
had received `NODE*` arguments, must now receive `ast_t*`.
* node.c, node.h: defines `ast_t` struct and related operations.
* gc.c, internal.h: defines `imemo_ast`.
* parse.y: makes `parser_params` struct have a reference to `ast_t`.
Instead of `rb_node_newnode`, use `rb_ast_newnode` to create a NODE.
* iseq.c, load.c, ruby.c, template/prelude.c.tmpl: modifies some
functions to handle `ast_t*` instead of `NODE*`.
* test/ruby/test_gc.rb: ad-hoc fix for a failed test. The test assumes
GC eden is increased at startup by NODE object creation. However,
this change now create no NODE object, so GC eden is not necessarily
increased.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60485 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This is just a preparation to manage AST NODEs out of GC.
Currently `add_mark_object` does nothing.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60481 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
imemo_alloc provides a memory buffer whose contents are marked by GC.
C code can access imemo_alloc buffer freely, so imemo_alloc must be
considered writebarrier-unprotected. But T_IMEMO is writebarrier-
protected by default, which caused a GC bug.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60427 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
The idiom `rb_gc_force_recycle((VALUE)...)` was used heavily.
This change defines `rb_discard_nod(NODE*)` and allows C compiler to
typecheck.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60391 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
The magic number was used to mark NODE_ATTRASGN when its receiver is
self. But the hack was refactored at r46366. So the remaining code
fragments are no longer needed.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60389 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* parse.y (new_op_assign_gen): new_op_assign_gen for
ripper does not need column information.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60289 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* parse.y (new_op_assign): Define the macro separately
for ruby parser and for ripper.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60286 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* internal.h (rb_imemo_alloc_struct), gc.c (gc_mark_imemo): turned
next into the pointer to chain.
* parse.y (NEWHEAP): needs a cast.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60244 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* parse.y (parse_numeric): Do not literal_flush if
float is read correctly. This will fix the first
column of float.
before:
```
1.2i1.1
^
```
after:
```
1.2i1.1
^~~
```
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60225 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* parse.y (new_cdecl_gen): Rename local variable name
from cdecl to nd_cdecl. We can not use cdecl on Windows.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60223 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* node.c (A_NODE_HEADER): Print column number.
* node.h (nd_column, nd_set_column): Define getter/setter
macros for column number of RNode.
* parse.y: Set first column number of tokens.
Notes:
* Use `@n` to get the location of token in parse.y.
* When we use `@n`, arguments of yyerror and yylex are
changed.
* Initialize column of nodes with -1 to make it easy
to detect nodes which we forget to set a column number.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60220 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* parse.y (parser_here_document): update indent at '#', which is
not a space. [ruby-core:83368] [Bug #14032]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60213 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* parse.y (rel_expr): warn sequence of comparisons, which would be
probably unintentional. [EXPERIMENTAL]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60207 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* parse.y (parser_prepare): set token_info_enabled flag first,
before returning at BOM. [ruby-dev:50288] [Bug #13998]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60160 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* parse.y (KWD2EID): should respect the previous callback result
for keywords. [ruby-core:83106] [Bug #13971]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60120 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
nobu
mame
yui-knk
ko1
svn