argument if optinal arguments are defined not only mandatory or
post arguments. [ruby-core:41557] [Bug #5730]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@34020 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
raise_method_missing().
* vm_eval.c (send_internal): remove inadvertent symbol creation
from public_send. based on a patch by Jeremy Evans <code AT
jeremyevans.net> in [ruby-core:38576]. [Feature #5112]
* vm_insnhelper.c (vm_call_method): remove inadvertent symbol
creation from send and __send__, too.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@33419 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
ancestors only when global scope. [ruby-core:39227] [Bug #5264]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@33163 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
also inherited constants for classes without superclass and
modules. [ruby-core:37698] [Bug #3423]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@33157 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
protect a VALUE from GC. It's not for general anti-optimizing
purpose.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@33126 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
What's the problem?
autoload is thread unsafe. When we define a constant to be
autoloaded, we expect the constant construction is invariant. But
current autoload implementation allows other threads to access the
constant while the first thread is loading a file.
What's happening inside?
The current implementation uses Qundef as a marker of autoload in
Constant table. Once the first thread find Qundef as a value at
constant lookup, it starts loading a defined feature. Generally a
loaded file overrides the Qundef in Constant table by module/class
declaration at very beginning lines of the file, so other threads
can see the new Module/Class object before feature loading is
finished. It breaks invariant construction.
How to solve?
To ensure invariant constant construction, we need to override
Qundef with defined Object after the feature loading. For keeping
Qundef in Constant table, I expanded autoload_data struct in
Module to have a slot for keeping the defined object while feature
loading. And changed Module's constant lookup/update logic a
little so that the slot is only visible from the thread which
invokes feature loading. (== the first thread which accessed the
autoload constant)
Evaluation?
All test passes (bootstrap test, test-all and RubySpec) and added
8 tests for threading behavior. Extra logics are executed only
when Qundef is found, so no perf drop should happen except
autoloading.
* variable.c (rb_autoload): Prepare new autoload_data struct.
* variable.c (rb_autoload_load): Load feature and update Constant
table after feature loading is finished.
* variable.c (rb_const_get_0): When the fetched constant is under
autoloading, it returns the object only for the thread which starts
autoloading.
* variable.c (rb_const_defined_0): Ditto.
* variable.c (rb_const_set): When the specified constant is under
autoloading, it sets the object only for the thread which starts
autoloading. Otherwise, simply overrides Qundef with constant
override warning.
* vm_insnhelper.c (vm_get_ev_const): Apply same change as
rb_const_get_0 in variable.c.
* test/ruby/test_autoload.rb: Added tests for threading behavior.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@33078 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Manage a redefinition of special methods for each classes.
A patch from Joel Gouly <joel.gouly@gmail.com>. Thanks!
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@33059 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
to solve an issue reported at [ruby-dev:44413] [Ruby 1.9 - Bug #5217].
Before this fix, each instruction has an information including
line number (iseq::iseq_insn_info_table). Instead of this data
structure, recording only line number changing places
(iseq::iseq_line_info_table).
The order of entries in iseq_line_info_table is ascending order of
iseq_line_info_table_entry::position. You can get a line number
by an iseq and a program counter with this data structure.
This fix reduces memory consumption of iseq (bytecode).
On my measurement, a rails application consumes 21.8MB for
iseq with this fix on the 32bit CPU. Without this fix, it
consumes 24.7MB for iseq [ruby-dev:44415].
* proc.c: ditto.
* vm_insnhelper.c: ditto.
* vm_method.c: ditto.
* vm.c (rb_vm_get_sourceline): change to use rb_iseq_line_no().
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@33046 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
of cfp consistency error problem on OS X 10.7 (Lion). It's
suspected llvm optimization bug.
[Bug #5076] [ruby-dev:44185]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@32628 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Fixes Bug #4648.
The patch is contributed by Kazuki Tsujimoto.
* bootstraptest/test_proc.rb: add tests for above.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@32495 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
for methods defined by define_method().
* thread.c (call_trace_proc): Fix to skip if class is not given (0).
Note that ID and Class object are passed for call/return event
if the called method was defined by define_method().
If you are author of tracer/profiler/debugger, this may be an
important change. You should check passed class as zero or
non-zero instead of checking the event type.
* test/ruby/test_settracefunc.rb: add a test for above.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@32334 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
stack overrun. currently super() in Proc created in a method
defined by Module#define_method raise NoMethodError. [Bug #4881]
* test/ruby/test_method.rb t_super_in_proc_from_define_method):
add test for it.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@32227 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
On clang -O, it is needed to avoid the optimization.
With this and llvm/clang's recent fix, clang 3.0 can
build ruby-trunk with -O option.
* cont.c (cont_capture): use for-loop.
* array.c (rb_ary_each): add volatile and use it.
* vm_insnhelper.c (vm_call_cfunc): ditto.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@32201 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* compile.c, eval.c, iseq.c, object.c, parse.y, proc.c, process.c,
thread.c, vm.c, vm_eval.c, vm_insnhelper.c, vm_method.c: don't
declare internal functions.
Note that rb_method_entry_eq() is defined in vm_method.c but
there was a declaration in proc.c with different const-ness.
Now it is declared in method.h with same const-ness to the
definition.
* object.c (rb_mod_module_exec): don't declare functions declared in
include/ruby/intern.h.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@32163 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
RSTRING_PTR() for function calls since it evaluates the argument
a couple of times.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@30922 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
private constant has been prohibited incorrectly.
* test/ruby/test_module.rb (test_toplevel_private_constant): add a
test for above.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@30718 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
not raise an exception even when the constant is private. Instead,
rb_public_const_get_* and rb_public_const_defined_* are introduced,
which raise an exception when the referring constant is private.
see [ruby-core:32912].
* vm_insnhelper.c (vm_get_ev_const): use rb_public_const_get_* instead
of rb_const_get_* to follow the constant visibility when user code
refers a constant.
* test/ruby/test_marshal.rb (test_marshal_private_class): add a test.
This test had failed because of incompatibility of rb_const_get.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@30713 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
vm_insnhelper.c (caller_setup_args): reverted r30241 and r30243
except for the test.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@30283 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
marcandre
nobu
naruse
nagachika
mame
tenderlove
kazu
nahi
usa
ko1
kosaki
akr
shugo