and x86_64 ABI.
* template/sizes.c.tmpl: Relax a pattern for types.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50920 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/date/date_parse.c (s3e, date_zone_to_diff, parse_ddd_cb):
use ALLOCV instead of ALLOCA get rid of stack overflow.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50919 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/date/date_strptime.c (read_digits): use ALLOCV instead of
ALLOCA get rid of stack overflow.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50918 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This reduces GC overhead and makes the API more consistent
with IO#read and IO#read_nonblock.
* ext/socket/basicsocket.c (bsock_recv): document outbuf
* ext/socket/unixsocket.c (unix_recvfrom): ditto
* ext/socket/init.c (rsock_strbuf, recvfrom_locktmp): new functions
(rsock_s_recvfrom): support destination buffer as 3rd arg
(rsock_s_recvfrom_nonblock): ditto
* string.c (rb_str_locktmp_ensure): export for internal ext
* test/socket/test_nonblock.rb: test recv_nonblock
* test/socket/test_unix.rb: test recv
[ruby-core:69543] [Feature #11242]
Benchmark results:
user system total real
alloc 0.130000 0.280000 0.410000 ( 0.420656)
extbuf 0.100000 0.220000 0.320000 ( 0.318708)
-------------------8<--------------------
require 'socket'
require 'benchmark'
nr = 100000
msg = ' ' * 16384
size = msg.bytesize
buf = ' ' * size
UNIXSocket.pair(:DGRAM) do |a, b|
Benchmark.bmbm do |x|
x.report('alloc') do
nr.times do
b.send(msg, 0)
a.recv(size, 0)
end
end
x.report('extbuf') do
nr.times do
b.send(msg, 0)
a.recv(size, 0, buf)
end
end
end
end
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50912 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
As documented before, exceptions are expensive and IO::Wait*able are too
common in socket applications to be the exceptional case. Datagram
sockets deserve the same API which stream sockets are allowed with
read_nonblock and write_nonblock.
Note: this does not offer a performance advantage under optimal
conditions when both ends are equally matched in speed, but it it
does make debug output cleaner by avoiding exceptions whenever
the receiver slows down.
* ext/socket/ancdata.c (bsock_sendmsg_internal, bsock_recvmsg_internal):
support "exception: false" kwarg
* ext/socket/init.c (rsock_s_recvfrom_nonblock):
ditto
* ext/socket/init.c (rsock_s_recvfrom_nonblock): use rsock_opt_false_p
* ext/socket/socket.c (sock_connect_nonblock): ditto
* ext/socket/rubysocket.h (rsock_opt_false_p): new function
* ext/socket/basicsocket.c (bsock_recv_nonblock): update rdoc
* ext/socket/udpsocket.c (udp_recvfrom_nonblock): ditto
* test/socket/test_nonblock.rb: new tests
[ruby-core:69542] [Feature #11229]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50910 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/socket/raddrinfo.c (parse_numeric_port): used only when
inet_pton() is available.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50904 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/-test-/gvl/call_without_gvl/call_without_gvl.c: sys/select.h
is included by ruby/define.h if available.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50903 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
unless main thread is already exited. Otherwise main thread could be
wrongly interrupted when it uses rb_thread_call_without_gvl(). Patch
by Takehiro Kubo. [Bug #11237][ruby-dev:49044][GH-898]
* test/-ext-/gvl/test_last_thread.rb: new test for the above fix.
* ext/-test-/gvl/call_without_gvl/call_without_gvl.c: new ext for the
above test.
* ext/-test-/gvl/call_without_gvl/extconf.rb: ditto.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50900 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
port overflow.
(numeric_getaddrinfo): Use parse_numeric_port.
numeric_getaddrinfo fails if port is too big now.
This makes rb_getaddrinfo invokes the real getaddrinfo()
on such condition.
This change is related to [ruby-core:69355] [Bug #11179].
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50890 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
it was not mentioned at its commit log and it caused a build error on
Windows ($(MAKE) is already quoted).
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50867 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/io/console/depend (win32_vk.inc): use offsetof() to suppress
warnings against cast from pointer to integer of different size.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50859 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
We call memset to zero `mh' a few lines above anyways, making
explicitly zero-ing each field unnecessary.
Compilers optimize this code away anyways, but it still costs me
4 lines of my small terminal and even smaller attention span.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50786 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
parsed even on systems without HAVE_STRUCT_MSGHDR_MSG_CONTROL
to prevent SEGV caused by passing Qnil to RARRAY_LENINT and
to preserve behavior before r50776.
[Bug #11224] [ruby-core:69468] [Bug #11225] [ruby-core:69469]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50784 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/socket/ancdata.c (bsock_sendmsg_internal): avoid msg_control
ptr if msg_controllen is zero to fix portability problems.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50781 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Systems without HAVE_STRUCT_MSGHDR_MSG_CONTROL still need
controls to be nil for the check after argument parsing.
Note: not actually tested on a real system without msg_control
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50780 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This should make the code more maintainable and make it easier to
add keyword argument support for "exception: false" in the future.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50776 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/tk/tkutil/tkutil.c (cbsubst_append_inf_key): extract a
function append a key in subst info to a string. make result
strings first and get rid of potential memory leak.
* ext/tk/tkutil/tkutil.c (cbsubst_get_subst_arg): allocate the
result buffer as a string to fix:
* memory leak when the argument key is not found:
loop {Tk::Event.subst_arg(:a) rescue nil}
* buffer overflow segfault when many arguments:
class T < TkUtil::CallbackSubst
_setup_subst_table([[?a, ?A, :_a]], [])
subst_arg(*[:_a]*1000).size
end
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50704 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/tk/tkutil/tkutil.c (cbsubst_get_subst_key): reduce
unnecessary buffer. the result string is one byte per one word.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50696 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
See doc/extension.rdoc for explanation.
[ruby-core:69419] [Feature #11198]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50685 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/openssl: make wrapper objects before allocating structs to
get rid of potential memory leaks.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50673 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/psych/psych_emitter.c (allocate): allocate structs with
making new wrapper objects and get rid of potential memory leak.
* ext/psych/psych_parser.c (allocate): ditto.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50672 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/socket/ancdata.c (bsock_sendmsg_internal,
bsock_recvmsg_internal):
avoid redundant fcntl on Linux
[ruby-core:69154] [Feature #11145]
* ext/socket/init.c (rsock_s_recvfrom_nonblock): ditto
* ext/socket/rubysocket.h (MSG_DONTWAIT_RELIABLE): new macro
MSG_DONTWAIT is enough to force non-blocking I/O under Linux,
so avoid changing the state of a socket. This will allow certain
threads to do a non-destructive non-blocking "peek" while others
block (without relying on an extra ppoll syscall).
We shall be conservative about enabling this feature since some
OSes may have incomplete support for MSG_DONTWAIT. I shall
defer to a FreeBSD expert to enable that for FreeBSD.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50666 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ObjectSpace.internal_class_of: return RBASIC_CLASS(obj).
* ObjectSpace.internal_super_of: return RCLASS_SUPER(cls).
* NEWS: add information about both methods.
* test/objspace/test_objspace.rb: add tests for both methods.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50662 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/json/generator/generator.c (cState_s_allocate): allocate
structs with making new wrapper objects and get rid of potential
memory leak.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50661 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/json/parser/parser.rl (cJSON_parser_s_allocate): allocate
structs with making new wrapper objects and get rid of potential
memory leak.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50660 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/win32ole/win32ole.c (Init_win32ole): make wrapper object
before making st_table.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50659 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/tk/tcltklib.c (ip_create_slave_core): allocate structs with
making new wrapper objects and get rid of potential memory leak.
* ext/tk/tkutil/tkutil.c (allocate_cbsubst_info): ditto.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50658 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/tk/tkutil/tkutil.c (cbsubst_get_subst_key): check arguments
type. implicit conversion is disabled to get rid of method
calls which can modify other arguments.
* ext/tk/tkutil/tkutil.c (cbsubst_table_setup): ditto.
* ext/tk/tkutil/tkutil.c (cbsubst_scan_args): ditto.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50657 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/tk/tkutil/tkutil.c (cbsubst_get_subst_key): ensure the
result type of a method call to be an array before accessing by
RARRAY macros.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50656 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/bigdecimal/bigdecimal.c (VpNewRbClass): make wrapper object
before result structs allocation and manage refcount for each
elements to get rid of potential memory leak.
* ext/bigdecimal/bigdecimal.c (BigDecimal_global_new): ditto.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50650 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
From doc/extension.rdoc:
>
> Using the RB_GC_GUARD macro is preferable to using the "volatile"
> keyword in C. RB_GC_GUARD has the following advantages:
>
> 1) the intent of the macro use is clear
>
> 2) RB_GC_GUARD only affects its call site, "volatile" generates some
> extra code every time the variable is used, hurting optimization.
>
> 3) "volatile" implementations may be buggy/inconsistent in some
> compilers and architectures. RB_GC_GUARD is customizable for broken
> systems/compilers without those without negatively affecting other
> systems.
* ext/openssl/ossl_asn1.c (ossl_asn1_traverse, ossl_asn1_decode,
ossl_asn1_decode_all): use RB_GC_GUARD instead of volatile
[ruby-core:69371] [Bug #11185]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50646 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/io/nonblock/nonblock.c (io_nonblock_set): return whether
nonblock flag was changed.
* ext/io/nonblock/nonblock.c (rb_io_nonblock_block): nothing to
restore but just yield unless nonblock flag is changed.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@50531 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
