All normal Ruby IO methods (IO#read, IO#gets, IO#write, ...) are
all capable of appearing to be "blocking" when presented with a
file description with the O_NONBLOCK flag set; so there is
little risk of incompatibility within Ruby-using programs.
The biggest compatibility risk is when spawning external
programs. As a result, stdin, stdout, and stderr are now always
made blocking before exec-family calls.
This change will make an event-oriented MJIT usable if it is
waiting on pipes on POSIX_like platforms.
It is ALSO necessary to take advantage of (proposed lightweight
concurrency (aka "auto-Fiber") or any similar proposal for
network concurrency: https://bugs.ruby-lang.org/issues/13618
Named-pipe (FIFO) are NOT yet non-blocking by default since
they are rarely-used and may introduce compatibility problems
and extra syscall overhead for a common path.
Please revert this commit if there are problems and if I am afk
since I am afk a lot, lately.
[ruby-core:89950] [Bug #14968]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@65922 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Running with TESTS=--no-retry, I sometimes get short reads
leading to failures due to the use of IO#readpartial instead of
IO#gets.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64248 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Apparently my tty can't keep up when I run test-all with
TESTS='-v --no-retry'
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64247 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
I'm not entirely sure why, but SIGTTOU pauses the test
when running test-all in parallel.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63583 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This should make portability problems more apparent when we try
using ppoll with rb_wait_for_single_fd on FreeBSD or other
non-Linux platforms.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63423 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* test/io/console/test_io_console.rb: flush to ensure the second
data is sent.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60049 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* test/io/console/test_io_console.rb (test_set_winsize_console):
split from test_get_winsize_console and enable the case when
attached to a console already
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@59842 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Not so good test, but it would be better than nothing, I guess...
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@59741 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* test/io/console/test_io_console.rb (test_winsize): on old CentOS5
window size seems unable to be set across a pty.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@57289 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/io/console/console.c (console_key_pressed_p): raise the same
exception, "unknown virtual key code", for names with nul chars.
though console_win32_vk() considers the length and can deal with
nul chars, rb_sprintf() raised at PRIsVALUE previously, so quote
it if it is unprintable.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@54589 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* test/io/console/test_io_console.rb: move conditions for method
definitions before the bodies.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@54588 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
select(2) in AIX returns "readable" for the write-side fd
of a pipe, so it is not possible to use a pipe to test
the read-write timeout of IO#wait on AIX.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@54047 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/io/wait/wait.c (io_wait_readwrite): [EXPERIMENTAL] allow to
wait for multiple modes, readable and writable, at once. the
arguments may change in the future. [Feature #12013]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@53642 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* test/io/wait/test_io_wait.rb: add tests for IO#wait_readable,
same as IO#wait.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@53640 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
When you change this to true, you may need to add more tests.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@53141 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* test/io/console/test_io_console.rb (test_getpass): skip for the
time being.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@53001 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* test/io/console/test_io_console.rb (test_getpass): wait for the
child process to read the input, not to read it by the master
itself.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@53000 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* test/io/console/test_io_console.rb (test_getpass): ignore echo
back of input from the master side. some systems seem echo back
but other may not.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52998 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
for a process. Fix Errno::ECHILD in TestIO_Console#test_close and
TestIO_Console#test_sync.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52957 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Reject IO::Console#getpass because it can't test portablly.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52945 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* test/io/console/test_io_console.rb (test_getpass): separate
master side and slave side in each threads. r52937 deadlocked
on OS X. [Bug #11780] [ruby-dev:49412]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52939 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
should be tested. Narrow ensure block. This reverts r52911.
[Bug #11780] [ruby-dev:49412]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52937 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
I wrote some code which relies on this nowadays, but Ruby <=2.2
and earlier behaved differently...
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@52231 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* time: Object#timeout has been deprecated a long time ago, use
Timeout.timeout.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51225 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/io/wait/wait.c (io_nread): wrap return value with INT2FIX
Thanks to Yura Sokolov <funny.falcon@gmail.com>
[ruby-core:68369] [Bug#10923]
* test/io/wait/test_io_wait.rb (test_nread_buffered):
fix broken test
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@49797 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* test/io/console/test_io_console.rb (test_noctty): use different
names for each tempfiles.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@49039 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/io/console/console.c (console_dev): send the given arguments
to the opened console. as a special case, do nothing if :close
is given.
* test/lib/leakchecker.rb (LeakChecker#check_fd_leak): close if
console.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48993 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* test/io/console/test_io_console.rb (run_pty): pty cannot read
all, should use gets only.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48989 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* test/io/console/test_io_console.rb (run_pty): IO#readlines
ignores the given block. need map to chomp the results.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48987 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/io/console/console.c (console_dev): id_console is not a
constant name, use rb_const_remove() to get rid of NameError.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@48982 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* test/io/console/test_io_console.rb (test_noctty): use
`assert_ruby_status` to show failure details.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@45903 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* test/io/console/test_io_console.rb (test_stringio_getch): use more
descriptive assertions.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@39030 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/io/console/console.c (rawmode_opt): min is minimum characters,
not tenths.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@39029 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* test/io/wait/test_io_wait.rb (TestIOWait#fill_pipe):
Errno::EWOULDBLOCK may not be the same as Errno::EAGAIN. patch by
phasis68 (Heesob Park) at [ruby-core:49894]. [Bug #7420]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@37984 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/io/wait/wait.c (io_wait_writable): this is easier to use than
IO.select for a single IO object and is immune to the
limitations/innefficiency of select() on platforms where poll/ppoll
is available. patched by Eric Wong. [Feature #4646]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@37785 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
for StringIO. https://github.com/nobu/io-console/issues/4
* ext/stringio/stringio.c: moved some methods to hidden modules.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@34376 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
OpenBSD seems not to detach the controlling terminal, when the
argument noclose is non-zero. ref: [Bug #5135]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@32848 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
test_wait}: give system some time to process the written data.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@31857 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
AF_UNIX, but windows socketpair doesn't support it. we can't
avoid platform check. sigh!
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@31472 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
PTY.open is not guaranteed to work. On my machine opening a
pty is prohibited via process control group. On those cases
exceptions shall occur, and that doesn't mean our fault.
Skip those tests on such situations.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@29281 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Kazuhiro NISHIYAMA
Nobuyoshi Nakada
k0kubun
nobu
normal
mame
odaira
naruse
ngoto
akr
usa
hsbt
shyouhei
kosaki
kazu