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
worker. It's very likely to be thread-unsafe and so it's better to avoid
using in MJIT worker to prevent surprises by race condition.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@65048 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
I'm simplifying the interface of ChildRecord as I'm going to complicate
it a little next.
I didn't drop hInput since leaving it would be more natural as its
interface.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@65044 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Prior to this commit, some of parent process's output was unintentionally
suppressed. We couldn't suppress only child process's output with spawnvp.
Instead of that, this commit uses CreateProcess directly to redirect stdout
and stderr only for child process.
As it's dealing with HANDLE returned from CreateProcess, now waitpid macro
needs to CloseHandle it.
win32/win32.c: Introduce rb_w32_start_process which is designed for MJIT
worker. Other similar functions can't be used since they are using ALLOCV
that may trigger GC, which should be avoided on MJIT worker.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@65033 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* win32/win32.c (constat_parse): split long buffer and limit write
size on a console, as well as rb_w32_write.
[ruby-dev:50597] [Bug #14942]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@64092 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* win32/win32.c (filetime_split, clock_gettime): keep the
precision as possible as the FILETIME format.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63636 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* win32/win32.c (constat_apply): CSI 'J' and 'K' are defaulted to
1, not 0. [ruby-core:86560] [Bug #14691]
* win32/win32.c (constat_apply): "delete before cursor" sequences
include the cursor position.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@63187 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* win32/win32.c (rb_w32_write): writing to closed pipe fails with
ERROR_NO_DATA but msvcrt maps it to EINVAL. map it to EPIPE.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@62150 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* win32/win32.c (w32_io_info, rb_w32_file_identical_p): fallback
to GetFileInformationByHandle if GetFileInformationByHandleEx
failed. it seems not working on network drives.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@61067 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Support nanosec file timestamp on Windows 8 or later.
Original patches are written by kubo (Kubo Takehiro).
Windows 7 and earlier also supports nanosec file timestamp, but it's too
accurate than system time. so, this feature is disabled on such versions.
[Feature #13726]
this change also includes [Misc #13702]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@61013 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* win32/win32.c (FILE_ID_128): it's not defined in SDK with VC10, but seems to
be defined in SDK with VC12.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@61007 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* win32/win32.c (vm_exit_handler): separate exit handler for
resources which must be released at exit of Ruby VM.
* win32/win32.c (socklist_insert, constat_handle): install the VM
exit handler.
* gc.c (ENABLE_VM_OBJSPACE): no longer needs process global object
space on Windows too.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60856 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* win32/win32.c (gettiemeofday, wutime): use GetSystemTimePreciseAsFileTime
instead of GetSystemTimeAsFileTime if it is available.
This patch is based on Takehiro Kubo <kubo@jiubao.org> 's one (change only
the name of wrapper function). Thanks! and sorry to late
[ruby-dev:50167] [Feature #13732]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@60240 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* win32/win32.c (rb_w32_set_thread_description): fix the condition
if the API function pointer is found.
[ruby-core:82494] [Bug #13845]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@59678 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* win32/win32.c (rb_w32_set_thread_description_str): return the
result when name is nil.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@59661 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* win32/win32.c (rb_dir_getwd_ospath): Windows implementation
moved from dir.c. get rid of freeing malloced memory by xfree.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@59061 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* win32/win32.c (skipspace, w32_cmdvector): get rid of iswspace(),
which is locale dependent.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@59019 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* win32/win32.c (winnt_stat): the return value of `get_final_path` is the
expected buffer length, not the actuall filled length.
* win32/win32.c (winnt_stat): `finalname` may be accessed in the outer block of
its definition via `path`.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@57135 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* win32/win32.c (rb_w32_home_dir): move from win32/file.c to try
special folders.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@56901 b2dd03c8-39d4-4d8f-98ff-823fe69b080e