noderb 0.0.4 → 0.0.6
Sign up to get free protection for your applications and to get access to all the features.
- data/ext/noderb_extension/libuv/AUTHORS +11 -1
- data/ext/noderb_extension/libuv/LICENSE +0 -8
- data/ext/noderb_extension/libuv/Makefile +5 -5
- data/ext/noderb_extension/libuv/README +11 -6
- data/ext/noderb_extension/libuv/common.gypi +158 -7
- data/ext/noderb_extension/libuv/config-unix.mk +34 -26
- data/ext/noderb_extension/libuv/gyp_uv +22 -4
- data/ext/noderb_extension/libuv/include/{eio.h → uv-private/eio.h} +2 -2
- data/ext/noderb_extension/libuv/include/{uv-unix.h → uv-private/uv-unix.h} +40 -0
- data/ext/noderb_extension/libuv/include/{uv-win.h → uv-private/uv-win.h} +96 -14
- data/ext/noderb_extension/libuv/include/uv.h +508 -83
- data/ext/noderb_extension/libuv/src/unix/cares.c +185 -0
- data/ext/noderb_extension/libuv/src/unix/core.c +828 -0
- data/ext/noderb_extension/libuv/src/{eio → unix/eio}/config_darwin.h +4 -0
- data/ext/noderb_extension/libuv/src/{eio → unix/eio}/config_linux.h +5 -15
- data/ext/noderb_extension/libuv/src/{eio → unix/eio}/eio.c +23 -2
- data/ext/noderb_extension/libuv/src/unix/error.c +102 -0
- data/ext/noderb_extension/libuv/src/{ev → unix/ev}/config_linux.h +26 -15
- data/ext/noderb_extension/libuv/src/unix/fs.c +562 -0
- data/ext/noderb_extension/libuv/src/unix/internal.h +75 -0
- data/ext/noderb_extension/libuv/src/unix/pipe.c +282 -0
- data/ext/noderb_extension/libuv/src/unix/process.c +287 -0
- data/ext/noderb_extension/libuv/src/unix/stream.c +727 -0
- data/ext/noderb_extension/libuv/src/unix/tcp.c +226 -0
- data/ext/noderb_extension/libuv/src/unix/udp.c +524 -0
- data/ext/noderb_extension/libuv/src/{uv-eio.c → unix/uv-eio.c} +48 -27
- data/ext/noderb_extension/libuv/src/{uv-eio.h → unix/uv-eio.h} +1 -1
- data/ext/noderb_extension/libuv/src/uv-common.c +24 -14
- data/ext/noderb_extension/libuv/src/uv-common.h +12 -7
- data/ext/noderb_extension/libuv/src/win/async.c +14 -16
- data/ext/noderb_extension/libuv/src/win/cares.c +64 -79
- data/ext/noderb_extension/libuv/src/win/core.c +105 -53
- data/ext/noderb_extension/libuv/src/win/error.c +23 -13
- data/ext/noderb_extension/libuv/src/win/fs.c +807 -0
- data/ext/noderb_extension/libuv/src/win/getaddrinfo.c +61 -41
- data/ext/noderb_extension/libuv/src/win/handle.c +56 -30
- data/ext/noderb_extension/libuv/src/win/internal.h +134 -95
- data/ext/noderb_extension/libuv/src/win/loop-watcher.c +21 -18
- data/ext/noderb_extension/libuv/src/win/pipe.c +313 -158
- data/ext/noderb_extension/libuv/src/win/process.c +117 -131
- data/ext/noderb_extension/libuv/src/win/req.c +55 -31
- data/ext/noderb_extension/libuv/src/win/stdio.c +5 -5
- data/ext/noderb_extension/libuv/src/win/stream.c +19 -14
- data/ext/noderb_extension/libuv/src/win/tcp.c +278 -336
- data/ext/noderb_extension/libuv/src/win/threadpool.c +73 -0
- data/ext/noderb_extension/libuv/src/win/timer.c +44 -37
- data/ext/noderb_extension/libuv/src/win/udp.c +592 -0
- data/ext/noderb_extension/libuv/src/win/util.c +20 -6
- data/ext/noderb_extension/libuv/src/win/winapi.c +23 -6
- data/ext/noderb_extension/libuv/src/win/winapi.h +4206 -0
- data/ext/noderb_extension/libuv/src/win/winsock.c +270 -0
- data/ext/noderb_extension/libuv/src/win/winsock.h +134 -0
- data/ext/noderb_extension/libuv/test/benchmark-ares.c +10 -6
- data/ext/noderb_extension/libuv/test/benchmark-getaddrinfo.c +10 -7
- data/ext/noderb_extension/libuv/test/benchmark-list.h +21 -0
- data/ext/noderb_extension/libuv/test/benchmark-ping-pongs.c +11 -7
- data/ext/noderb_extension/libuv/test/benchmark-pound.c +147 -58
- data/ext/noderb_extension/libuv/test/benchmark-pump.c +29 -23
- data/ext/noderb_extension/libuv/test/benchmark-spawn.c +13 -10
- data/ext/noderb_extension/libuv/test/benchmark-udp-packet-storm.c +250 -0
- data/ext/noderb_extension/libuv/test/dns-server.c +11 -6
- data/ext/noderb_extension/libuv/test/echo-server.c +30 -22
- data/ext/noderb_extension/libuv/test/test-async.c +3 -3
- data/ext/noderb_extension/libuv/test/test-callback-stack.c +6 -6
- data/ext/noderb_extension/libuv/test/test-connection-fail.c +6 -5
- data/ext/noderb_extension/libuv/test/test-delayed-accept.c +13 -13
- data/ext/noderb_extension/libuv/test/test-fs.c +715 -0
- data/ext/noderb_extension/libuv/test/test-getaddrinfo.c +11 -8
- data/ext/noderb_extension/libuv/test/test-gethostbyname.c +8 -9
- data/ext/noderb_extension/libuv/test/test-getsockname.c +142 -16
- data/ext/noderb_extension/libuv/test/test-idle.c +4 -3
- data/ext/noderb_extension/libuv/test/test-list.h +29 -2
- data/ext/noderb_extension/libuv/test/test-loop-handles.c +9 -8
- data/ext/noderb_extension/libuv/test/test-ping-pong.c +9 -9
- data/ext/noderb_extension/libuv/test/test-pipe-bind-error.c +18 -14
- data/ext/noderb_extension/libuv/test/test-ref.c +17 -16
- data/ext/noderb_extension/libuv/test/test-shutdown-eof.c +5 -5
- data/ext/noderb_extension/libuv/test/test-spawn.c +17 -17
- data/ext/noderb_extension/libuv/test/test-tcp-bind-error.c +24 -18
- data/ext/noderb_extension/libuv/test/test-tcp-bind6-error.c +19 -14
- data/ext/noderb_extension/libuv/test/test-tcp-writealot.c +6 -5
- data/ext/noderb_extension/libuv/test/test-threadpool.c +59 -0
- data/ext/noderb_extension/libuv/test/test-timer-again.c +15 -12
- data/ext/noderb_extension/libuv/test/test-timer.c +8 -8
- data/ext/noderb_extension/libuv/test/test-udp-dgram-too-big.c +88 -0
- data/ext/noderb_extension/libuv/test/test-udp-ipv6.c +158 -0
- data/ext/noderb_extension/libuv/test/test-udp-send-and-recv.c +210 -0
- data/ext/noderb_extension/libuv/{all.gyp → uv.gyp} +75 -77
- data/ext/noderb_extension/libuv/vcbuild.bat +93 -0
- data/ext/noderb_extension/noderb.c +13 -0
- data/ext/noderb_extension/noderb_common.h +2 -1
- data/ext/noderb_extension/noderb_defer.c +54 -0
- data/ext/noderb_extension/noderb_defer.h +15 -0
- data/ext/noderb_extension/noderb_dns.c +1 -1
- data/ext/noderb_extension/noderb_fs.c +277 -0
- data/ext/noderb_extension/noderb_fs.h +27 -0
- data/ext/noderb_extension/noderb_process.c +5 -5
- data/ext/noderb_extension/noderb_tcp.c +10 -6
- data/ext/noderb_extension/noderb_timers.c +1 -1
- data/ext/noderb_extension/noderb_tools.c +42 -8
- data/ext/noderb_extension/noderb_tools.h +3 -1
- data/lib/noderb/defer.rb +25 -0
- data/lib/noderb/file.rb +119 -0
- data/lib/noderb/fs.rb +72 -0
- data/lib/noderb/version.rb +1 -1
- data/lib/noderb.rb +8 -0
- metadata +102 -266
- data/ext/noderb_extension/libuv/BSDmakefile +0 -2
- data/ext/noderb_extension/libuv/create-msvs-files.bat +0 -21
- data/ext/noderb_extension/libuv/deps/pthread-win32/ANNOUNCE +0 -482
- data/ext/noderb_extension/libuv/deps/pthread-win32/BUGS +0 -141
- data/ext/noderb_extension/libuv/deps/pthread-win32/Bmakefile +0 -268
- data/ext/noderb_extension/libuv/deps/pthread-win32/CONTRIBUTORS +0 -140
- data/ext/noderb_extension/libuv/deps/pthread-win32/COPYING +0 -150
- data/ext/noderb_extension/libuv/deps/pthread-win32/COPYING.LIB +0 -504
- data/ext/noderb_extension/libuv/deps/pthread-win32/ChangeLog +0 -5194
- data/ext/noderb_extension/libuv/deps/pthread-win32/FAQ +0 -451
- data/ext/noderb_extension/libuv/deps/pthread-win32/GNUmakefile +0 -593
- data/ext/noderb_extension/libuv/deps/pthread-win32/MAINTAINERS +0 -4
- data/ext/noderb_extension/libuv/deps/pthread-win32/Makefile +0 -516
- data/ext/noderb_extension/libuv/deps/pthread-win32/NEWS +0 -1245
- data/ext/noderb_extension/libuv/deps/pthread-win32/Nmakefile +0 -24
- data/ext/noderb_extension/libuv/deps/pthread-win32/Nmakefile.tests +0 -260
- data/ext/noderb_extension/libuv/deps/pthread-win32/PROGRESS +0 -4
- data/ext/noderb_extension/libuv/deps/pthread-win32/README +0 -601
- data/ext/noderb_extension/libuv/deps/pthread-win32/README.Borland +0 -57
- data/ext/noderb_extension/libuv/deps/pthread-win32/README.CV +0 -3036
- data/ext/noderb_extension/libuv/deps/pthread-win32/README.NONPORTABLE +0 -783
- data/ext/noderb_extension/libuv/deps/pthread-win32/README.Watcom +0 -62
- data/ext/noderb_extension/libuv/deps/pthread-win32/README.WinCE +0 -6
- data/ext/noderb_extension/libuv/deps/pthread-win32/TODO +0 -7
- data/ext/noderb_extension/libuv/deps/pthread-win32/WinCE-PORT +0 -222
- data/ext/noderb_extension/libuv/deps/pthread-win32/attr.c +0 -53
- data/ext/noderb_extension/libuv/deps/pthread-win32/autostatic.c +0 -69
- data/ext/noderb_extension/libuv/deps/pthread-win32/barrier.c +0 -47
- data/ext/noderb_extension/libuv/deps/pthread-win32/build/all.gyp +0 -207
- data/ext/noderb_extension/libuv/deps/pthread-win32/builddmc.bat +0 -9
- data/ext/noderb_extension/libuv/deps/pthread-win32/cancel.c +0 -44
- data/ext/noderb_extension/libuv/deps/pthread-win32/cleanup.c +0 -148
- data/ext/noderb_extension/libuv/deps/pthread-win32/condvar.c +0 -50
- data/ext/noderb_extension/libuv/deps/pthread-win32/config.h +0 -153
- data/ext/noderb_extension/libuv/deps/pthread-win32/context.h +0 -74
- data/ext/noderb_extension/libuv/deps/pthread-win32/create.c +0 -308
- data/ext/noderb_extension/libuv/deps/pthread-win32/dll.c +0 -92
- data/ext/noderb_extension/libuv/deps/pthread-win32/errno.c +0 -94
- data/ext/noderb_extension/libuv/deps/pthread-win32/exit.c +0 -44
- data/ext/noderb_extension/libuv/deps/pthread-win32/fork.c +0 -39
- data/ext/noderb_extension/libuv/deps/pthread-win32/global.c +0 -107
- data/ext/noderb_extension/libuv/deps/pthread-win32/implement.h +0 -944
- data/ext/noderb_extension/libuv/deps/pthread-win32/misc.c +0 -50
- data/ext/noderb_extension/libuv/deps/pthread-win32/mutex.c +0 -62
- data/ext/noderb_extension/libuv/deps/pthread-win32/need_errno.h +0 -145
- data/ext/noderb_extension/libuv/deps/pthread-win32/nonportable.c +0 -47
- data/ext/noderb_extension/libuv/deps/pthread-win32/private.c +0 -54
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread.c +0 -66
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread.dsp +0 -142
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread.dsw +0 -29
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread.h +0 -1368
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_attr_destroy.c +0 -79
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_attr_getdetachstate.c +0 -86
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_attr_getinheritsched.c +0 -51
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_attr_getschedparam.c +0 -52
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_attr_getschedpolicy.c +0 -61
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_attr_getscope.c +0 -54
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_attr_getstackaddr.c +0 -97
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_attr_getstacksize.c +0 -100
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_attr_init.c +0 -117
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_attr_setdetachstate.c +0 -91
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_attr_setinheritsched.c +0 -57
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_attr_setschedparam.c +0 -63
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_attr_setschedpolicy.c +0 -55
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_attr_setscope.c +0 -62
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_attr_setstackaddr.c +0 -97
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_attr_setstacksize.c +0 -110
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_barrier_destroy.c +0 -103
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_barrier_init.c +0 -69
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_barrier_wait.c +0 -104
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_barrierattr_destroy.c +0 -83
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_barrierattr_getpshared.c +0 -95
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_barrierattr_init.c +0 -85
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_barrierattr_setpshared.c +0 -119
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_cancel.c +0 -189
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_cond_destroy.c +0 -253
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_cond_init.c +0 -167
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_cond_signal.c +0 -231
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_cond_wait.c +0 -567
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_condattr_destroy.c +0 -86
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_condattr_getpshared.c +0 -97
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_condattr_init.c +0 -87
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_condattr_setpshared.c +0 -117
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_delay_np.c +0 -172
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_detach.c +0 -136
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_equal.c +0 -76
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_exit.c +0 -106
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_getconcurrency.c +0 -45
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_getschedparam.c +0 -75
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_getspecific.c +0 -87
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_getunique_np.c +0 -47
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_getw32threadhandle_np.c +0 -65
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_join.c +0 -157
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_key_create.c +0 -108
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_key_delete.c +0 -125
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_kill.c +0 -105
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_mutex_consistent.c +0 -187
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_mutex_destroy.c +0 -148
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_mutex_init.c +0 -130
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_mutex_lock.c +0 -269
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_mutex_timedlock.c +0 -324
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_mutex_trylock.c +0 -154
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_mutex_unlock.c +0 -175
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_mutexattr_destroy.c +0 -83
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_mutexattr_getkind_np.c +0 -44
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_mutexattr_getpshared.c +0 -95
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_mutexattr_getrobust.c +0 -113
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_mutexattr_gettype.c +0 -56
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_mutexattr_init.c +0 -86
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_mutexattr_setkind_np.c +0 -44
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_mutexattr_setpshared.c +0 -119
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_mutexattr_setrobust.c +0 -119
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_mutexattr_settype.c +0 -143
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_num_processors_np.c +0 -56
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_once.c +0 -79
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_rwlock_destroy.c +0 -143
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_rwlock_init.c +0 -109
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_rwlock_rdlock.c +0 -102
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_rwlock_timedrdlock.c +0 -109
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_rwlock_timedwrlock.c +0 -139
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_rwlock_tryrdlock.c +0 -102
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_rwlock_trywrlock.c +0 -122
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_rwlock_unlock.c +0 -93
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_rwlock_wrlock.c +0 -133
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_rwlockattr_destroy.c +0 -84
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_rwlockattr_getpshared.c +0 -97
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_rwlockattr_init.c +0 -83
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_rwlockattr_setpshared.c +0 -120
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_self.c +0 -141
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_setcancelstate.c +0 -125
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_setcanceltype.c +0 -126
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_setconcurrency.c +0 -53
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_setschedparam.c +0 -123
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_setspecific.c +0 -167
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_spin_destroy.c +0 -111
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_spin_init.c +0 -123
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_spin_lock.c +0 -80
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_spin_trylock.c +0 -77
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_spin_unlock.c +0 -71
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_testcancel.c +0 -103
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_timechange_handler_np.c +0 -108
- data/ext/noderb_extension/libuv/deps/pthread-win32/pthread_win32_attach_detach_np.c +0 -258
- data/ext/noderb_extension/libuv/deps/pthread-win32/ptw32_MCS_lock.c +0 -278
- data/ext/noderb_extension/libuv/deps/pthread-win32/ptw32_callUserDestroyRoutines.c +0 -232
- data/ext/noderb_extension/libuv/deps/pthread-win32/ptw32_calloc.c +0 -56
- data/ext/noderb_extension/libuv/deps/pthread-win32/ptw32_cond_check_need_init.c +0 -78
- data/ext/noderb_extension/libuv/deps/pthread-win32/ptw32_getprocessors.c +0 -91
- data/ext/noderb_extension/libuv/deps/pthread-win32/ptw32_is_attr.c +0 -47
- data/ext/noderb_extension/libuv/deps/pthread-win32/ptw32_mutex_check_need_init.c +0 -92
- data/ext/noderb_extension/libuv/deps/pthread-win32/ptw32_new.c +0 -94
- data/ext/noderb_extension/libuv/deps/pthread-win32/ptw32_processInitialize.c +0 -92
- data/ext/noderb_extension/libuv/deps/pthread-win32/ptw32_processTerminate.c +0 -105
- data/ext/noderb_extension/libuv/deps/pthread-win32/ptw32_relmillisecs.c +0 -132
- data/ext/noderb_extension/libuv/deps/pthread-win32/ptw32_reuse.c +0 -151
- data/ext/noderb_extension/libuv/deps/pthread-win32/ptw32_rwlock_cancelwrwait.c +0 -50
- data/ext/noderb_extension/libuv/deps/pthread-win32/ptw32_rwlock_check_need_init.c +0 -77
- data/ext/noderb_extension/libuv/deps/pthread-win32/ptw32_semwait.c +0 -135
- data/ext/noderb_extension/libuv/deps/pthread-win32/ptw32_spinlock_check_need_init.c +0 -78
- data/ext/noderb_extension/libuv/deps/pthread-win32/ptw32_threadDestroy.c +0 -79
- data/ext/noderb_extension/libuv/deps/pthread-win32/ptw32_threadStart.c +0 -357
- data/ext/noderb_extension/libuv/deps/pthread-win32/ptw32_throw.c +0 -189
- data/ext/noderb_extension/libuv/deps/pthread-win32/ptw32_timespec.c +0 -83
- data/ext/noderb_extension/libuv/deps/pthread-win32/ptw32_tkAssocCreate.c +0 -118
- data/ext/noderb_extension/libuv/deps/pthread-win32/ptw32_tkAssocDestroy.c +0 -114
- data/ext/noderb_extension/libuv/deps/pthread-win32/rwlock.c +0 -51
- data/ext/noderb_extension/libuv/deps/pthread-win32/sched.c +0 -53
- data/ext/noderb_extension/libuv/deps/pthread-win32/sched.h +0 -183
- data/ext/noderb_extension/libuv/deps/pthread-win32/sched_get_priority_max.c +0 -134
- data/ext/noderb_extension/libuv/deps/pthread-win32/sched_get_priority_min.c +0 -135
- data/ext/noderb_extension/libuv/deps/pthread-win32/sched_getscheduler.c +0 -71
- data/ext/noderb_extension/libuv/deps/pthread-win32/sched_setscheduler.c +0 -83
- data/ext/noderb_extension/libuv/deps/pthread-win32/sched_yield.c +0 -71
- data/ext/noderb_extension/libuv/deps/pthread-win32/sem_close.c +0 -58
- data/ext/noderb_extension/libuv/deps/pthread-win32/sem_destroy.c +0 -144
- data/ext/noderb_extension/libuv/deps/pthread-win32/sem_getvalue.c +0 -110
- data/ext/noderb_extension/libuv/deps/pthread-win32/sem_init.c +0 -169
- data/ext/noderb_extension/libuv/deps/pthread-win32/sem_open.c +0 -58
- data/ext/noderb_extension/libuv/deps/pthread-win32/sem_post.c +0 -128
- data/ext/noderb_extension/libuv/deps/pthread-win32/sem_post_multiple.c +0 -142
- data/ext/noderb_extension/libuv/deps/pthread-win32/sem_timedwait.c +0 -238
- data/ext/noderb_extension/libuv/deps/pthread-win32/sem_trywait.c +0 -117
- data/ext/noderb_extension/libuv/deps/pthread-win32/sem_unlink.c +0 -58
- data/ext/noderb_extension/libuv/deps/pthread-win32/sem_wait.c +0 -187
- data/ext/noderb_extension/libuv/deps/pthread-win32/semaphore.c +0 -69
- data/ext/noderb_extension/libuv/deps/pthread-win32/semaphore.h +0 -169
- data/ext/noderb_extension/libuv/deps/pthread-win32/signal.c +0 -179
- data/ext/noderb_extension/libuv/deps/pthread-win32/spin.c +0 -46
- data/ext/noderb_extension/libuv/deps/pthread-win32/sync.c +0 -43
- data/ext/noderb_extension/libuv/deps/pthread-win32/tsd.c +0 -44
- data/ext/noderb_extension/libuv/deps/pthread-win32/version.rc +0 -388
- data/ext/noderb_extension/libuv/deps/pthread-win32/w32_CancelableWait.c +0 -161
- data/ext/noderb_extension/libuv/doc/iocp-links.html +0 -574
- data/ext/noderb_extension/libuv/src/uv-unix.c +0 -2421
- data/ext/noderb_extension/libuv/src/win/ntdll.h +0 -130
- /data/ext/noderb_extension/libuv/include/{ev.h → uv-private/ev.h} +0 -0
- /data/ext/noderb_extension/libuv/include/{ngx-queue.h → uv-private/ngx-queue.h} +0 -0
- /data/ext/noderb_extension/libuv/include/{tree.h → uv-private/tree.h} +0 -0
- /data/ext/noderb_extension/libuv/src/{uv-cygwin.c → unix/cygwin.c} +0 -0
- /data/ext/noderb_extension/libuv/src/{uv-darwin.c → unix/darwin.c} +0 -0
- /data/ext/noderb_extension/libuv/src/{eio → unix/eio}/Changes +0 -0
- /data/ext/noderb_extension/libuv/src/{eio → unix/eio}/LICENSE +0 -0
- /data/ext/noderb_extension/libuv/src/{eio → unix/eio}/Makefile.am +0 -0
- /data/ext/noderb_extension/libuv/src/{eio → unix/eio}/aclocal.m4 +0 -0
- /data/ext/noderb_extension/libuv/src/{eio → unix/eio}/autogen.sh +0 -0
- /data/ext/noderb_extension/libuv/src/{eio → unix/eio}/config.h.in +0 -0
- /data/ext/noderb_extension/libuv/src/{eio → unix/eio}/config_cygwin.h +0 -0
- /data/ext/noderb_extension/libuv/src/{eio → unix/eio}/config_freebsd.h +0 -0
- /data/ext/noderb_extension/libuv/src/{eio → unix/eio}/config_sunos.h +0 -0
- /data/ext/noderb_extension/libuv/src/{eio → unix/eio}/configure.ac +0 -0
- /data/ext/noderb_extension/libuv/src/{eio → unix/eio}/demo.c +0 -0
- /data/ext/noderb_extension/libuv/src/{eio → unix/eio}/ecb.h +0 -0
- /data/ext/noderb_extension/libuv/src/{eio → unix/eio}/eio.3 +0 -0
- /data/ext/noderb_extension/libuv/src/{eio → unix/eio}/eio.pod +0 -0
- /data/ext/noderb_extension/libuv/src/{eio → unix/eio}/libeio.m4 +0 -0
- /data/ext/noderb_extension/libuv/src/{eio → unix/eio}/xthread.h +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/Changes +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/LICENSE +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/Makefile.am +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/Makefile.in +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/README +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/aclocal.m4 +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/autogen.sh +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/config.guess +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/config.h.in +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/config.sub +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/config_cygwin.h +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/config_darwin.h +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/config_freebsd.h +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/config_sunos.h +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/configure +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/configure.ac +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/depcomp +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/ev++.h +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/ev.3 +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/ev.c +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/ev.pod +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/ev_epoll.c +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/ev_kqueue.c +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/ev_poll.c +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/ev_port.c +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/ev_select.c +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/ev_vars.h +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/ev_win32.c +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/ev_wrap.h +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/event.c +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/event.h +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/install-sh +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/libev.m4 +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/ltmain.sh +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/missing +0 -0
- /data/ext/noderb_extension/libuv/src/{ev → unix/ev}/mkinstalldirs +0 -0
- /data/ext/noderb_extension/libuv/src/{uv-freebsd.c → unix/freebsd.c} +0 -0
- /data/ext/noderb_extension/libuv/src/{uv-linux.c → unix/linux.c} +0 -0
- /data/ext/noderb_extension/libuv/src/{uv-sunos.c → unix/sunos.c} +0 -0
@@ -1,1245 +0,0 @@
|
|
1
|
-
CURRENT CVS HEAD Version
|
2
|
-
RELEASE 2.9.0 pending
|
3
|
-
-------------
|
4
|
-
(2011-??-??)
|
5
|
-
|
6
|
-
General
|
7
|
-
-------
|
8
|
-
New bug fixes in this release since 2.8.0 have NOT been applied to the
|
9
|
-
1.x.x series.
|
10
|
-
|
11
|
-
Version 2.8.0 may be the last release for some older Windows systems.
|
12
|
-
Some changes post 2011-02-26 in CVS may not be compatible with pre
|
13
|
-
Windows 2000 systems.
|
14
|
-
|
15
|
-
Version 1 no longer maintained
|
16
|
-
------------------------------
|
17
|
-
The 1.x.x series is no longer maintained. However, if you really need a
|
18
|
-
version 1, the differences between 1.11.0 and 2.7.0 are very small, mainly
|
19
|
-
revolving around the pthread_once_t_ struct. Those differences applied
|
20
|
-
as a patch to the current 2.x.x should work. Don't forget to change
|
21
|
-
the version numbering in pthread.h before building. If you distribute
|
22
|
-
such a version 1.x.x please bear in mind that your numbers may clash
|
23
|
-
with those of others doing the same thing. Please consider also making
|
24
|
-
identifying changes in version.rc to differentiate your build.
|
25
|
-
|
26
|
-
Testing and verification
|
27
|
-
------------------------
|
28
|
-
The current CVS head version has been tested on an SMP architecture
|
29
|
-
(AMD Phenom 9750 Quad Core) by running the MinGW32 (GCC) builds against
|
30
|
-
the full test suite, stress tests and benchmarks.
|
31
|
-
|
32
|
-
New Features
|
33
|
-
------------
|
34
|
-
(MSC and GNU builds) The statically linked library now automatically
|
35
|
-
initialises and cleans up on program start/exit, i.e. statically linked
|
36
|
-
applications need not call the routines pthread_win32_process_attach_np()
|
37
|
-
and pthread_win32_process_detach_np() explicitly. The per-thread routine
|
38
|
-
pthread_win32_thread_detach_np() is also called at program exit to cleanup
|
39
|
-
POSIX resources acquired by the primary Windows native thread (if I (RJ)
|
40
|
-
understand the process correctly). Other Windows native threads that call
|
41
|
-
POSIX API routines may need to call the thread detach routine on thread
|
42
|
-
exit if the application depends on reclaimed POSIX resources or running
|
43
|
-
POSIX TSD (TLS) destructors.
|
44
|
-
See README.NONPORTABLE for descriptions of these routines.
|
45
|
-
- Ramiro Polla
|
46
|
-
|
47
|
-
Robust mutexes are implemented within the PROCESS_PRIVATE scope. NOTE that
|
48
|
-
pthread_mutex_* functions may return different error codes for robust
|
49
|
-
mutexes than they otherwise do in normal usage, e.g. pthread_mutex_unlock
|
50
|
-
is required to check ownership for all mutex types when the mutex is
|
51
|
-
robust, whereas this does not occur for the "normal" non-robust mutex type.
|
52
|
-
- Ross Johnson
|
53
|
-
|
54
|
-
pthread_getunique_np is implemented for source level compatibility
|
55
|
-
with some other implementations. This routine returns a 64 bit
|
56
|
-
sequence number that is uniquely associated with a thread. It can be
|
57
|
-
used by applications to order or hash POSIX thread handles.
|
58
|
-
- Ross Johnson
|
59
|
-
|
60
|
-
Bug fixes
|
61
|
-
---------
|
62
|
-
Many more changes for 64 bit systems.
|
63
|
-
- Kai Tietz
|
64
|
-
|
65
|
-
Various modifications and fixes to build and test for WinCE.
|
66
|
-
- Marcel Ruff, Sinan Kaya
|
67
|
-
|
68
|
-
Fix pthread_cond_destroy() - should not be a cancellation point. Other
|
69
|
-
minor build problems fixed.
|
70
|
-
- Romano Paolo Tenca
|
71
|
-
|
72
|
-
Remove potential deadlock condition from pthread_cond_destroy().
|
73
|
-
- Eric Berge
|
74
|
-
|
75
|
-
Various modifications to build and test for Win64.
|
76
|
-
- Kip Streithorst
|
77
|
-
|
78
|
-
Various fixes to the QueueUserAPCEx async cancellation helper DLL
|
79
|
-
(this is a separate download) and pthreads code cleanups.
|
80
|
-
- Sebastian Gottschalk
|
81
|
-
|
82
|
-
Removed potential NULL pointer reference.
|
83
|
-
- Robert Kindred
|
84
|
-
|
85
|
-
Removed the requirement that applications restrict the number of threads
|
86
|
-
calling pthread_barrier_wait to just the barrier count. Also reduced the
|
87
|
-
contention between barrier_wait and barrier_destroy. This change will have
|
88
|
-
slowed barriers down slightly but halves the number of semaphores consumed
|
89
|
-
per barrier to one.
|
90
|
-
- Ross Johnson
|
91
|
-
|
92
|
-
Fixed a handle leak in sched_[gs]etscheduler.
|
93
|
-
- Mark Pizzolato
|
94
|
-
|
95
|
-
Removed all of the POSIX re-entrant function compatibility macros from pthread.h.
|
96
|
-
Some were simply not semanticly correct.
|
97
|
-
- Igor Lubashev
|
98
|
-
|
99
|
-
Threads no longer attempt to pass uncaught exceptions out of thread scope (C++
|
100
|
-
and SEH builds only). Uncaught exceptions now cause the thread to exit with
|
101
|
-
the return code PTHREAD_CANCELED.
|
102
|
-
- Ross Johnson
|
103
|
-
|
104
|
-
Lots of casting fixes particularly for x64, Interlocked fixes and reworking
|
105
|
-
for x64.
|
106
|
-
- Daniel Richard G., John Kamp
|
107
|
-
|
108
|
-
Other changes
|
109
|
-
-------------
|
110
|
-
Dependence on the winsock library is now discretionary via
|
111
|
-
#define RETAIN_WSALASTERROR in config.h. It is undefined by default unless
|
112
|
-
WINCE is defined (because RJ is unsure of the dependency there).
|
113
|
-
- Ramiro Polla
|
114
|
-
|
115
|
-
Several static POSIX mutexes used for internal management were replaced by
|
116
|
-
MCS queue-based locks to reduce resource consumption, in particular use of Win32
|
117
|
-
objects.
|
118
|
-
- Ross Johnson
|
119
|
-
|
120
|
-
For security, the QuserEx.dll if used must now be installed in the Windows System
|
121
|
-
folder.
|
122
|
-
- Ross Johnson
|
123
|
-
|
124
|
-
New tests
|
125
|
-
---------
|
126
|
-
robust[1-5].c - Robust mutexes
|
127
|
-
sequence1.c - per-thread unique sequence numbers
|
128
|
-
|
129
|
-
Modified tests and benchtests
|
130
|
-
-----------------------------
|
131
|
-
All mutex*.c tests wherever appropriate have been modified to also test
|
132
|
-
robust mutexes under the same conditions.
|
133
|
-
Added robust mutex benchtests to benchtest*.c wherever appropriate.
|
134
|
-
|
135
|
-
|
136
|
-
RELEASE 2.8.0
|
137
|
-
-------------
|
138
|
-
(2006-12-22)
|
139
|
-
|
140
|
-
General
|
141
|
-
-------
|
142
|
-
New bug fixes in this release since 2.7.0 have not been applied to the
|
143
|
-
version 1.x.x series. It is probably time to drop version 1.
|
144
|
-
|
145
|
-
Testing and verification
|
146
|
-
------------------------
|
147
|
-
This release has not yet been tested on SMP architechtures. All tests pass
|
148
|
-
on a uni-processor system.
|
149
|
-
|
150
|
-
Bug fixes
|
151
|
-
---------
|
152
|
-
Sem_destroy could return EBUSY even though no threads were waiting on the
|
153
|
-
semaphore. Other races around invalidating semaphore structs (internally)
|
154
|
-
have been removed as well.
|
155
|
-
|
156
|
-
New tests
|
157
|
-
---------
|
158
|
-
semaphore5.c - tests the bug fix referred to above.
|
159
|
-
|
160
|
-
|
161
|
-
RELEASE 2.7.0
|
162
|
-
-------------
|
163
|
-
(2005-06-04)
|
164
|
-
|
165
|
-
General
|
166
|
-
-------
|
167
|
-
All new features in this release have been back-ported in release 1.11.0,
|
168
|
-
including the incorporation of MCS locks in pthread_once, however, versions
|
169
|
-
1 and 2 remain incompatible even though they are now identical in
|
170
|
-
performance and functionality.
|
171
|
-
|
172
|
-
Testing and verification
|
173
|
-
------------------------
|
174
|
-
This release has been tested (passed the test suite) on both uni-processor
|
175
|
-
and multi-processor systems.
|
176
|
-
- Tim Theisen
|
177
|
-
|
178
|
-
Bug fixes
|
179
|
-
---------
|
180
|
-
Pthread_once has been re-implemented to remove priority boosting and other
|
181
|
-
complexity to improve robustness. Races for Win32 handles that are not
|
182
|
-
recycle-unique have been removed. The general form of pthread_once is now
|
183
|
-
the same as that suggested earlier by Alexander Terekhov, but instead of the
|
184
|
-
'named mutex', a queue-based lock has been implemented which has the required
|
185
|
-
properties of dynamic self initialisation and destruction. This lock is also
|
186
|
-
efficient. The ABI is unaffected in as much as the size of pthread_once_t has
|
187
|
-
not changed and PTHREAD_ONCE_INIT has not changed, however, applications that
|
188
|
-
peek inside pthread_once_t, which is supposed to be opaque, will break.
|
189
|
-
- Vladimir Kliatchko
|
190
|
-
|
191
|
-
New features
|
192
|
-
------------
|
193
|
-
* Support for Mingw cross development tools added to GNUmakefile.
|
194
|
-
Mingw cross tools allow building the libraries on Linux.
|
195
|
-
- Mikael Magnusson
|
196
|
-
|
197
|
-
|
198
|
-
RELEASE 2.6.0
|
199
|
-
-------------
|
200
|
-
(2005-05-19)
|
201
|
-
|
202
|
-
General
|
203
|
-
-------
|
204
|
-
All of the bug fixes and new features in this release have been
|
205
|
-
back-ported in release 1.10.0.
|
206
|
-
|
207
|
-
Testing and verification
|
208
|
-
------------------------
|
209
|
-
This release has been tested (passed the test suite) on both uni-processor
|
210
|
-
and multi-processor systems. Thanks to Tim Theisen at TomoTherapy for
|
211
|
-
exhaustively running the MP tests and for providing crutial observations
|
212
|
-
and data when faults are detected.
|
213
|
-
|
214
|
-
Bugs fixed
|
215
|
-
----------
|
216
|
-
|
217
|
-
* pthread_detach() now reclaims remaining thread resources if called after
|
218
|
-
the target thread has terminated. Previously, this routine did nothing in
|
219
|
-
this case.
|
220
|
-
|
221
|
-
New tests
|
222
|
-
---------
|
223
|
-
|
224
|
-
* detach1.c - tests that pthread_detach properly invalidates the target
|
225
|
-
thread, which indicates that the thread resources have been reclaimed.
|
226
|
-
|
227
|
-
|
228
|
-
RELEASE 2.5.0
|
229
|
-
-------------
|
230
|
-
(2005-05-09)
|
231
|
-
|
232
|
-
General
|
233
|
-
-------
|
234
|
-
|
235
|
-
The package now includes a reference documentation set consisting of
|
236
|
-
HTML formatted Unix-style manual pages that have been edited for
|
237
|
-
consistency with Pthreads-w32. The set can also be read online at:
|
238
|
-
http://sources.redhat.com/pthreads-win32/manual/index.html
|
239
|
-
|
240
|
-
Thanks again to Tim Theisen for running the test suite pre-release
|
241
|
-
on an MP system.
|
242
|
-
|
243
|
-
All of the bug fixes and new features in this release have been
|
244
|
-
back-ported in release 1.9.0.
|
245
|
-
|
246
|
-
Bugs fixed
|
247
|
-
----------
|
248
|
-
|
249
|
-
* Thread Specific Data (TSD) key management has been ammended to
|
250
|
-
eliminate a source of (what was effectively) resource leakage (a HANDLE
|
251
|
-
plus memory for each key destruct routine/thread association). This was
|
252
|
-
not a true leak because these resources were eventually reclaimed when
|
253
|
-
pthread_key_delete was run AND each thread referencing the key had exited.
|
254
|
-
The problem was that these two conditions are often not met until very
|
255
|
-
late, and often not until the process is about to exit.
|
256
|
-
|
257
|
-
The ammended implementation avoids the need for the problematic HANDLE
|
258
|
-
and reclaims the memory as soon as either the key is deleted OR the
|
259
|
-
thread exits, whichever is first.
|
260
|
-
|
261
|
-
Thanks to Richard Hughes at Aculab for identifying and locating the leak.
|
262
|
-
|
263
|
-
* TSD key destructors are now processed up to PTHREAD_DESTRUCTOR_ITERATIONS
|
264
|
-
times instead of just once. PTHREAD_DESTRUCTOR_ITERATIONS has been
|
265
|
-
defined in pthread.h for some time but not used.
|
266
|
-
|
267
|
-
* Fix a semaphore accounting race between sem_post/sem_post_multiple
|
268
|
-
and sem_wait cancellation. This is the same issue as with
|
269
|
-
sem_timedwait that was fixed in the last release.
|
270
|
-
|
271
|
-
* sem_init, sem_post, and sem_post_multiple now check that the
|
272
|
-
semaphore count never exceeds _POSIX_SEM_VALUE_MAX.
|
273
|
-
|
274
|
-
* Although sigwait() is nothing more than a no-op, it should at least
|
275
|
-
be a cancellation point to be consistent with the standard.
|
276
|
-
|
277
|
-
New tests
|
278
|
-
---------
|
279
|
-
|
280
|
-
* stress1.c - attempts to expose problems in condition variable
|
281
|
-
and semaphore timed wait logic. This test was inspired by Stephan
|
282
|
-
Mueller's sample test code used to identify the sem_timedwait bug
|
283
|
-
from the last release. It's not a part of the regular test suite
|
284
|
-
because it can take awhile to run. To run it:
|
285
|
-
nmake clean VC-stress
|
286
|
-
|
287
|
-
* tsd2.c - tests that key destructors are re-run if the tsd key value is
|
288
|
-
not NULL after the destructor routine has run. Also tests that
|
289
|
-
pthread_setspecific() and pthread_getspecific() are callable from
|
290
|
-
destructors.
|
291
|
-
|
292
|
-
|
293
|
-
RELEASE 2.4.0
|
294
|
-
-------------
|
295
|
-
(2005-04-26)
|
296
|
-
|
297
|
-
General
|
298
|
-
-------
|
299
|
-
|
300
|
-
There is now no plan to release a version 3.0.0 to fix problems in
|
301
|
-
pthread_once(). Other possible implementations of pthread_once
|
302
|
-
will still be investigated for a possible future release in an attempt
|
303
|
-
to reduce the current implementation's complexity.
|
304
|
-
|
305
|
-
All of the bug fixes and new features in this release have been
|
306
|
-
back-ported for release 1.8.0.
|
307
|
-
|
308
|
-
Bugs fixed
|
309
|
-
----------
|
310
|
-
|
311
|
-
* Fixed pthread_once race (failures on an MP system). Thanks to
|
312
|
-
Tim Theisen for running exhaustive pre-release testing on his MP system
|
313
|
-
using a range of compilers:
|
314
|
-
VC++ 6
|
315
|
-
VC++ 7.1
|
316
|
-
Intel C++ version 8.0
|
317
|
-
All tests passed.
|
318
|
-
Some minor speed improvements were also done.
|
319
|
-
|
320
|
-
* Fix integer overrun error in pthread_mutex_timedlock() - missed when
|
321
|
-
sem_timedwait() was fixed in release 2.2.0. This routine no longer returns
|
322
|
-
ENOTSUP when NEED_SEM is defined - it is supported (NEED_SEM is only
|
323
|
-
required for WinCE versions prior to 3.0).
|
324
|
-
|
325
|
-
* Fix timeout bug in sem_timedwait().
|
326
|
-
- Thanks to Stephan Mueller for reporting, providing diagnostic output
|
327
|
-
and test code.
|
328
|
-
|
329
|
-
* Fix several problems in the NEED_SEM conditionally included code.
|
330
|
-
NEED_SEM included code is provided for systems that don't implement W32
|
331
|
-
semaphores, such as WinCE prior to version 3.0. An alternate implementation
|
332
|
-
of POSIX semaphores is built using W32 events for these systems when
|
333
|
-
NEED_SEM is defined. This code has been completely rewritten in this
|
334
|
-
release to reuse most of the default POSIX semaphore code, and particularly,
|
335
|
-
to implement all of the sem_* routines supported by pthreads-win32. Tim
|
336
|
-
Theisen also run the test suite over the NEED_SEM code on his MP system. All
|
337
|
-
tests passed.
|
338
|
-
|
339
|
-
* The library now builds without errors for the Borland Builder 5.5 compiler.
|
340
|
-
|
341
|
-
New features
|
342
|
-
------------
|
343
|
-
|
344
|
-
* pthread_mutex_timedlock() and all sem_* routines provided by
|
345
|
-
pthreads-win32 are now implemented for WinCE versions prior to 3.0. Those
|
346
|
-
versions did not implement W32 semaphores. Define NEED_SEM in config.h when
|
347
|
-
building the library for these systems.
|
348
|
-
|
349
|
-
Known issues in this release
|
350
|
-
----------------------------
|
351
|
-
|
352
|
-
* pthread_once is too complicated - but it works as far as testing can
|
353
|
-
determine..
|
354
|
-
|
355
|
-
* The Borland version of the dll fails some of the tests with a memory read
|
356
|
-
exception. The cause is not yet known but a compiler bug has not been ruled
|
357
|
-
out.
|
358
|
-
|
359
|
-
|
360
|
-
RELEASE 2.3.0
|
361
|
-
-------------
|
362
|
-
(2005-04-12)
|
363
|
-
|
364
|
-
General
|
365
|
-
-------
|
366
|
-
|
367
|
-
Release 1.7.0 is a backport of features and bug fixes new in
|
368
|
-
this release. See earlier notes under Release 2.0.0/General.
|
369
|
-
|
370
|
-
Bugs fixed
|
371
|
-
----------
|
372
|
-
|
373
|
-
* Fixed pthread_once potential for post once_routine cancellation
|
374
|
-
hanging due to starvation. See comments in pthread_once.c.
|
375
|
-
Momentary priority boosting is used to ensure that, after a
|
376
|
-
once_routine is cancelled, the thread that will run the
|
377
|
-
once_routine is not starved by higher priority waiting threads at
|
378
|
-
critical times. Priority boosting occurs only AFTER a once_routine
|
379
|
-
cancellation, and is applied only to that once_control. The
|
380
|
-
once_routine is run at the thread's normal base priority.
|
381
|
-
|
382
|
-
New tests
|
383
|
-
---------
|
384
|
-
|
385
|
-
* once4.c: Aggressively tests pthread_once() under realtime
|
386
|
-
conditions using threads with varying priorities. Windows'
|
387
|
-
random priority boosting does not occur for threads with realtime
|
388
|
-
priority levels.
|
389
|
-
|
390
|
-
|
391
|
-
RELEASE 2.2.0
|
392
|
-
-------------
|
393
|
-
(2005-04-04)
|
394
|
-
|
395
|
-
General
|
396
|
-
-------
|
397
|
-
|
398
|
-
* Added makefile targets to build static link versions of the library.
|
399
|
-
Both MinGW and MSVC. Please note that this does not imply any change
|
400
|
-
to the LGPL licensing, which still imposes psecific conditions on
|
401
|
-
distributing software that has been statically linked with this library.
|
402
|
-
|
403
|
-
* There is a known bug in pthread_once(). Cancellation of the init_routine
|
404
|
-
exposes a potential starvation (i.e. deadlock) problem if a waiting thread
|
405
|
-
has a higher priority than the initting thread. This problem will be fixed
|
406
|
-
in version 3.0.0 of the library.
|
407
|
-
|
408
|
-
Bugs fixed
|
409
|
-
----------
|
410
|
-
|
411
|
-
* Fix integer overrun error in sem_timedwait().
|
412
|
-
Kevin Lussier
|
413
|
-
|
414
|
-
* Fix preprocessor directives for static linking.
|
415
|
-
Dimitar Panayotov
|
416
|
-
|
417
|
-
|
418
|
-
RELEASE 2.1.0
|
419
|
-
-------------
|
420
|
-
(2005-03-16)
|
421
|
-
|
422
|
-
Bugs fixed
|
423
|
-
----------
|
424
|
-
|
425
|
-
* Reverse change to pthread_setcancelstate() in 2.0.0.
|
426
|
-
|
427
|
-
|
428
|
-
RELEASE 2.0.0
|
429
|
-
-------------
|
430
|
-
(2005-03-16)
|
431
|
-
|
432
|
-
General
|
433
|
-
-------
|
434
|
-
|
435
|
-
This release represents an ABI change and the DLL version naming has
|
436
|
-
incremented from 1 to 2, e.g. pthreadVC2.dll.
|
437
|
-
|
438
|
-
Version 1.4.0 back-ports the new functionality included in this
|
439
|
-
release. Please distribute DLLs built from that version with updates
|
440
|
-
to applications built on pthreads-win32 version 1.x.x.
|
441
|
-
|
442
|
-
The package naming has changed, replacing the snapshot date with
|
443
|
-
the version number + descriptive information. E.g. this
|
444
|
-
release is "pthreads-w32-2-0-0-release".
|
445
|
-
|
446
|
-
Bugs fixed
|
447
|
-
----------
|
448
|
-
|
449
|
-
* pthread_setcancelstate() no longer checks for a pending
|
450
|
-
async cancel event if the library is using alertable async
|
451
|
-
cancel. See the README file (Prerequisites section) for info
|
452
|
-
on adding alertable async cancelation.
|
453
|
-
|
454
|
-
New features
|
455
|
-
------------
|
456
|
-
|
457
|
-
* pthread_once() now supports init_routine cancellability.
|
458
|
-
|
459
|
-
New tests
|
460
|
-
---------
|
461
|
-
|
462
|
-
* Agressively test pthread_once() init_routine cancellability.
|
463
|
-
|
464
|
-
|
465
|
-
SNAPSHOT 2005-03-08
|
466
|
-
-------------------
|
467
|
-
Version 1.3.0
|
468
|
-
|
469
|
-
Bug reports (fixed)
|
470
|
-
-------------------
|
471
|
-
|
472
|
-
* Implicitly created threads leave Win32 handles behind after exiting.
|
473
|
-
- Dmitrii Semii
|
474
|
-
|
475
|
-
* pthread_once() starvation problem.
|
476
|
-
- Gottlob Frege
|
477
|
-
|
478
|
-
New tests
|
479
|
-
---------
|
480
|
-
|
481
|
-
* More intense testing of pthread_once().
|
482
|
-
|
483
|
-
|
484
|
-
SNAPSHOT 2005-01-25
|
485
|
-
-------------------
|
486
|
-
Version 1.2.0
|
487
|
-
|
488
|
-
Bug fixes
|
489
|
-
---------
|
490
|
-
|
491
|
-
* Attempted acquisition of a recursive mutex could cause waiting threads
|
492
|
-
to not be woken when the mutex was released.
|
493
|
-
- Ralf Kubis <RKubis at mc.com>
|
494
|
-
|
495
|
-
* Various package omissions have been fixed.
|
496
|
-
|
497
|
-
|
498
|
-
SNAPSHOT 2005-01-03
|
499
|
-
-------------------
|
500
|
-
Version 1.1.0
|
501
|
-
|
502
|
-
Bug fixes
|
503
|
-
---------
|
504
|
-
|
505
|
-
* Unlocking recursive or errorcheck mutexes would sometimes
|
506
|
-
unexpectedly return an EPERM error (bug introduced in
|
507
|
-
snapshot-2004-11-03).
|
508
|
-
- Konstantin Voronkov <beowinkle at yahoo.com>
|
509
|
-
|
510
|
-
|
511
|
-
SNAPSHOT 2004-11-22
|
512
|
-
-------------------
|
513
|
-
Version 1.0.0
|
514
|
-
|
515
|
-
This snapshot primarily fixes the condvar bug introduced in
|
516
|
-
snapshot-2004-11-03. DLL versioning has also been included to allow
|
517
|
-
applications to runtime check the Microsoft compatible DLL version
|
518
|
-
information, and to extend the DLL naming system for ABI and major
|
519
|
-
(non-backward compatible) API changes. See the README file for details.
|
520
|
-
|
521
|
-
Bug fixes
|
522
|
-
---------
|
523
|
-
|
524
|
-
* Condition variables no longer deadlock (bug introduced in
|
525
|
-
snapshot-2004-11-03).
|
526
|
-
- Alexander Kotliarov and Nicolas at saintmac
|
527
|
-
|
528
|
-
* DLL naming extended to avoid 'DLL hell' in the future, and to
|
529
|
-
accommodate the ABI change introduced in snapshot-2004-11-03. Snapshot
|
530
|
-
2004-11-03 will be removed from FTP sites.
|
531
|
-
|
532
|
-
New features
|
533
|
-
------------
|
534
|
-
|
535
|
-
* A Microsoft-style version resource has been added to the DLL for
|
536
|
-
applications that wish to check DLL compatibility at runtime.
|
537
|
-
|
538
|
-
* Pthreads-win32 DLL naming has been extended to allow incompatible DLL
|
539
|
-
versions to co-exist in the same filesystem. See the README file for details,
|
540
|
-
but briefly: while the version information inside the DLL will change with
|
541
|
-
each release from now on, the DLL version names will only change if the new
|
542
|
-
DLL is not backward compatible with older applications.
|
543
|
-
|
544
|
-
The versioning scheme has been borrowed from GNU Libtool, and the DLL
|
545
|
-
naming scheme is from Cygwin. Provided the Libtool-style numbering rules are
|
546
|
-
honoured, the Cygwin DLL naming scheme automatcally ensures that DLL name
|
547
|
-
changes are minimal and that applications will not load an incompatible
|
548
|
-
pthreads-win32 DLL.
|
549
|
-
|
550
|
-
Those who use the pre-built DLLs will find that the DLL/LIB names have a new
|
551
|
-
suffix (1) in this snapshot. E.g. pthreadVC1.dll etc.
|
552
|
-
|
553
|
-
* The POSIX thread ID reuse uniqueness feature introduced in the last snapshot
|
554
|
-
has been kept as default, but the behaviour can now be controlled when the DLL
|
555
|
-
is built to effectively switch it off. This makes the library much more
|
556
|
-
sensitive to applications that assume that POSIX thread IDs are unique, i.e.
|
557
|
-
are not strictly compliant with POSIX. See the PTW32_THREAD_ID_REUSE_INCREMENT
|
558
|
-
macro comments in config.h for details.
|
559
|
-
|
560
|
-
Other changes
|
561
|
-
-------------
|
562
|
-
Certain POSIX macros have changed.
|
563
|
-
|
564
|
-
These changes are intended to conform to the Single Unix Specification version 3,
|
565
|
-
which states that, if set to 0 (zero) or not defined, then applications may use
|
566
|
-
sysconf() to determine their values at runtime. Pthreads-win32 does not
|
567
|
-
implement sysconf().
|
568
|
-
|
569
|
-
The following macros are no longer undefined, but defined and set to -1
|
570
|
-
(not implemented):
|
571
|
-
|
572
|
-
_POSIX_THREAD_ATTR_STACKADDR
|
573
|
-
_POSIX_THREAD_PRIO_INHERIT
|
574
|
-
_POSIX_THREAD_PRIO_PROTECT
|
575
|
-
_POSIX_THREAD_PROCESS_SHARED
|
576
|
-
|
577
|
-
The following macros are defined and set to 200112L (implemented):
|
578
|
-
|
579
|
-
_POSIX_THREADS
|
580
|
-
_POSIX_THREAD_SAFE_FUNCTIONS
|
581
|
-
_POSIX_THREAD_ATTR_STACKSIZE
|
582
|
-
_POSIX_THREAD_PRIORITY_SCHEDULING
|
583
|
-
_POSIX_SEMAPHORES
|
584
|
-
_POSIX_READER_WRITER_LOCKS
|
585
|
-
_POSIX_SPIN_LOCKS
|
586
|
-
_POSIX_BARRIERS
|
587
|
-
|
588
|
-
The following macros are defined and set to appropriate values:
|
589
|
-
|
590
|
-
_POSIX_THREAD_THREADS_MAX
|
591
|
-
_POSIX_SEM_VALUE_MAX
|
592
|
-
_POSIX_SEM_NSEMS_MAX
|
593
|
-
PTHREAD_DESTRUCTOR_ITERATIONS
|
594
|
-
PTHREAD_KEYS_MAX
|
595
|
-
PTHREAD_STACK_MIN
|
596
|
-
PTHREAD_THREADS_MAX
|
597
|
-
|
598
|
-
|
599
|
-
SNAPSHOT 2004-11-03
|
600
|
-
-------------------
|
601
|
-
|
602
|
-
DLLs produced from this snapshot cannot be used with older applications without
|
603
|
-
recompiling the application, due to a change to pthread_t to provide unique POSIX
|
604
|
-
thread IDs.
|
605
|
-
|
606
|
-
Although this snapshot passes the extended test suite, many of the changes are
|
607
|
-
fairly major, and some applications may show different behaviour than previously,
|
608
|
-
so adopt with care. Hopefully, any changed behaviour will be due to the library
|
609
|
-
being better at it's job, not worse.
|
610
|
-
|
611
|
-
Bug fixes
|
612
|
-
---------
|
613
|
-
|
614
|
-
* pthread_create() no longer accepts NULL as the thread reference arg.
|
615
|
-
A segfault (memory access fault) will result, and no thread will be
|
616
|
-
created.
|
617
|
-
|
618
|
-
* pthread_barrier_wait() no longer acts as a cancelation point.
|
619
|
-
|
620
|
-
* Fix potential race condition in pthread_once()
|
621
|
-
- Tristan Savatier <tristan at mpegtv.com>
|
622
|
-
|
623
|
-
* Changes to pthread_cond_destroy() exposed some coding weaknesses in several
|
624
|
-
test suite mini-apps because pthread_cond_destroy() now returns EBUSY if the CV
|
625
|
-
is still in use.
|
626
|
-
|
627
|
-
New features
|
628
|
-
------------
|
629
|
-
|
630
|
-
* Added for compatibility:
|
631
|
-
PTHREAD_RECURSIVE_MUTEX_INITIALIZER,
|
632
|
-
PTHREAD_ERRORCHECK_MUTEX_INITIALIZER,
|
633
|
-
PTHREAD_RECURSIVE_MUTEX_INITIALIZER_NP,
|
634
|
-
PTHREAD_ERRORCHECK_MUTEX_INITIALIZER_NP
|
635
|
-
|
636
|
-
* Initial support for Digital Mars compiler
|
637
|
-
- Anuj Goyal <anuj.goyal at gmail.com>
|
638
|
-
|
639
|
-
* Faster Mutexes. These have been been rewritten following a model provided by
|
640
|
-
Alexander Terekhov that reduces kernel space checks, and eliminates some additional
|
641
|
-
critical sections used to manage a race between timedlock expiration and unlock.
|
642
|
-
Please be aware that the new mutexes do not enforce strict absolute FIFO scheduling
|
643
|
-
of mutexes, however any out-of-order lock acquisition should be very rare.
|
644
|
-
|
645
|
-
* Faster semaphores. Following a similar model to mutexes above, these have been
|
646
|
-
rewritten to use preliminary users space checks.
|
647
|
-
|
648
|
-
* sem_getvalue() now returns the number of waiters.
|
649
|
-
|
650
|
-
* The POSIX thread ID now has much stronger uniqueness characteristics. The library
|
651
|
-
garrantees not to reuse the same thread ID for at least 2^(wordsize) thread
|
652
|
-
destruction/creation cycles.
|
653
|
-
|
654
|
-
New tests
|
655
|
-
---------
|
656
|
-
|
657
|
-
* semaphore4.c: Tests cancelation of the new sem_wait().
|
658
|
-
|
659
|
-
* semaphore4t.c: Likewise for sem_timedwait().
|
660
|
-
|
661
|
-
* rwlock8.c: Tests and times the slow execution paths of r/w locks, and the CVs,
|
662
|
-
mutexes, and semaphores that they're built on.
|
663
|
-
|
664
|
-
|
665
|
-
SNAPSHOT 2004-05-16
|
666
|
-
-------------------
|
667
|
-
|
668
|
-
Attempt to add Watcom to the list of compilers that can build the library.
|
669
|
-
This failed in the end due to it's non-thread-aware errno. The library
|
670
|
-
builds but the test suite fails. See README.Watcom for more details.
|
671
|
-
|
672
|
-
Bug fixes
|
673
|
-
---------
|
674
|
-
* Bug and memory leak in sem_init()
|
675
|
-
- Alex Blanco <Alex.Blanco at motorola.com>
|
676
|
-
|
677
|
-
* ptw32_getprocessors() now returns CPU count of 1 for WinCE.
|
678
|
-
- James Ewing <james.ewing at sveasoft.com>
|
679
|
-
|
680
|
-
* pthread_cond_wait() could be canceled at a point where it should not
|
681
|
-
be cancelable. Fixed.
|
682
|
-
- Alexander Terekhov <TEREKHOV at de.ibm.com>
|
683
|
-
|
684
|
-
* sem_timedwait() had an incorrect timeout calculation.
|
685
|
-
- Philippe Di Cristo <philipped at voicebox.com>
|
686
|
-
|
687
|
-
* Fix a memory leak left behind after threads are destroyed.
|
688
|
-
- P. van Bruggen <pietvb at newbridges.nl>
|
689
|
-
|
690
|
-
New features
|
691
|
-
------------
|
692
|
-
* Ported to AMD64.
|
693
|
-
- Makoto Kato <raven at oldskool.jp>
|
694
|
-
|
695
|
-
* True pre-emptive asynchronous cancelation of threads. This is optional
|
696
|
-
and requires that Panagiotis E. Hadjidoukas's QueueUserAPCEx package be
|
697
|
-
installed. This package is included in the pthreads-win32 self-unpacking
|
698
|
-
Zip archive starting from this snapshot. See the README.txt file inside
|
699
|
-
the package for installation details.
|
700
|
-
|
701
|
-
Note: If you don't use async cancelation in your application, or don't need
|
702
|
-
to cancel threads that are blocked on system resources such as network I/O,
|
703
|
-
then the default non-preemptive async cancelation is probably good enough.
|
704
|
-
However, pthreads-win32 auto-detects the availability of these components
|
705
|
-
at run-time, so you don't need to rebuild the library from source if you
|
706
|
-
change your mind later.
|
707
|
-
|
708
|
-
All of the advice available in books and elsewhere on the undesirability
|
709
|
-
of using async cancelation in any application still stands, but this
|
710
|
-
feature is a welcome addition with respect to the library's conformance to
|
711
|
-
the POSIX standard.
|
712
|
-
|
713
|
-
SNAPSHOT 2003-09-18
|
714
|
-
-------------------
|
715
|
-
|
716
|
-
Cleanup of thread priority management. In particular, setting of thread
|
717
|
-
priority now attempts to map invalid Win32 values within the range returned
|
718
|
-
by sched_get_priority_min/max() to useful values. See README.NONPORTABLE
|
719
|
-
under "Thread priority".
|
720
|
-
|
721
|
-
Bug fixes
|
722
|
-
---------
|
723
|
-
* pthread_getschedparam() now returns the priority given by the most recent
|
724
|
-
call to pthread_setschedparam() or established by pthread_create(), as
|
725
|
-
required by the standard. Previously, pthread_getschedparam() incorrectly
|
726
|
-
returned the running thread priority at the time of the call, which may have
|
727
|
-
been adjusted or temporarily promoted/demoted.
|
728
|
-
|
729
|
-
* sched_get_priority_min() and sched_get_priority_max() now return -1 on error
|
730
|
-
and set errno. Previously, they incorrectly returned the error value directly.
|
731
|
-
|
732
|
-
|
733
|
-
SNAPSHOT 2003-09-04
|
734
|
-
-------------------
|
735
|
-
|
736
|
-
Bug fixes
|
737
|
-
---------
|
738
|
-
* ptw32_cancelableWait() now allows cancelation of waiting implicit POSIX
|
739
|
-
threads.
|
740
|
-
|
741
|
-
New test
|
742
|
-
--------
|
743
|
-
* cancel8.c tests cancelation of Win32 threads waiting at a POSIX cancelation
|
744
|
-
point.
|
745
|
-
|
746
|
-
|
747
|
-
SNAPSHOT 2003-09-03
|
748
|
-
-------------------
|
749
|
-
|
750
|
-
Bug fixes
|
751
|
-
---------
|
752
|
-
* pthread_self() would free the newly created implicit POSIX thread handle if
|
753
|
-
DuplicateHandle failed instead of recycle it (very unlikely).
|
754
|
-
|
755
|
-
* pthread_exit() was neither freeing nor recycling the POSIX thread struct
|
756
|
-
for implicit POSIX threads.
|
757
|
-
|
758
|
-
New feature - Cancelation of/by Win32 (non-POSIX) threads
|
759
|
-
---------------------------------------------------------
|
760
|
-
Since John Bossom's original implementation, the library has allowed non-POSIX
|
761
|
-
initialised threads (Win32 threads) to call pthreads-win32 routines and
|
762
|
-
therefore interact with POSIX threads. This is done by creating an on-the-fly
|
763
|
-
POSIX thread ID for the Win32 thread that, once created, allows fully
|
764
|
-
reciprical interaction. This did not extend to thread cancelation (async or
|
765
|
-
deferred). Now it does.
|
766
|
-
|
767
|
-
Any thread can be canceled by any other thread (Win32 or POSIX) if the former
|
768
|
-
thread's POSIX pthread_t value is known. It's TSD destructors and POSIX
|
769
|
-
cleanup handlers will be run before the thread exits with an exit code of
|
770
|
-
PTHREAD_CANCELED (retrieved with GetExitCodeThread()).
|
771
|
-
|
772
|
-
This allows a Win32 thread to, for example, call POSIX CV routines in the same way
|
773
|
-
that POSIX threads would/should, with pthread_cond_wait() cancelability and
|
774
|
-
cleanup handlers (pthread_cond_wait() is a POSIX cancelation point).
|
775
|
-
|
776
|
-
By adding cancelation, Win32 threads should now be able to call all POSIX
|
777
|
-
threads routines that make sense including semaphores, mutexes, condition
|
778
|
-
variables, read/write locks, barriers, spinlocks, tsd, cleanup push/pop,
|
779
|
-
cancelation, pthread_exit, scheduling, etc.
|
780
|
-
|
781
|
-
Note that these on-the-fly 'implicit' POSIX thread IDs are initialised as detached
|
782
|
-
(not joinable) with deferred cancelation type. The POSIX thread ID will be created
|
783
|
-
automatically by any POSIX routines that need a POSIX handle (unless the routine
|
784
|
-
needs a pthread_t as a parameter of course). A Win32 thread can discover it's own
|
785
|
-
POSIX thread ID by calling pthread_self(), which will create the handle if
|
786
|
-
necessary and return the pthread_t value.
|
787
|
-
|
788
|
-
New tests
|
789
|
-
---------
|
790
|
-
Test the above new feature.
|
791
|
-
|
792
|
-
|
793
|
-
SNAPSHOT 2003-08-19
|
794
|
-
-------------------
|
795
|
-
|
796
|
-
This snapshot fixes some accidental corruption to new test case sources.
|
797
|
-
There are no changes to the library source code.
|
798
|
-
|
799
|
-
|
800
|
-
SNAPSHOT 2003-08-15
|
801
|
-
-------------------
|
802
|
-
|
803
|
-
Bug fixes
|
804
|
-
---------
|
805
|
-
|
806
|
-
* pthread.dsp now uses correct compile flags (/MD).
|
807
|
-
- Viv <vcotirlea@hotmail.com>
|
808
|
-
|
809
|
-
* pthread_win32_process_detach_np() fixed memory leak.
|
810
|
-
- Steven Reddie <Steven.Reddie@ca.com>
|
811
|
-
|
812
|
-
* pthread_mutex_destroy() fixed incorrect return code.
|
813
|
-
- Nicolas Barry <boozai@yahoo.com>
|
814
|
-
|
815
|
-
* pthread_spin_destroy() fixed memory leak.
|
816
|
-
- Piet van Bruggen <pietvb@newbridges.nl>
|
817
|
-
|
818
|
-
* Various changes to tighten arg checking, and to work with later versions of
|
819
|
-
MinGW32 and MsysDTK.
|
820
|
-
|
821
|
-
* pthread_getschedparam() etc, fixed dangerous thread validity checking.
|
822
|
-
- Nicolas Barry <boozai@yahoo.com>
|
823
|
-
|
824
|
-
* POSIX thread handles are now reused and their memory is not freed on thread exit.
|
825
|
-
This allows for stronger thread validity checking.
|
826
|
-
|
827
|
-
New standard routine
|
828
|
-
--------------------
|
829
|
-
|
830
|
-
* pthread_kill() added to provide thread validity checking to applications.
|
831
|
-
It does not accept any non zero values for the signal arg.
|
832
|
-
|
833
|
-
New test cases
|
834
|
-
--------------
|
835
|
-
|
836
|
-
* New test cases to confirm validity checking, pthread_kill(), and thread reuse.
|
837
|
-
|
838
|
-
|
839
|
-
SNAPSHOT 2003-05-10
|
840
|
-
-------------------
|
841
|
-
|
842
|
-
Bug fixes
|
843
|
-
---------
|
844
|
-
|
845
|
-
* pthread_mutex_trylock() now returns correct error values.
|
846
|
-
pthread_mutex_destroy() will no longer destroy a recursively locked mutex.
|
847
|
-
pthread_mutex_lock() is no longer inadvertantly behaving as a cancelation point.
|
848
|
-
- Thomas Pfaff <tpfaff@gmx.net>
|
849
|
-
|
850
|
-
* pthread_mutex_timedlock() no longer occasionally sets incorrect mutex
|
851
|
-
ownership, causing deadlocks in some applications.
|
852
|
-
- Robert Strycek <strycek@posam.sk> and Alexander Terekhov <TEREKHOV@de.ibm.com>
|
853
|
-
|
854
|
-
|
855
|
-
SNAPSHOT 2002-11-04
|
856
|
-
-------------------
|
857
|
-
|
858
|
-
Bug fixes
|
859
|
-
---------
|
860
|
-
|
861
|
-
* sem_getvalue() now returns the correct value under Win NT and WinCE.
|
862
|
-
- Rob Fanner <rfanner@stonethree.com>
|
863
|
-
|
864
|
-
* sem_timedwait() now uses tighter checks for unreasonable
|
865
|
-
abstime values - that would result in unexpected timeout values.
|
866
|
-
|
867
|
-
* ptw32_cond_wait_cleanup() no longer mysteriously consumes
|
868
|
-
CV signals but may produce more spurious wakeups. It is believed
|
869
|
-
that the sem_timedwait() call is consuming a CV signal that it
|
870
|
-
shouldn't.
|
871
|
-
- Alexander Terekhov <TEREKHOV@de.ibm.com>
|
872
|
-
|
873
|
-
* Fixed a memory leak in ptw32_threadDestroy() for implicit threads.
|
874
|
-
|
875
|
-
* Fixed potential for deadlock in pthread_cond_destroy().
|
876
|
-
A deadlock could occur for statically declared CVs (PTHREAD_COND_INITIALIZER),
|
877
|
-
when one thread is attempting to destroy the condition variable while another
|
878
|
-
is attempting to dynamically initialize it.
|
879
|
-
- Michael Johnson <michaelj@maine.rr.com>
|
880
|
-
|
881
|
-
|
882
|
-
SNAPSHOT 2002-03-02
|
883
|
-
-------------------
|
884
|
-
|
885
|
-
Cleanup code default style. (IMPORTANT)
|
886
|
-
----------------------------------------------------------------------
|
887
|
-
Previously, if not defined, the cleanup style was determined automatically
|
888
|
-
from the compiler/language, and one of the following was defined accordingly:
|
889
|
-
|
890
|
-
__CLEANUP_SEH MSVC only
|
891
|
-
__CLEANUP_CXX C++, including MSVC++, GNU G++
|
892
|
-
__CLEANUP_C C, including GNU GCC, not MSVC
|
893
|
-
|
894
|
-
These defines determine the style of cleanup (see pthread.h) and,
|
895
|
-
most importantly, the way that cancelation and thread exit (via
|
896
|
-
pthread_exit) is performed (see the routine ptw32_throw() in private.c).
|
897
|
-
|
898
|
-
In short, the exceptions versions of the library throw an exception
|
899
|
-
when a thread is canceled or exits (via pthread_exit()), which is
|
900
|
-
caught by a handler in the thread startup routine, so that the
|
901
|
-
the correct stack unwinding occurs regardless of where the thread
|
902
|
-
is when it's canceled or exits via pthread_exit().
|
903
|
-
|
904
|
-
In this and future snapshots, unless the build explicitly defines (e.g.
|
905
|
-
via a compiler option) __CLEANUP_SEH, __CLEANUP_CXX, or __CLEANUP_C, then
|
906
|
-
the build NOW always defaults to __CLEANUP_C style cleanup. This style
|
907
|
-
uses setjmp/longjmp in the cancelation and pthread_exit implementations,
|
908
|
-
and therefore won't do stack unwinding even when linked to applications
|
909
|
-
that have it (e.g. C++ apps). This is for consistency with most
|
910
|
-
current commercial Unix POSIX threads implementations. Compaq's TRU64
|
911
|
-
may be an exception (no pun intended) and possible future trend.
|
912
|
-
|
913
|
-
Although it was not clearly documented before, it is still necessary to
|
914
|
-
build your application using the same __CLEANUP_* define as was
|
915
|
-
used for the version of the library that you link with, so that the
|
916
|
-
correct parts of pthread.h are included. That is, the possible
|
917
|
-
defines require the following library versions:
|
918
|
-
|
919
|
-
__CLEANUP_SEH pthreadVSE.dll
|
920
|
-
__CLEANUP_CXX pthreadVCE.dll or pthreadGCE.dll
|
921
|
-
__CLEANUP_C pthreadVC.dll or pthreadGC.dll
|
922
|
-
|
923
|
-
E.g. regardless of whether your app is C or C++, if you link with
|
924
|
-
pthreadVC.lib or libpthreadGC.a, then you must define __CLEANUP_C.
|
925
|
-
|
926
|
-
|
927
|
-
THE POINT OF ALL THIS IS: if you have not been defining one of these
|
928
|
-
explicitly, then the defaults as described at the top of this
|
929
|
-
section were being used.
|
930
|
-
|
931
|
-
THIS NOW CHANGES, as has been explained above, but to try to make this
|
932
|
-
clearer here's an example:
|
933
|
-
|
934
|
-
If you were building your application with MSVC++ i.e. using C++
|
935
|
-
exceptions and not explicitly defining one of __CLEANUP_*, then
|
936
|
-
__CLEANUP_C++ was automatically defined for you in pthread.h.
|
937
|
-
You should have been linking with pthreadVCE.dll, which does
|
938
|
-
stack unwinding.
|
939
|
-
|
940
|
-
If you now build your application as you had before, pthread.h will now
|
941
|
-
automatically set __CLEANUP_C as the default style, and you will need to
|
942
|
-
link with pthreadVC.dll. Stack unwinding will now NOT occur when a thread
|
943
|
-
is canceled, or the thread calls pthread_exit().
|
944
|
-
|
945
|
-
Your application will now most likely behave differently to previous
|
946
|
-
versions, and in non-obvious ways. Most likely is that locally
|
947
|
-
instantiated objects may not be destroyed or cleaned up after a thread
|
948
|
-
is canceled.
|
949
|
-
|
950
|
-
If you want the same behaviour as before, then you must now define
|
951
|
-
__CLEANUP_C++ explicitly using a compiler option and link with
|
952
|
-
pthreadVCE.dll as you did before.
|
953
|
-
|
954
|
-
|
955
|
-
WHY ARE WE MAKING THE DEFAULT STYLE LESS EXCEPTION-FRIENDLY?
|
956
|
-
Because no commercial Unix POSIX threads implementation allows you to
|
957
|
-
choose to have stack unwinding. Therefore, providing it in pthread-win32
|
958
|
-
as a default is dangerous. We still provide the choice but unless
|
959
|
-
you consciously choose to do otherwise, your pthreads applications will
|
960
|
-
now run or crash in similar ways irrespective of the threads platform
|
961
|
-
you use. Or at least this is the hope.
|
962
|
-
|
963
|
-
|
964
|
-
WHY NOT REMOVE THE EXCEPTIONS VERSIONS OF THE LIBRARY ALTOGETHER?
|
965
|
-
There are a few reasons:
|
966
|
-
- because there are well respected POSIX threads people who believe
|
967
|
-
that POSIX threads implementations should be exceptions aware and
|
968
|
-
do the expected thing in that context. (There are equally respected
|
969
|
-
people who believe it should not be easily accessible, if it's there
|
970
|
-
at all, for unconditional conformity to other implementations.)
|
971
|
-
- because pthreads-win32 is one of the few implementations that has
|
972
|
-
the choice, perhaps the only freely available one, and so offers
|
973
|
-
a laboratory to people who may want to explore the effects;
|
974
|
-
- although the code will always be around somewhere for anyone who
|
975
|
-
wants it, once it's removed from the current version it will not be
|
976
|
-
nearly as visible to people who may have a use for it.
|
977
|
-
|
978
|
-
|
979
|
-
Source module splitting
|
980
|
-
-----------------------
|
981
|
-
In order to enable smaller image sizes to be generated
|
982
|
-
for applications that link statically with the library,
|
983
|
-
most routines have been separated out into individual
|
984
|
-
source code files.
|
985
|
-
|
986
|
-
This is being done in such a way as to be backward compatible.
|
987
|
-
The old source files are reused to congregate the individual
|
988
|
-
routine files into larger translation units (via a bunch of
|
989
|
-
# includes) so that the compiler can still optimise wherever
|
990
|
-
possible, e.g. through inlining, which can only be done
|
991
|
-
within the same translation unit.
|
992
|
-
|
993
|
-
It is also possible to build the entire library by compiling
|
994
|
-
the single file named "pthread.c", which just #includes all
|
995
|
-
the secondary congregation source files. The compiler
|
996
|
-
may be able to use this to do more inlining of routines.
|
997
|
-
|
998
|
-
Although the GNU compiler is able to produce libraries with
|
999
|
-
the necessary separation (the -ffunction-segments switch),
|
1000
|
-
AFAIK, the MSVC and other compilers don't have this feature.
|
1001
|
-
|
1002
|
-
Finally, since I use makefiles and command-line compilation,
|
1003
|
-
I don't know what havoc this reorganisation may wreak amongst
|
1004
|
-
IDE project file users. You should be able to continue
|
1005
|
-
using your existing project files without modification.
|
1006
|
-
|
1007
|
-
|
1008
|
-
New non-portable functions
|
1009
|
-
--------------------------
|
1010
|
-
pthread_num_processors_np():
|
1011
|
-
Returns the number of processors in the system that are
|
1012
|
-
available to the process, as determined from the processor
|
1013
|
-
affinity mask.
|
1014
|
-
|
1015
|
-
pthread_timechange_handler_np():
|
1016
|
-
To improve tolerance against operator or time service initiated
|
1017
|
-
system clock changes.
|
1018
|
-
|
1019
|
-
This routine can be called by an application when it
|
1020
|
-
receives a WM_TIMECHANGE message from the system. At present
|
1021
|
-
it broadcasts all condition variables so that waiting threads
|
1022
|
-
can wake up and re-evaluate their conditions and restart
|
1023
|
-
their timed waits if required.
|
1024
|
-
- Suggested by Alexander Terekhov
|
1025
|
-
|
1026
|
-
|
1027
|
-
Platform dependence
|
1028
|
-
-------------------
|
1029
|
-
As Win95 doesn't provide one, the library now contains
|
1030
|
-
it's own InterlockedCompareExchange() routine, which is used
|
1031
|
-
whenever Windows doesn't provide it. InterlockedCompareExchange()
|
1032
|
-
is used to implement spinlocks and barriers, and also in mutexes.
|
1033
|
-
This routine relies on the CMPXCHG machine instruction which
|
1034
|
-
is not available on i386 CPUs. This library (from snapshot
|
1035
|
-
20010712 onwards) is therefore no longer supported on i386
|
1036
|
-
processor platforms.
|
1037
|
-
|
1038
|
-
|
1039
|
-
New standard routines
|
1040
|
-
---------------------
|
1041
|
-
For source code portability only - rwlocks cannot be process shared yet.
|
1042
|
-
|
1043
|
-
pthread_rwlockattr_init()
|
1044
|
-
pthread_rwlockattr_destroy()
|
1045
|
-
pthread_rwlockattr_setpshared()
|
1046
|
-
pthread_rwlockattr_getpshared()
|
1047
|
-
|
1048
|
-
As defined in the new POSIX standard, and the Single Unix Spec version 3:
|
1049
|
-
|
1050
|
-
sem_timedwait()
|
1051
|
-
pthread_mutex_timedlock() - Alexander Terekhov and Thomas Pfaff
|
1052
|
-
pthread_rwlock_timedrdlock() - adapted from pthread_rwlock_rdlock()
|
1053
|
-
pthread_rwlock_timedwrlock() - adapted from pthread_rwlock_wrlock()
|
1054
|
-
|
1055
|
-
|
1056
|
-
pthread.h no longer includes windows.h
|
1057
|
-
--------------------------------------
|
1058
|
-
[Not yet for G++]
|
1059
|
-
|
1060
|
-
This was done to prevent conflicts.
|
1061
|
-
|
1062
|
-
HANDLE, DWORD, and NULL are temporarily defined within pthread.h if
|
1063
|
-
they are not already.
|
1064
|
-
|
1065
|
-
|
1066
|
-
pthread.h, sched.h and semaphore.h now use dllexport/dllimport
|
1067
|
-
--------------------------------------------------------------
|
1068
|
-
Not only to avoid the need for the pthread.def file, but to
|
1069
|
-
improve performance. Apparently, declaring functions with dllimport
|
1070
|
-
generates a direct call to the function and avoids the overhead
|
1071
|
-
of a stub function call.
|
1072
|
-
|
1073
|
-
Bug fixes
|
1074
|
-
---------
|
1075
|
-
* Fixed potential NULL pointer dereferences in pthread_mutexattr_init,
|
1076
|
-
pthread_mutexattr_getpshared, pthread_barrierattr_init,
|
1077
|
-
pthread_barrierattr_getpshared, and pthread_condattr_getpshared.
|
1078
|
-
- Scott McCaskill <scott@magruder.org>
|
1079
|
-
|
1080
|
-
* Removed potential race condition in pthread_mutex_trylock and
|
1081
|
-
pthread_mutex_lock;
|
1082
|
-
- Alexander Terekhov <TEREKHOV@de.ibm.com>
|
1083
|
-
|
1084
|
-
* The behaviour of pthread_mutex_trylock in relation to
|
1085
|
-
recursive mutexes was inconsistent with commercial implementations.
|
1086
|
-
Trylock would return EBUSY if the lock was owned already by the
|
1087
|
-
calling thread regardless of mutex type. Trylock now increments the
|
1088
|
-
recursion count and returns 0 for RECURSIVE mutexes, and will
|
1089
|
-
return EDEADLK rather than EBUSY for ERRORCHECK mutexes. This is
|
1090
|
-
consistent with Solaris.
|
1091
|
-
- Thomas Pfaff <tpfaff@gmx.net>
|
1092
|
-
|
1093
|
-
* Found a fix for the library and workaround for applications for
|
1094
|
-
the known bug #2, i.e. where __CLEANUP_CXX or __CLEANUP_SEH is defined.
|
1095
|
-
See the "Known Bugs in this snapshot" section below.
|
1096
|
-
|
1097
|
-
This could be made transparent to applications by replacing the macros that
|
1098
|
-
define the current C++ and SEH versions of pthread_cleanup_push/pop
|
1099
|
-
with the C version, but AFAIK cleanup handlers would not then run in the
|
1100
|
-
correct sequence with destructors and exception cleanup handlers when
|
1101
|
-
an exception occurs.
|
1102
|
-
|
1103
|
-
* Cancelation once started in a thread cannot now be inadvertantly
|
1104
|
-
double canceled. That is, once a thread begins it's cancelation run,
|
1105
|
-
cancelation is disabled and a subsequent cancel request will
|
1106
|
-
return an error (ESRCH).
|
1107
|
-
|
1108
|
-
* errno: An incorrect compiler directive caused a local version
|
1109
|
-
of errno to be used instead of the Win32 errno. Both instances are
|
1110
|
-
thread-safe but applications checking errno after a pthreads-win32
|
1111
|
-
call would be wrong. Fixing this also fixed a bad compiler
|
1112
|
-
option in the testsuite (/MT should have been /MD) which is
|
1113
|
-
needed to link with the correct library MSVCRT.LIB.
|
1114
|
-
|
1115
|
-
|
1116
|
-
SNAPSHOT 2001-07-12
|
1117
|
-
-------------------
|
1118
|
-
|
1119
|
-
To be added
|
1120
|
-
|
1121
|
-
|
1122
|
-
SNAPSHOT 2001-07-03
|
1123
|
-
-------------------
|
1124
|
-
|
1125
|
-
To be added
|
1126
|
-
|
1127
|
-
|
1128
|
-
SNAPSHOT 2000-08-13
|
1129
|
-
-------------------
|
1130
|
-
|
1131
|
-
New:
|
1132
|
-
- Renamed DLL and LIB files:
|
1133
|
-
pthreadVSE.dll (MS VC++/Structured EH)
|
1134
|
-
pthreadVSE.lib
|
1135
|
-
pthreadVCE.dll (MS VC++/C++ EH)
|
1136
|
-
pthreadVCE.lib
|
1137
|
-
pthreadGCE.dll (GNU G++/C++ EH)
|
1138
|
-
libpthreadw32.a
|
1139
|
-
|
1140
|
-
Both your application and the pthread dll should use the
|
1141
|
-
same exception handling scheme.
|
1142
|
-
|
1143
|
-
Bugs fixed:
|
1144
|
-
- MSVC++ C++ exception handling.
|
1145
|
-
|
1146
|
-
Some new tests have been added.
|
1147
|
-
|
1148
|
-
|
1149
|
-
SNAPSHOT 2000-08-10
|
1150
|
-
-------------------
|
1151
|
-
|
1152
|
-
New:
|
1153
|
-
- asynchronous cancelation on X86 (Jason Nye)
|
1154
|
-
- Makefile compatible with MS nmake to replace
|
1155
|
-
buildlib.bat
|
1156
|
-
- GNUmakefile for Mingw32
|
1157
|
-
- tests/Makefile for MS nmake replaces runall.bat
|
1158
|
-
- tests/GNUmakefile for Mingw32
|
1159
|
-
|
1160
|
-
Bugs fixed:
|
1161
|
-
- kernel32 load/free problem
|
1162
|
-
- attempt to hide internel exceptions from application
|
1163
|
-
exception handlers (__try/__except and try/catch blocks)
|
1164
|
-
- Win32 thread handle leakage bug
|
1165
|
-
(David Baggett/Paul Redondo/Eyal Lebedinsky)
|
1166
|
-
|
1167
|
-
Some new tests have been added.
|
1168
|
-
|
1169
|
-
|
1170
|
-
SNAPSHOT 1999-11-02
|
1171
|
-
-------------------
|
1172
|
-
|
1173
|
-
Bugs fixed:
|
1174
|
-
- ctime_r macro had an incorrect argument (Erik Hensema),
|
1175
|
-
- threads were not being created
|
1176
|
-
PTHREAD_CANCEL_DEFERRED. This should have
|
1177
|
-
had little effect as deferred is the only
|
1178
|
-
supported type. (Ross Johnson).
|
1179
|
-
|
1180
|
-
Some compatibility improvements added, eg.
|
1181
|
-
- pthread_setcancelstate accepts NULL pointer
|
1182
|
-
for the previous value argument. Ditto for
|
1183
|
-
pthread_setcanceltype. This is compatible
|
1184
|
-
with Solaris but should not affect
|
1185
|
-
standard applications (Erik Hensema)
|
1186
|
-
|
1187
|
-
Some new tests have been added.
|
1188
|
-
|
1189
|
-
|
1190
|
-
SNAPSHOT 1999-10-17
|
1191
|
-
-------------------
|
1192
|
-
|
1193
|
-
Bug fix - Cancelation of threads waiting on condition variables
|
1194
|
-
now works properly (Lorin Hochstein and Peter Slacik)
|
1195
|
-
|
1196
|
-
|
1197
|
-
SNAPSHOT 1999-08-12
|
1198
|
-
-------------------
|
1199
|
-
|
1200
|
-
Fixed exception stack cleanup if calling pthread_exit()
|
1201
|
-
- (Lorin Hochstein and John Bossom).
|
1202
|
-
|
1203
|
-
Fixed bugs in condition variables - (Peter Slacik):
|
1204
|
-
- additional contention checks
|
1205
|
-
- properly adjust number of waiting threads after timed
|
1206
|
-
condvar timeout.
|
1207
|
-
|
1208
|
-
|
1209
|
-
SNAPSHOT 1999-05-30
|
1210
|
-
-------------------
|
1211
|
-
|
1212
|
-
Some minor bugs have been fixed. See the ChangeLog file for details.
|
1213
|
-
|
1214
|
-
Some more POSIX 1b functions are now included but ony return an
|
1215
|
-
error (ENOSYS) if called. They are:
|
1216
|
-
|
1217
|
-
sem_open
|
1218
|
-
sem_close
|
1219
|
-
sem_unlink
|
1220
|
-
sem_getvalue
|
1221
|
-
|
1222
|
-
|
1223
|
-
SNAPSHOT 1999-04-07
|
1224
|
-
-------------------
|
1225
|
-
|
1226
|
-
Some POSIX 1b functions which were internally supported are now
|
1227
|
-
available as exported functions:
|
1228
|
-
|
1229
|
-
sem_init
|
1230
|
-
sem_destroy
|
1231
|
-
sem_wait
|
1232
|
-
sem_trywait
|
1233
|
-
sem_post
|
1234
|
-
sched_yield
|
1235
|
-
sched_get_priority_min
|
1236
|
-
sched_get_priority_max
|
1237
|
-
|
1238
|
-
Some minor bugs have been fixed. See the ChangeLog file for details.
|
1239
|
-
|
1240
|
-
|
1241
|
-
SNAPSHOT 1999-03-16
|
1242
|
-
-------------------
|
1243
|
-
|
1244
|
-
Initial release.
|
1245
|
-
|