status-im
/
nim-chronos
mirror of https://github.com/status-im/nim-chronos.git
2
0
Fork 0
Code Issues Projects Releases Wiki Activity
nim-chronos/chronos/ioselects/ioselectors_kqueue.nim

626 lines
21 KiB
Nim
Raw Normal View History

exception tracking (#166) * exception tracking This PR adds minimal exception tracking to chronos, moving the goalpost one step further. In particular, it becomes invalid to raise exceptions from `callSoon` callbacks: this is critical for writing correct error handling because there's no reasonable way that a user of chronos can possibly _reason_ about exceptions coming out of there: the event loop will be in an indeterminite state when the loop is executing an _random_ callback. As expected, there are several issues in the error handling of chronos: in particular, it will end up in an inconsistent internal state whenever the selector loop operations fail, because the internal state update functions are not written in an exception-safe way. This PR turns this into a Defect, which probably is not the optimal way of handling things - expect more work to be done here. Some API have no way of reporting back errors to callers - for example, when something fails in the accept loop, there's not much it can do, and no way to report it back to the user of the API - this has been fixed with the new accept flow - the old one should be deprecated. Finally, there is information loss in the API: in composite operations like `poll` and `waitFor` there's no way to differentiate internal errors from user-level errors originating from callbacks. * store `CatchableError` in future * annotate proc's with correct raises information * `selectors2` to avoid non-CatchableError IOSelectorsException * `$` should never raise * remove unnecessary gcsafe annotations * fix exceptions leaking out of timer waits * fix some imports * functions must signal raising the union of all exceptions across all platforms to enable cross-platform code * switch to unittest2 * add `selectors2` which supercedes the std library version and fixes several exception handling issues in there * fixes * docs, platform-independent eh specifiers for some functions * add feature flag for strict exception mode also bump version to 3.0.0 - _most_ existing code should be compatible with this version of exception handling but some things might need fixing - callbacks, existing raises specifications etc. * fix AsyncCheck for non-void T
2021-03-24 09:08:33 +00:00
#
#
# Nim's Runtime Library
# (c) Copyright 2016 Eugene Kabanov
#
# See the file "copying.txt", included in this
# distribution, for details about the copyright.
#
# This module implements BSD kqueue().
import posix, times, kqueue
const
# Maximum number of events that can be returned.
MAX_KQUEUE_EVENTS = 64
# SIG_IGN and SIG_DFL declared in posix.nim as variables, but we need them
# to be constants and GC-safe.
SIG_DFL = cast[proc(x: cint) {.raises: [],noconv,gcsafe.}](0)
SIG_IGN = cast[proc(x: cint) {.raises: [],noconv,gcsafe.}](1)
when defined(kqcache):
const CACHE_EVENTS = true
when defined(macosx) or defined(freebsd) or defined(dragonfly):
when defined(macosx):
const MAX_DESCRIPTORS_ID = 29 # KERN_MAXFILESPERPROC (MacOS)
else:
const MAX_DESCRIPTORS_ID = 27 # KERN_MAXFILESPERPROC (FreeBSD)
proc sysctl(name: ptr cint, namelen: cuint, oldp: pointer, oldplen: ptr csize_t,
newp: pointer, newplen: csize_t): cint
{.importc: "sysctl",header: """#include <sys/types.h>
#include <sys/sysctl.h>"""}
elif defined(netbsd) or defined(openbsd):
# OpenBSD and NetBSD don't have KERN_MAXFILESPERPROC, so we are using
# KERN_MAXFILES, because KERN_MAXFILES is always bigger,
# than KERN_MAXFILESPERPROC.
const MAX_DESCRIPTORS_ID = 7 # KERN_MAXFILES
proc sysctl(name: ptr cint, namelen: cuint, oldp: pointer, oldplen: ptr csize_t,
newp: pointer, newplen: csize_t): cint
{.importc: "sysctl",header: """#include <sys/param.h>
#include <sys/sysctl.h>"""}
when hasThreadSupport:
type
SelectorImpl[T] = object
kqFD: cint
maxFD: int
changes: ptr SharedArray[KEvent]
fds: ptr SharedArray[SelectorKey[T]]
count: int
changesLock: Lock
changesSize: int
changesLength: int
sock: cint
Selector*[T] = ptr SelectorImpl[T]
else:
type
SelectorImpl[T] = object
kqFD: cint
maxFD: int
changes: seq[KEvent]
fds: seq[SelectorKey[T]]
count: int
sock: cint
Selector*[T] = ref SelectorImpl[T]
type
SelectEventImpl = object
rfd: cint
wfd: cint
SelectEvent* = ptr SelectEventImpl
# SelectEvent is declared as `ptr` to be placed in `shared memory`,
# so you can share one SelectEvent handle between threads.
proc getUnique[T](s: Selector[T]): int {.inline.} =
# we create duplicated handles to get unique indexes for our `fds` array.
result = posix.fcntl(s.sock, F_DUPFD, s.sock)
if result == -1:
raiseIOSelectorsError(osLastError())
proc newSelector*[T](): owned(Selector[T]) =
var maxFD = 0.cint
var size = csize_t(sizeof(cint))
var namearr = [1.cint, MAX_DESCRIPTORS_ID.cint]
# Obtain maximum number of opened file descriptors for process
if sysctl(addr(namearr[0]), 2, cast[pointer](addr maxFD), addr size,
nil, 0) != 0:
raiseIOSelectorsError(osLastError())
var kqFD = kqueue()
if kqFD < 0:
raiseIOSelectorsError(osLastError())
# we allocating empty socket to duplicate it handle in future, to get unique
# indexes for `fds` array. This is needed to properly identify
# {Event.Timer, Event.Signal, Event.Process} events.
let usock = posix.socket(posix.AF_INET, posix.SOCK_STREAM,
posix.IPPROTO_TCP).cint
if usock == -1:
let err = osLastError()
discard posix.close(kqFD)
raiseIOSelectorsError(err)
when hasThreadSupport:
result = cast[Selector[T]](allocShared0(sizeof(SelectorImpl[T])))
result.fds = allocSharedArray[SelectorKey[T]](maxFD)
result.changes = allocSharedArray[KEvent](MAX_KQUEUE_EVENTS)
result.changesSize = MAX_KQUEUE_EVENTS
initLock(result.changesLock)
else:
result = Selector[T]()
result.fds = newSeq[SelectorKey[T]](maxFD)
result.changes = newSeqOfCap[KEvent](MAX_KQUEUE_EVENTS)
for i in 0 ..< maxFD:
result.fds[i].ident = InvalidIdent
result.sock = usock
result.kqFD = kqFD
result.maxFD = maxFD.int
proc close*[T](s: Selector[T]) =
let res1 = posix.close(s.kqFD)
let res2 = posix.close(s.sock)
when hasThreadSupport:
deinitLock(s.changesLock)
deallocSharedArray(s.fds)
deallocShared(cast[pointer](s))
if res1 != 0 or res2 != 0:
raiseIOSelectorsError(osLastError())
proc newSelectEvent*(): SelectEvent =
var fds: array[2, cint]
if posix.pipe(fds) != 0:
raiseIOSelectorsError(osLastError())
setNonBlocking(fds[0])
setNonBlocking(fds[1])
result = cast[SelectEvent](allocShared0(sizeof(SelectEventImpl)))
result.rfd = fds[0]
result.wfd = fds[1]
proc trigger*(ev: SelectEvent) =
var data: uint64 = 1
if posix.write(ev.wfd, addr data, sizeof(uint64)) != sizeof(uint64):
raiseIOSelectorsError(osLastError())
proc close*(ev: SelectEvent) =
let res1 = posix.close(ev.rfd)
let res2 = posix.close(ev.wfd)
deallocShared(cast[pointer](ev))
if res1 != 0 or res2 != 0:
raiseIOSelectorsError(osLastError())
template checkFd(s, f) =
if f >= s.maxFD:
raiseIOSelectorsError("Maximum number of descriptors is exhausted!")
when hasThreadSupport:
template withChangeLock[T](s: Selector[T], body: untyped) =
acquire(s.changesLock)
{.locks: [s.changesLock].}:
try:
body
finally:
release(s.changesLock)
else:
template withChangeLock(s, body: untyped) =
body
when hasThreadSupport:
template modifyKQueue[T](s: Selector[T], nident: uint, nfilter: cshort,
nflags: cushort, nfflags: cuint, ndata: int,
nudata: pointer) =
mixin withChangeLock
s.withChangeLock():
if s.changesLength == s.changesSize:
# if cache array is full, we allocating new with size * 2
let newSize = s.changesSize shl 1
let rdata = allocSharedArray[KEvent](newSize)
copyMem(rdata, s.changes, s.changesSize * sizeof(KEvent))
s.changesSize = newSize
s.changes[s.changesLength] = KEvent(ident: nident,
filter: nfilter, flags: nflags,
fflags: nfflags, data: ndata,
udata: nudata)
inc(s.changesLength)
when not declared(CACHE_EVENTS):
template flushKQueue[T](s: Selector[T]) =
mixin withChangeLock
s.withChangeLock():
if s.changesLength > 0:
if kevent(s.kqFD, addr(s.changes[0]), cint(s.changesLength),
nil, 0, nil) == -1:
raiseIOSelectorsError(osLastError())
s.changesLength = 0
else:
template modifyKQueue[T](s: Selector[T], nident: uint, nfilter: cshort,
nflags: cushort, nfflags: cuint, ndata: int,
nudata: pointer) =
s.changes.add(KEvent(ident: nident,
filter: nfilter, flags: nflags,
fflags: nfflags, data: ndata,
udata: nudata))
when not declared(CACHE_EVENTS):
template flushKQueue[T](s: Selector[T]) =
let length = cint(len(s.changes))
if length > 0:
if kevent(s.kqFD, addr(s.changes[0]), length,
nil, 0, nil) == -1:
raiseIOSelectorsError(osLastError())
s.changes.setLen(0)
proc registerHandle*[T](s: Selector[T], fd: int | SocketHandle,
events: set[Event], data: T) =
let fdi = int(fd)
s.checkFd(fdi)
doAssert(s.fds[fdi].ident == InvalidIdent)
s.setKey(fdi, events, 0, data)
if events != {}:
if Event.Read in events:
modifyKQueue(s, uint(fdi), EVFILT_READ, EV_ADD, 0, 0, nil)
inc(s.count)
if Event.Write in events:
modifyKQueue(s, uint(fdi), EVFILT_WRITE, EV_ADD, 0, 0, nil)
inc(s.count)
when not declared(CACHE_EVENTS):
flushKQueue(s)
proc updateHandle*[T](s: Selector[T], fd: int | SocketHandle,
events: set[Event]) =
let maskEvents = {Event.Timer, Event.Signal, Event.Process, Event.Vnode,
Event.User, Event.Oneshot, Event.Error}
let fdi = int(fd)
s.checkFd(fdi)
var pkey = addr(s.fds[fdi])
doAssert(pkey.ident != InvalidIdent,
"Descriptor $# is not registered in the queue!" % $fdi)
doAssert(pkey.events * maskEvents == {})
if pkey.events != events:
if (Event.Read in pkey.events) and (Event.Read notin events):
modifyKQueue(s, fdi.uint, EVFILT_READ, EV_DELETE, 0, 0, nil)
dec(s.count)
if (Event.Write in pkey.events) and (Event.Write notin events):
modifyKQueue(s, fdi.uint, EVFILT_WRITE, EV_DELETE, 0, 0, nil)
dec(s.count)
if (Event.Read notin pkey.events) and (Event.Read in events):
modifyKQueue(s, fdi.uint, EVFILT_READ, EV_ADD, 0, 0, nil)
inc(s.count)
if (Event.Write notin pkey.events) and (Event.Write in events):
modifyKQueue(s, fdi.uint, EVFILT_WRITE, EV_ADD, 0, 0, nil)
inc(s.count)
when not declared(CACHE_EVENTS):
flushKQueue(s)
pkey.events = events
proc registerTimer*[T](s: Selector[T], timeout: int, oneshot: bool,
data: T): int {.discardable.} =
let fdi = getUnique(s)
s.checkFd(fdi)
doAssert(s.fds[fdi].ident == InvalidIdent)
let events = if oneshot: {Event.Timer, Event.Oneshot} else: {Event.Timer}
let flags: cushort = if oneshot: EV_ONESHOT or EV_ADD else: EV_ADD
s.setKey(fdi, events, 0, data)
# EVFILT_TIMER on Open/Net(BSD) has granularity of only milliseconds,
# but MacOS and FreeBSD allow use `0` as `fflags` to use milliseconds
# too
modifyKQueue(s, fdi.uint, EVFILT_TIMER, flags, 0, cint(timeout), nil)
when not declared(CACHE_EVENTS):
flushKQueue(s)
inc(s.count)
result = fdi
proc registerSignal*[T](s: Selector[T], signal: int,
data: T): int {.discardable.} =
let fdi = getUnique(s)
s.checkFd(fdi)
doAssert(s.fds[fdi].ident == InvalidIdent)
s.setKey(fdi, {Event.Signal}, signal, data)
var nmask, omask: Sigset
discard sigemptyset(nmask)
discard sigemptyset(omask)
discard sigaddset(nmask, cint(signal))
blockSignals(nmask, omask)
# to be compatible with linux semantic we need to "eat" signals
posix.signal(cint(signal), SIG_IGN)
modifyKQueue(s, signal.uint, EVFILT_SIGNAL, EV_ADD, 0, 0,
cast[pointer](fdi))
when not declared(CACHE_EVENTS):
flushKQueue(s)
inc(s.count)
result = fdi
proc registerProcess*[T](s: Selector[T], pid: int,
data: T): int {.discardable.} =
let fdi = getUnique(s)
s.checkFd(fdi)
doAssert(s.fds[fdi].ident == InvalidIdent)
var kflags: cushort = EV_ONESHOT or EV_ADD
setKey(s, fdi, {Event.Process, Event.Oneshot}, pid, data)
modifyKQueue(s, pid.uint, EVFILT_PROC, kflags, NOTE_EXIT, 0,
cast[pointer](fdi))
when not declared(CACHE_EVENTS):
flushKQueue(s)
inc(s.count)
result = fdi
proc registerEvent*[T](s: Selector[T], ev: SelectEvent, data: T) =
let fdi = ev.rfd.int
doAssert(s.fds[fdi].ident == InvalidIdent, "Event is already registered in the queue!")
setKey(s, fdi, {Event.User}, 0, data)
modifyKQueue(s, fdi.uint, EVFILT_READ, EV_ADD, 0, 0, nil)
when not declared(CACHE_EVENTS):
flushKQueue(s)
inc(s.count)
template processVnodeEvents(events: set[Event]): cuint =
var rfflags = 0.cuint
if events == {Event.VnodeWrite, Event.VnodeDelete, Event.VnodeExtend,
Event.VnodeAttrib, Event.VnodeLink, Event.VnodeRename,
Event.VnodeRevoke}:
rfflags = NOTE_DELETE or NOTE_WRITE or NOTE_EXTEND or NOTE_ATTRIB or
NOTE_LINK or NOTE_RENAME or NOTE_REVOKE
else:
if Event.VnodeDelete in events: rfflags = rfflags or NOTE_DELETE
if Event.VnodeWrite in events: rfflags = rfflags or NOTE_WRITE
if Event.VnodeExtend in events: rfflags = rfflags or NOTE_EXTEND
if Event.VnodeAttrib in events: rfflags = rfflags or NOTE_ATTRIB
if Event.VnodeLink in events: rfflags = rfflags or NOTE_LINK
if Event.VnodeRename in events: rfflags = rfflags or NOTE_RENAME
if Event.VnodeRevoke in events: rfflags = rfflags or NOTE_REVOKE
rfflags
proc registerVnode*[T](s: Selector[T], fd: cint, events: set[Event], data: T) =
let fdi = fd.int
setKey(s, fdi, {Event.Vnode} + events, 0, data)
var fflags = processVnodeEvents(events)
modifyKQueue(s, fdi.uint, EVFILT_VNODE, EV_ADD or EV_CLEAR, fflags, 0, nil)
when not declared(CACHE_EVENTS):
flushKQueue(s)
inc(s.count)
proc unregister*[T](s: Selector[T], fd: int|SocketHandle) =
let fdi = int(fd)
s.checkFd(fdi)
var pkey = addr(s.fds[fdi])
doAssert(pkey.ident != InvalidIdent,
"Descriptor [" & $fdi & "] is not registered in the queue!")
if pkey.events != {}:
if pkey.events * {Event.Read, Event.Write} != {}:
if Event.Read in pkey.events:
modifyKQueue(s, uint(fdi), EVFILT_READ, EV_DELETE, 0, 0, nil)
dec(s.count)
if Event.Write in pkey.events:
modifyKQueue(s, uint(fdi), EVFILT_WRITE, EV_DELETE, 0, 0, nil)
dec(s.count)
when not declared(CACHE_EVENTS):
flushKQueue(s)
elif Event.Timer in pkey.events:
if Event.Finished notin pkey.events:
modifyKQueue(s, uint(fdi), EVFILT_TIMER, EV_DELETE, 0, 0, nil)
when not declared(CACHE_EVENTS):
flushKQueue(s)
dec(s.count)
if posix.close(cint(pkey.ident)) != 0:
raiseIOSelectorsError(osLastError())
elif Event.Signal in pkey.events:
var nmask, omask: Sigset
let signal = cint(pkey.param)
discard sigemptyset(nmask)
discard sigemptyset(omask)
discard sigaddset(nmask, signal)
unblockSignals(nmask, omask)
posix.signal(signal, SIG_DFL)
modifyKQueue(s, uint(pkey.param), EVFILT_SIGNAL, EV_DELETE, 0, 0, nil)
when not declared(CACHE_EVENTS):
flushKQueue(s)
dec(s.count)
if posix.close(cint(pkey.ident)) != 0:
raiseIOSelectorsError(osLastError())
elif Event.Process in pkey.events:
if Event.Finished notin pkey.events:
modifyKQueue(s, uint(pkey.param), EVFILT_PROC, EV_DELETE, 0, 0, nil)
when not declared(CACHE_EVENTS):
flushKQueue(s)
dec(s.count)
if posix.close(cint(pkey.ident)) != 0:
raiseIOSelectorsError(osLastError())
elif Event.Vnode in pkey.events:
modifyKQueue(s, uint(fdi), EVFILT_VNODE, EV_DELETE, 0, 0, nil)
when not declared(CACHE_EVENTS):
flushKQueue(s)
dec(s.count)
elif Event.User in pkey.events:
modifyKQueue(s, uint(fdi), EVFILT_READ, EV_DELETE, 0, 0, nil)
when not declared(CACHE_EVENTS):
flushKQueue(s)
dec(s.count)
clearKey(pkey)
proc unregister*[T](s: Selector[T], ev: SelectEvent) =
let fdi = int(ev.rfd)
s.checkFd(fdi)
var pkey = addr(s.fds[fdi])
doAssert(pkey.ident != InvalidIdent, "Event is not registered in the queue!")
doAssert(Event.User in pkey.events)
modifyKQueue(s, uint(fdi), EVFILT_READ, EV_DELETE, 0, 0, nil)
when not declared(CACHE_EVENTS):
flushKQueue(s)
clearKey(pkey)
dec(s.count)
proc selectInto*[T](s: Selector[T], timeout: int,
results: var openArray[ReadyKey]): int =
var
tv: Timespec
resTable: array[MAX_KQUEUE_EVENTS, KEvent]
ptv = addr tv
maxres = MAX_KQUEUE_EVENTS
verifySelectParams(timeout)
if timeout != -1:
if timeout >= 1000:
tv.tv_sec = posix.Time(timeout div 1_000)
tv.tv_nsec = (timeout %% 1_000) * 1_000_000
else:
tv.tv_sec = posix.Time(0)
tv.tv_nsec = timeout * 1_000_000
else:
ptv = nil
if maxres > len(results):
maxres = len(results)
var count = 0
when not declared(CACHE_EVENTS):
count = kevent(s.kqFD, nil, cint(0), addr(resTable[0]), cint(maxres), ptv)
else:
when hasThreadSupport:
s.withChangeLock():
if s.changesLength > 0:
count = kevent(s.kqFD, addr(s.changes[0]), cint(s.changesLength),
addr(resTable[0]), cint(maxres), ptv)
s.changesLength = 0
else:
count = kevent(s.kqFD, nil, cint(0), addr(resTable[0]), cint(maxres),
ptv)
else:
let length = cint(len(s.changes))
if length > 0:
count = kevent(s.kqFD, addr(s.changes[0]), length,
addr(resTable[0]), cint(maxres), ptv)
s.changes.setLen(0)
else:
count = kevent(s.kqFD, nil, cint(0), addr(resTable[0]), cint(maxres),
ptv)
if count < 0:
result = 0
let err = osLastError()
if cint(err) != EINTR:
raiseIOSelectorsError(err)
elif count == 0:
result = 0
else:
var i = 0
var k = 0 # do not delete this, because `continue` used in cycle.
var pkey: ptr SelectorKey[T]
while i < count:
let kevent = addr(resTable[i])
var rkey = ReadyKey(fd: int(kevent.ident), events: {})
if (kevent.flags and EV_ERROR) != 0:
rkey.events = {Event.Error}
rkey.errorCode = OSErrorCode(kevent.data)
case kevent.filter:
of EVFILT_READ:
pkey = addr(s.fds[int(kevent.ident)])
rkey.events.incl(Event.Read)
if Event.User in pkey.events:
var data: uint64 = 0
if posix.read(cint(kevent.ident), addr data,
sizeof(uint64)) != sizeof(uint64):
let err = osLastError()
if err == OSErrorCode(EAGAIN):
# someone already consumed event data
inc(i)
continue
else:
raiseIOSelectorsError(err)
rkey.events = {Event.User}
of EVFILT_WRITE:
pkey = addr(s.fds[int(kevent.ident)])
rkey.events.incl(Event.Write)
rkey.events = {Event.Write}
of EVFILT_TIMER:
pkey = addr(s.fds[int(kevent.ident)])
if Event.Oneshot in pkey.events:
# we will not clear key until it will be unregistered, so
# application can obtain data, but we will decrease counter,
# because kqueue is empty.
dec(s.count)
# we are marking key with `Finished` event, to avoid double decrease.
pkey.events.incl(Event.Finished)
rkey.events.incl(Event.Timer)
of EVFILT_VNODE:
pkey = addr(s.fds[int(kevent.ident)])
rkey.events.incl(Event.Vnode)
if (kevent.fflags and NOTE_DELETE) != 0:
rkey.events.incl(Event.VnodeDelete)
if (kevent.fflags and NOTE_WRITE) != 0:
rkey.events.incl(Event.VnodeWrite)
if (kevent.fflags and NOTE_EXTEND) != 0:
rkey.events.incl(Event.VnodeExtend)
if (kevent.fflags and NOTE_ATTRIB) != 0:
rkey.events.incl(Event.VnodeAttrib)
if (kevent.fflags and NOTE_LINK) != 0:
rkey.events.incl(Event.VnodeLink)
if (kevent.fflags and NOTE_RENAME) != 0:
rkey.events.incl(Event.VnodeRename)
if (kevent.fflags and NOTE_REVOKE) != 0:
rkey.events.incl(Event.VnodeRevoke)
of EVFILT_SIGNAL:
pkey = addr(s.fds[cast[int](kevent.udata)])
rkey.fd = cast[int](kevent.udata)
rkey.events.incl(Event.Signal)
of EVFILT_PROC:
rkey.fd = cast[int](kevent.udata)
pkey = addr(s.fds[cast[int](kevent.udata)])
# we will not clear key, until it will be unregistered, so
# application can obtain data, but we will decrease counter,
# because kqueue is empty.
dec(s.count)
# we are marking key with `Finished` event, to avoid double decrease.
pkey.events.incl(Event.Finished)
rkey.events.incl(Event.Process)
else:
doAssert(true, "Unsupported kqueue filter in the queue!")
if (kevent.flags and EV_EOF) != 0:
# TODO this error handling needs to be rethought.
# `fflags` can sometimes be `0x80000000` and thus we use 'cast'
# here:
if kevent.fflags != 0:
rkey.errorCode = cast[OSErrorCode](kevent.fflags)
else:
# This assumes we are dealing with sockets.
# TODO: For future-proofing it might be a good idea to give the
# user access to the raw `kevent`.
rkey.errorCode = OSErrorCode(ECONNRESET)
rkey.events.incl(Event.Error)
results[k] = rkey
inc(k)
inc(i)
result = k
proc select*[T](s: Selector[T], timeout: int): seq[ReadyKey] =
result = newSeq[ReadyKey](MAX_KQUEUE_EVENTS)
let count = selectInto(s, timeout, result)
result.setLen(count)
template isEmpty*[T](s: Selector[T]): bool =
(s.count == 0)
proc contains*[T](s: Selector[T], fd: SocketHandle|int): bool {.inline.} =
let fdi = fd.int
fdi < s.maxFD and s.fds[fd.int].ident != InvalidIdent
proc setData*[T](s: Selector[T], fd: SocketHandle|int, data: T): bool =
let fdi = int(fd)
if fdi in s:
s.fds[fdi].data = data
result = true
template withData*[T](s: Selector[T], fd: SocketHandle|int, value,
body: untyped) =
let fdi = int(fd)
if fdi in s:
var value = addr(s.fds[fdi].data)
body
template withData*[T](s: Selector[T], fd: SocketHandle|int, value, body1,
body2: untyped) =
let fdi = int(fd)
if fdi in s:
var value = addr(s.fds[fdi].data)
body1
else:
body2
proc getFd*[T](s: Selector[T]): int =
return s.kqFD.int