Update TODO items within the code [skip ci]

weave/channels/event_notifiers.nim

@@ -63,7 +63,6 @@ const supportsFutex = defined(linux) or defined(windows)
 # that do not happen on MacOS.
 # Instead we use raw futex to replace both locks and condition variables.
 # They uses 20x less space and do not add useless logic for our use case.
-# TODO: Windows futex. Unfortunately BSDs including Mac do not provide a futex.
 
 type
   EventNotifier* = object

weave/datatypes/prell_deques.nim

@@ -19,15 +19,6 @@ type
     # task has a "fn" field with the proc to run
     task.fn is proc (param: pointer) {.nimcall.}
 
-    # TODO: checkout the liftLocal macro
-    #       to reimplement closures and allow
-    #
-    # var myArray: ptr UncheckedArray[int]
-    # parallel_loop(i, 0, 100000):
-    #   myArray[i] = i
-    #
-    # with "myArray" implicitly captured.
-
   PrellDeque*[T: StealableTask] = object
     ## Private Intrusive Work-Stealing Deque
     ## from PhD Thesis

weave/datatypes/sync_types.nim

@@ -54,12 +54,9 @@ type
   # Steal requests
   # ----------------------------------------------------------------------------------
 
-  # Padding shouldn't be needed as steal requests are used as value types
-  # and deep-copied between threads
   StealRequest* = ptr object
-    # TODO: padding to cache line
     # TODO: Remove workaround generic atomics bug: https://github.com/nim-lang/Nim/issues/12695
-    next*: Atomic[pointer]                        # For intrusive lists and queues
+    next*{.align:WV_CacheLinePadding.}: Atomic[pointer]                        # For intrusive lists and queues
     thiefAddr*: ptr ChannelSpscSinglePtr[Task]    # Channel for sending tasks back to the thief
     thiefID*: WorkerID
     retry*: int32                                 # 0 <= retry <= num_workers

weave/memory/memory_access_patterns.md

@@ -187,10 +187,7 @@ In that case, they should have padding so that within the array, fields of 2 con
 #### Channels for steal requests
 
 Multi-Producer Single-Consumer channels with a dynamic but bounded capacity:
-  - "MaxSteal * number_of_workers" for a spawned thread
-  - "MaxSteal * number_of_workers * 2" for the master thread
-    (TODO: the 2x is/was needed when workers sent their state to a manager
-           thread for termination detection but a Djikstra toen-passing algorithm as been replaced by a dedicated tree algorithm by Dinan et al)
+  - "MaxSteal * number_of_workers"
 
 They serve as a mailbox for steal requests and are stored in a global array `array[MaxWorkers, Channel[StealRequest]]`.
 

weave/memory/memory_pools.nim

@@ -493,8 +493,7 @@ proc recycle*[T](p: ptr T) {.gcsafe.} =
   ## block was returned, the main thread should now
   ## have ownership of the related arenas and can deallocate them.
 
-  # TODO: sink ptr T - parsing bug to raise
-  #   similar to https://github.com/nim-lang/Nim/issues/12091
+  # TODO: sink ptr T - parsing bug https://github.com/nim-lang/Nim/issues/12757
   preCondition: not p.isNil
 
   let p = cast[ptr MemBlock](p)

weave/primitives/barriers_posix.nim

@@ -5,8 +5,7 @@
 #   * Apache v2 license (license terms in the root directory or at http://www.apache.org/licenses/LICENSE-2.0).
 # at your option. This file may not be copied, modified, or distributed except according to those terms.
 
-# Abstractions over OS-specific barrier implementations
-# TODO: support Windows (check iOS, Android, Genode, Haiku, Nintendo Switch)
+# Abstractions over POSIX barriers (non-)implementations
 
 when not compileOption("threads"):
   {.error: "This requires --threads:on compilation flag".}
@@ -49,4 +48,4 @@ else:
     ## Wait on `barrier`
     ## Returns PTHREAD_BARRIER_SERIAL_THREAD for a single arbitrary thread
     ## Returns 0 for the other
-    ## Returns Errno if there is an error
+    ## Returns Errno if there is an error

weave/scheduler.nim

@@ -291,5 +291,3 @@ proc schedule*(task: sink Task) =
     dispatchElseDecline(req)
 
   profile_start(enq_deq_task)
-
-  ## TODO steal early