I spotted a few minor typos in the TPL Dataflow doc.
Dataflow.chm
=================================
IDataflowBlock.Complete
Change:
but rather will initiaite the request
To:
but rather will initiate the request
-----------------
IDataflowBlock.Fault
Change:
uprocessed input messages as well as unoffered output messages
To:
unprocessed input messages as well as unoffered output messages
---------------------
DataflowBlock.OutputAvailableAsync
Change:
if it'sResult is true
To:
if its Result is true
----------------------------------
DataflowBlock.SendAsync
Change:
with itsResultindicating whether
To:
with its Result indicating whether
---------------------
ISourceBlock.ConsumeMessage
Change (reserve for consume):
Only ITargetBlock`1 instances linked to this ISourceBlock`1 instance may
use ConsumeMessage, and it must only be used to reserve DataflowMessageHeader
instances previously offered by this source to the target.
To:
Only ITargetBlock`1 instances linked to this ISourceBlock`1 instance may
use ConsumeMessage, and it must only be used to consume DataflowMessageHeader
instances previously offered by this source to the target.
-------------------
DataflowBlock.Choose
Some overloads monitor two sources, while others monitor 3, yet all say the
"two", so perhaps all should be changed, as follows:
Change:
Monitors two dataflow sources
To:
Monitors multiple dataflow sources
===============================
TPL Dataflow Introduction
-------------------------
Change:
utilize the DataflowBlockExtensions’ Encapsulate method to
To:
utilize the DataflowBlock’s Encapsulate method to
Change:
return DataflowBlockExtensions.Encapsulate(target, source);
To:
return DataflowBlock.Encapsulate(target, source);
Change:
(such as those defined on DataflowBlockExtensions)
To:
(such as those defined on DataflowBlock extension methods)
-----------------------
WriteOnceBLock's has no default constructor, so it requires a parameter.
Change:
WriteOnceBlock()
To:
WriteOnceBlock(null)
(This needs to be done for two of the four WriteOnceBlock samples.
For the other two of the four WriteOnceBlock samples, i=>i is passed in as
the parameter. I wonder would it be better to pass in null here too, on
efficiency grounds. It avoids the construction of the cached anonymous delegate
from the lambda. It is not important when dealing with 10 messages, but would be
important when dealing with a million.
------------------------------------------
minor typos in the TPL Dataflow doc
