Fix spelling in comments

Should be non-functional, except for some error message strings.
This commit is contained in:
Michael Davidsaver
2021-08-15 09:54:21 -07:00
parent 42d06d6a38
commit e34b6c5c0c
186 changed files with 419 additions and 419 deletions

View File

@@ -27,17 +27,17 @@
// 1) Allow sufficent headroom so that users will be able to perform
// a reasonable amount of IO within CA callbacks without experiencing
// a push/pull deadlock. If a potential push/pull deadlock situation
// occurs then detect and avoid it and provide diagnotic to the user
// occurs then detect and avoid it and provide diagnostic to the user
// via special status.
// 2) Return status to the user when there is insufficent memory to
// 2) Return status to the user when there is insufficient memory to
// queue a complete message.
// 3) return status to the user when a message cant be flushed because
// a connection dropped.
// 4) Do not allocate too much memory in exception situatons (such as
// 4) Do not allocate too much memory in exception situations (such as
// after a circuit disconnect).
// 5) Avoid allocating more memory than is absolutely necessary to meet
// the above requirements.
// 6) Message fragments must never be sent to the IOC when there isnt
// 6) Message fragments must never be sent to the IOC when there isn't
// enough memory to queue part of a message (we also must not force
// a disconnect because the client is starved for memory).
// 7) avoid the need to check status for each byte pushed into the
@@ -45,7 +45,7 @@
//
// Implementation:
// 1) When queuing a complete message, first test to see if a flush is
// required. If it is a receive thread scheduals the flush with the
// required. If it is a receive thread schedules the flush with the
// send thread, and otherwise directly execute the system call. The
// send thread must run at a higher priority than the receive thread
// if we are to minimize memory consumption.
@@ -58,9 +58,9 @@
// a) A user is queuing more requests that demand a response from a
// callback than are removed by the response that initiated the
// callback, and this situation persists for many callbacks until
// all buffering in the system is exausted.
// all buffering in the system is exhausted.
// b) A user is queuing many requests that demand a response from one
// callback until all buffering in the system is exausted.
// callback until all buffering in the system is exhausted.
// c) Some combination of both (a) nad (b).
//
//