views:

461

answers:

3

Why the peekMessage statement is required before Getmessage() for creating message queue?

+13  A: 

It's not required.

What you'll sometimes see, though, is a thread that isn't ready to process messages yet, but it wants to be able to receive them in its message queue. New threads do not have message queues right away, but calling PeekMessage is sufficient to create the message queue. It returns immediately since there is no message, and that allows the thread to continue getting itself ready. In the meantime, other threads can start queuing messages for the new thread. Once the new thread is ready, it calls GetMessage to either retrieve the first message off the queue, or to wait for a message to be put onto the queue.

Rob Kennedy
Thanks, it answers my question.
Josh
@Josh: Welcome to StackOverflow! If this is a good answer to your question, click the checkmark under the voting widget to mark it as the accepted answer. That way, the rest of us know that this question has been answered satisfactorily.
Mason Wheeler
+14  A: 

It's not. The two functions do different things.

PeekMessage(...) doesn't wait for a message to appear -- it gets the first one if it's there, optionally removing it from the queue as well, but returns false immediately there isn't one. It's more common in apps where you're doing some processing while waiting for messages, and can't just sit there and wait forever for the next message. Real-time games and such easily fall into this category.

GetMessage(...) waits til there's a message, and gets it. It's more efficient CPUwise, cause it's not constantly polling, but it will pause if there aren't any messages. It's more common in formy apps and other programs that don't require constant real-time processing to be going on.

cHao
Here's another reference, which notes that it's not necessarily the wisest way of doing background work (it seems Delphi has threads): http://windows-programming.suite101.com/article.cfm/win32_message_processing_primer
Hostile Fork
+1  A: 

There are multiple reasons for using PeekMessage before/instead of GetMessage:

  1. Ensuring the program will not hang until a message arrives - that's a bit redundant, cause you can directly use PeekMessage with the PM_REMOVE flag to poll the message queue and leave out GetMessage altogether.
  2. Using the function with PM_NOREMOVE and deciding if you want to process and/or remove the message from the queue, or not.
  3. Calling IsWindowUnicode on the returned messages' window handle and selecting either PeekMessageA or PeekMessageW.
  4. Multiple of the above.
Viktor Svub