[Info-vax] QIO Writes to a mailbox

[Michael Moroney]

JF Mezei <jfmezei.spamnot at vaxination.ca> writes:

>Michael Moroney wrote:

>> Not true, unless IO$M_NOW was specified on the write.  Otherwise, the 
>> write QIO(W) won't complete until another process issues a read, and that
>> could be a *significant* delay, if the reader was delayed.
>> 


>Are you sure of that for the mailbox driver ? You can do multiple writes
>to a mailbox (even with QIOW which I what I have used for writes) before
>the reading process collects the messages. This is why you declare a
>buffer size when you create a mailbox.

It works as you state if and only if you specify the IO$M_NOW modifier.

Without it, writes will wait until another process issues a read for the
same data.

Remember, the mailbox IOSB specifies the PID of the writer process (for 
reads) and the PID of the reader (for writes).  Not possible for both
to happen unless everything happens at once.

There is also a stream modifier that may act similarly, I am not familiar
with it.