To programmatically set and control typeahead mode you use
the FDEVICECONTROL intrinsic, specifying one of three control directives
in the parm1 value.
A control directive of 51 is used
to specify whether or not typeahead mode is on for the device. Use
this call with the buffer parameter set to
1 to enable typeahead mode, or with the buffer
parameter set to 0 to disable typeahead mode (parm2
set to 2 or 3).
A control directive of 60 allows you to discard
any data in the typeahead buffer and read directly from the device.
A control directive of 61 allows you to bypass the
typeahead buffer for the next read while saving the data already
in the typeahead buffer.
A control directive of 63 is used to enable or disable
single echo typeahead. Use this call with the buffer
parameter set to 1 to enable typeahead mode, or with the buffer
parameter set to 0 to disable single echo typeahead mode (parm2
set to 2 or 3).
The code fragments that follow illustrate how these calls
might look in your program.
Setting Typeahead Mode |
 |
The code fragment shown in Figure 6-1 “Program Fragment Setting Typeahead
Mode” illustrates how you can use the FDEVICECONTROL
intrinsic to allow typeahead processing for your program. A control
directive of 51, set typeahead mode, is passed in parm1,
and parm2 is set to 2 to specify Write access.
The buffer parameter contains a variable
called value, set to 1 in this example,
which will cause typeahead mode to be enabled. (A value of 0 would
disable typeahead mode.)
In this example, the Pascal/iX ccode
function checks for errors that might occur during execution of
the intrinsic. If an error occurs, the program reports the file
system error number associated with the error. If no errors occur,
the program informs the user that typeahead is now on.
Figure 6-1 Program Fragment Setting Typeahead
Mode
Flushing the Typeahead Buffer |
|
Figure 6-2 “Program Fragment to Flush Buffer” illustrates how
FDEVICECONTROL can be used to flush the typeahead buffer. For example,
the buffer can be flushed if an error is detected in the data that
was typed ahead. This FDEVICECONTROL is only valid if type ahead
is enabled. The default is not to flush the typeahead buffer.
An FDEVICECONTROL call with a control directive of 60 will
cause the buffer to be flushed at the time of the next read. The
read can then be satisfied by data entered directly from the device.
The typeahead buffer is flushed for a single read only; normal typeahead
mode processing will resume for any following reads.
In the example, the parm1 value is
set to 60, flush typeahead buffer, and the buffer
parameter (value) passes a value of 1.
This specifies that all data currently in the typeahead buffer should
be discarded, and the next read should access the device directly.
Write access is specified in parm2, and the
program checks for errors using the Pascal/iX ccode
function.
Figure 6-2 Program Fragment to Flush Buffer
Bypassing the Typeahead Buffer |
|
Figure 6-3 “Program Fragment to Bypass Buffer” is an example
of using FDEVICECONTROL to bypass the typeahead buffer. You could
use this capability to send a status request and obtain a reply
directly from a terminal without affecting the data already in the
typeahead buffer. This FDEVICECONTROL is only valid if typeahead
is enabled. The default is not to bypass the typeahead buffer.
This FDEVICECONTROL call will cause the typeahead buffer to
be bypassed for a single read only; normal typeahead mode processing
will resume for any following reads, and data will be taken from
the typeahead buffer if data is present.
When you use this request, you call FDEVICECONTROL with parm1
set to 61, bypass typeahead mode, and pass a value of 1 in the buffer
parameter (value :=1). As in the
previous examples, Write access is specified in parm2,
and the Pascal/iX ccode function is used
for error checking.
|
| |
|
 | NOTE: In cases where there is no read trigger (read trigger
set to null) bypassing the typeahead buffer may have unexpected
results. If, for example, your program is swapped out between the
FWRITE and FREAD of a terminal status request, the terminal may
appear to become "hung", since there is no read
trigger to synchronize terminal response with the posting of the
read. To avoid this problem, instead of bypassing the typeahead
buffer to send a status request when no read trigger is present,
the following should be the sequence that occurs: = Disable typeahead mode = Post a 0 byte read = Enable typeahead mode = (Lock keyboard) (FWRITE) Send status request = FREAD of status = FWRITE (unlock keyboard)
|
|
| |
|
Figure 6-3 Program Fragment to Bypass Buffer
