Properly handling BREAK within CIO handler?

[tschak909]

Am curious as to how I am supposed to handle processing the BREAK key in the CIO handler, am I supposed to (at any potentially wait-y spot):

 

* Check BREAK key via $11

* If set, set Y to $80 and jump out?

 

-Thom

[Rybags]

Depends on what the device does and whether you can be subject to hangs or unwanted long timeouts.

 

For a device doing external IO you'd probably want to be able to abort part way through a record read or write rather than waiting on a potentially long timeout.

S: checks it on each iteration e.g. per pixel when doing a line or fill, which is sort of overkill although a fill operation could be something that takes some time.

(though that said, the nature of internal handlers is that the PUT code doesn't necessarily know if it's doing a single operation only or one that's part of thousands)

 

Not sure if the OS portion of CIO checks it - might be worth looking at some source listings there.

 

Also, further down the line you have retry logic - a good example there is with DOS - pressing Break can sometimes interrupt an operation but often just causes a retry and multiple presses are needed to abort.

Edited March 21, 2021 by Rybags

[thorfdbg]

On 3/21/2021 at 7:39 AM, Rybags said:

Not sure if the OS portion of CIO checks it - might be worth looking at some source listings there.

No, CIO itself does not check the break flag. That is up to the handler to do.