Using non blocking wait

[kolia]

Hello,

I have a handler "updateStatus" that runs every 10 seconds. At the very beginning of this handler we have

Code: Select all

setUpdateIsRunning true

and at the end

Code: Select all

setUpdateIsRunning false

User may click on any interface control at any time. At the beginning of the control handler, we have to make sure that updateStatus is not running in order to avoid concurrent calls to a hardware device. I have this instruction :

Code: Select all

   wait while getUpdateStatusRunning() is true with messages 

Can anyone tell me that it does what I'm expecting, that is the control handler will resume execution when the updateStatus handler hits the end, in order to avoid possible concurrent calls?
Thank you

Just for reference the code of the two routines

Code: Select all

local updateStatusRunning
command setUpdateStatusRunning pValue
   put pValue into updateStatusRunning
end setUpdateStatusRunning

function getUpdateStatusRunning
   if updateStatusRunning is empty then put false into updateStatusRunning 
   return updateStatusRunning
end getUpdateStatusRunning

[dunbarx]

Hi.

It is hard to comment, because, for example, you do not say what or where the handler "setUpdateIsRunning" is, or what it does. If I tried to reconstruct the handful of code smippets you mentioned, I would get a runtime error. There is no such handler, and I cannot go any further.

Perhaps come back with complete handlers? I hope they are not too long. I generally do not read long handler submissions, unless I can extract the salient portions or rewrite them into shorter examples.

Craig Newman

[kolia]

Hi Craig,
Thanks for getting involved in my question. I'll try to be as concise as possible.

updateStatus runs every 10 sec. At the beginning of the routine it sets a flag (setUpdateStatusRunning true) and at the end it sets the same flag to false (setUpdateStatusRunning false)

Whenever a user click, at the beginning of the handler the status of the flag is checked and the handler is paused (non blocking) until the flag is set to false by updateStatus. The instruction at the beginning this handler is:

Code: Select all

wait while getUpdateStatusRunning() is true with messages.

Both handlers, updateStatus and onMouseUp, access a remote hardware device and I wan't to avoid concurrent accesses. So I wan't to make sure that "wait while getUpdateStatusRunning() is true with messages" will pause onMouseUp until updateStatus reaches the end.

Nicolas

[WaltBrown]

What is the difference between "setUpdateStatusRunning" and "setUpdateIsRunning"? You gave the code for the former but not the latter. Typo?
Walt

[zaxos]

Code: Select all

wait while getUpdateStatusRunning() is true with messages.

No need for that, instead:
e.x. Code of button 1

Code: Select all

global isRunning
on mouseUp
if isRunning then
doSomething
else
-- do nothing
end if
end mouseUp

Code in stack

Code: Select all

global isRunning
on updateStatus
put true into isRunning
do stuff
do stuff
do stuff
do stuff
put false into isRunning
send isRunning to me in 10 sec
end updateStatus

You can use custom properties instead of a global, you can actualy have all the scripts in the card, but you get the big picture.

[kolia]

What is the difference between "setUpdateStatusRunning" and "setUpdateIsRunning"? You gave the code for the former but not the latter. Typo?
Walt

Yes sorry typo error the routine is setUpdateStatusRunning

[kolia]

Thanks Xanos for your contribution. But this is not the idea. The idea is to have the onMouseUp handler resuming execution when the updateStatus reaches the end.

[bn]

Hi Nicolas,

as far as I see when you run your "update" handler there is no way for another message to get processed since you don't e.g. wait with messages in that handler.

In theory you should not even have to check if the update handler runs because once it runs it blocks all other requests. Once the update handler is finished mouseUp should fire because it should be queued.

Unless your "update" handler is non-blocking

I made a little stack that simulates both situations.

On top 3 buttons. Click "simulate non blocking update" (runs 5 seconds) then "my MouseUp" repeatedly, in the field you will see the status of the simulate handler with the milliseconds attached to see something is going on.
myMouseUp2 button has a handler that checks the status of the non-blocking button and only proceeds when the non-blocking buttonn is done.

at the bottom 2 buttons for a blocking update handler
click "simulate blocking update" button (the button just waits for 5 seconds) and then click the myMouseUP3 button. The button does not react but the action is queued. Once the blocking handler is done you see the milliseconds change in the field.

By the way when checking the non-blocking handler with button myMouseUp2 I first tried to use "wait 1 second with messages" in a repeat loop: That did not work, apparently the non-blocking handler, which also waited with messages then never ended waiting. I had to use "send in time" to do this.


Kind regards
Bernd

[kolia]

Hi Bernd
I really appreciate your valuable input. It explains another behavior I noticed without being able to understand it. This makes very clear how LC handles interruption if I may say so. I will check the stack you attached as soon as I can and I will post what I understood.
Thanks again
Nicolas

[FourthWorld]

Only peripherally related, but this modest demonstration stack shows a couple ways to simulate some limited forms of threading in LC:
http://fourthworld.net/channels/lc/IdleHour.rev

[kolia]

Thanks Richard, I'll look at this one as well