if not semID := locknew ??

Roger Lee

RE: "The official Guide" Programming and Customizing...

Timestamp Dev (ASM) example on page 113

I'm working through this example for PASD debugger and came to this line

if not semID := locknew

1) It's a conditional
2) negative logic ?
3) assignment operation ?
4) check out a lock

What is this line doing?

PUB TestStart

  m   := 10
  s   := 59
  ms  := 999
  dt  := clkfreq/1000

  if not semID := locknew       ' check out a lock
    cognew(@entry, @m)

'{'

  dbg.start(31, 30, @entry)     ' <--- Add for debugger
  
'}'

  repeat                        ' keep the cog running

kuroneko

Roger Lee said...
What is this line doing?

Probably not quite what was intended? It boils down to calling cognew only when the checked out lock ID is 0, for failure (-1) or lock IDs 1..7 it does nothing (but checks out a lock regardless).

Roger Lee

Hmmm, very interesting.

Thanks.

kuroneko

Roger Lee said...
Hmmm, very interesting.

I don't have the book. What's the context, i.e. maybe it's supposed to work like this? Then again the consumption of locks seems a bit suspicious.

I found something like this in the TimeStamp Object but judging by the associated comment that's completely broken.

Post Edited (kuroneko) : 5/1/2010 2:36:41 AM GMT

Mike Green

You have to think of ":=" as an operator (which it is in Spin). The assignment gets done first, then the unary operator "not" gets applied, then the "if" test is done on the final result.

Roger Lee

Thank you Mike,
makes sense when you say it like that. Just what I needed.

So, then the result is:

check to see if a lock is available before cognew(@entry, @m)

Phil Pilgrim (PhiPi)

kuroneko said...
Probably not quite what was intended? It boils down to calling cognew only when the checked out lock ID is 0, for failure (-1) or lock IDs 1..7 it does nothing (but checks out a lock regardless).

The programmer probably meant if ! semID := locknew instead, which is true only on a successful lock acquisition.

-Phil