A LB quirk?

[tenochtitlanuk]

Someone posting on Rosetta Code with a LB solution included the following line..

rd=gn=bl=int(rnd(1)*255)

Multiple attributions are available in some languages, written this way. Not in LB.

What do you think LB makes of it...??

[tsh73]

I pretty sure it does not assign anything but outer variable
Inner used in comparison, resulting in 0 (false) or 1 (true)
But in what order?
This test

x=1:y=1:z=1 '1'1'1
x=1:y=2:z=2 '0'1'0
x=2:y=2:z=1 '1'0'1
print
print (x=y)=z
print x=(y=z)
print x=y=z

suggests that
x=y=z
evaluates as
(x=y)=z

So we get
rd=((gn=bl)=int(rnd(1)*255))
(gn=bl) is 0 or 1
int(rnd(1)*255)) is 0 or 1 with probability 1/256

So rd is 1 with probability 1/256, else 0
gn,bl is not changed, and is not affect random result (! ha ! nice obsfucation)

Let's test it

N=1000000
for i = 1 to N
rd=gn=bl=int(rnd(1)*255)
if rd then s=s+1
'print i, rd
next
print s, N/s

I got on a test run
4047          247.096615
- pretty close I think.
(other runs sure will differ)

[tenochtitlanuk]

Yup. I did a similar test-
ifor i =1 to 1024
  rd=gn=bl=int(rnd(1)*255)
  if rd =0 then print "0"; else print: print rd
next