Testing a variable inside a for loop

[wizard]

The objective is to end the for loop when the variable done = 1, which should happen after 10 seconds, but instead the loop continues, not sure why.

Thanks
wizard

Code: Select all

#! /bin/bash
done=0

for ((i=1; i<=15; i++)) {
    if [[ $done == 1 ]];then
    break
    fi
    echo $i
    sleep 1
} &

sleep 10 
done=1
echo $done
echo done=1

Output:

forloop-test.jpg (12.77 KiB) Viewed 3096 times

[AntonioPt]

@wizard you forget to increment done

((done++))

Code: Select all

#! /bin/bash
done=0

for ((i=1; i<=15; i++)) {
    if [[ $done == 1 ]];then
        done=1 # set with one or $i
        break
    fi

    echo $i
    sleep 1
    ((done++)) 
} &

sleep 10 

echo $done
echo done=1

The objective is to end the for loop when the variable done = 1, which should happen after 10 seconds, but instead the loop continues, not sure why.

Thanks
wizard

Code: Select all

#! /bin/bash
done=0

for ((i=1; i<=15; i++)) {
    if [[ $done == 1 ]];then
    break
    fi
    echo $i
    sleep 1
} &

sleep 10 
done=1
echo $done
echo done=1

Output:
forloop-test.jpg

[wizard]

@AntonioPt

@wizard you forget to increment done

Actually, done is incremented here:

Code: Select all

sleep 10 
done=1

Your code exits the loop in 1 second.

Thanks
wizard

[AntonioPt]

But @wizard its outside of the loop lol so will never works like you could see

[wizard]

@AntonioPt

But @wizard its outside of the loop lol so will never works like you could see

So if the for loop can't read variables outside the loop, then why is this able to stop the loop?

Thanks
wizard

Code: Select all

#! /bin/bash

done=1

for ((i=1; i<=15; i++)) {
    if [[ $done == 1 ]];then
    break
    fi
    echo $i
    sleep 1
    
}

[AntonioPt]

for your example right you wanna test done and it starts with 0 in order to check if done become 1 or you increment done inside the loop or you add done=1 all so inside the loop if done=1 is outside the loop will do what you are asking that is loop 16 times because there is no increment inside loop

Code: Select all

done=0 # starting point

for ((i=1; i<=15; i++)) {
    if [[ $done == 1 ]];then
     break
    fi
    echo $i
    sleep 1
   # done=1 or ((done++))
   # if you add 1
   done=1 # it will stop loop but if you add a different number in the condition will loop 16 times again 
   ((done++)) # and if you add increment it will stop when condition match (any number from 0 till 15 ) get it ? 
}

[AntonioPt]

but @wizard you can do some better

Code: Select all

start=0  # starting point 
stop=2  # stop loop when it reach 2

# lets loop 16 times
for ((i=0; i<=15; i++)) {
	# let do a echo to see our final result
	echo " i = $i "
	echo " start = $start "
	echo " stop = $stop "
	# now lets check if  start match with stop
	if [ $start == $stop ] ; then 
		break # lets stop loop
	fi
	# lets increment start
	((start++))
}

output would be
i = 0
start = 0
stop = 2
i = 1
start = 1
stop = 2
i = 2
start = 2
stop = 2

[AntonioPt]

sorry @wizard just realize your real question now

so in your first code because you add --> } & <-- when it reach 10 it stops and reads your done=1 and im gonna guess has to do with & and im gonna guess bash will read 10 as 1 and 0 separate and this is why it stops and then continues loop again

but anyways in loops & never should be there and done=1 or ((done++)) it was outside loop there for will never do what you wanna do at the end

[rcrsn51]

I think that you want something like this, using PIDs.

Code: Select all

for ((i=1; i<=15; i++)) { echo $i; sleep 1; } &
PID=$!
sleep 10
kill $PID

[Trapster]

It seems to work after removing the "&"

Code: Select all

done=0
for ((i=1; i<=15; i++)) {
    if [[ $done == 1 ]];then
    break
    fi
    echo $i
    sleep 1
}

sleep 10 
done=1
echo $done
echo done=1

Code: Select all

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
1
done=1
/mnt/sda2$ 

EDIT: copy missed the "done=0" at the start of script

[rcrsn51]

It seems to work after removing the "&"

Really?

which should happen after 10 seconds, but instead the loop continues

[Trapster]

See edit to my post

[Chelsea80]

@wizard

Bearing in mind I'm only starting out with trying to learn bash script -

So I wonder if this could be adapted to what you want to achieve -

#!/bin/sh

# Define a variable (this could be set dynamically)
my_var="Hello"

# Start a loop (for example, loop through numbers 1 to 5)
for i in {1..5}
do
# Test if the variable is equal to "Hello"
if [ "$my_var" = "Hello" ]; then
echo "The variable my_var is 'Hello' on iteration $i."
else
echo "The variable my_var is not 'Hello' on iteration $i."
fi

# You can change the variable in each loop iteration if needed
my_var="Goodbye"
done

[rcrsn51]

@wizard Are you still interested in understanding why the "done" method fails?

[wizard]

@rcrsn51

Are you still interested in understanding why the "done" method fails?

Yes

Do like your code

Code: Select all

for ((i=1; i<=15; i++)) { echo $i; sleep 1; } &
PID=$!
sleep 10
kill $PID

My original code attempts to stop the loop with the if statement inside the loop, yours just kills the process which stops the loop.

Many of the replie above don't follow the intent of the original code which should have been explained better. Pseudo code for it would be something like:
-Start a timer
-Run a process
-Stop the timer when the process is complete

Have not found a good answer for these two cases.

In this case the variable "done" is defined and read from outside the loop and stops the loop.

Code: Select all

#! /bin/bash

done=1

for ((i=1; i<=15; i++)) {
    if [[ $done == 1 ]];then
    break
    fi
    echo $i
    sleep 1
    
}

Whereas, if the variable is incremeted after the loop is running it isn't read and does not work.

Code: Select all

#! /bin/bash
done=0

for ((i=1; i<=15; i++)) {
    if [[ $done == 1 ]];then
    break
    fi
    echo $i
    sleep 1
} &

sleep 10 
done=1
echo $done
echo done=1

Thanks
wizard

[rcrsn51]

You are assuming that the "done" variable acts like a C global variable. You want it to be visible to the background process started by the &. So when the foreground process sets "done=1", the loop in the background would detect this and terminate.

But bash variables don't work that way.

One work-around would be for the foreground to write the value of "done" into a file like "/tmp/done.txt". The background could read the contents of the file each time it goes around the loop and terminate when it detects the change. But that's clearly inefficient.

[MochiMoppel]

So if the for loop can't read variables outside the loop, then why is this able to stop the loop?

By using '&' your script started the loop in a subshell.
A subshell inherites values of the parent shell, or better: it makes copies of the variables. This is why the loop knows that done=0 because it was defined before starting the subshell. Once the subshell starts, it cant "see" changes made to the (original) variables of the parent shell.This is why your changes to done=1 is never detected by the subshell. Note that also variables changed in the subshell don't propagate back to the parent shell. This is why in following demo variable i keeps a value of 100 even though the subshell changed the value (of its copy!) to 5.

Code: Select all

#! /bin/sh
done=0
i=100
for ((i=1; i<=5; i++)) {
    echo "Subshell:$done $i"
    sleep 1
} &
done=1	# change value while for loop still running, for loop is unimpressed ;-)
echo "Parent while subshell running:$done $i"
sleep 7
echo "Parent after subshell stopped:$done $i"

Result:

Code: Select all

Parent while subshell running:1 100
Subshell:0 1
Subshell:0 2
Subshell:0 3
Subshell:0 4
Subshell:0 5
Parent after subshell stopped:1 100

[rcrsn51]

@MochiMoppel In your above example, why does "done" not need to be exported in order for the background process to initially see it? What is the difference between this situation and a gtkdialog <action> directive where you must export the variables that you want to pass into the environment of the subshell?

There seems to be a distinction between a subprocess and a subshell that I don't understand.

[Chelsea80]

@wizard

-Start a timer
-Run a process
-Stop the timer when the process is complete

OK, this might help (or not) -

#!/bin/bash

# Start the timer
SECONDS=0

# Run your process here
echo "Running your process..."
sleep 10 # Replace this with your actual process

# Stop the timer and calculate elapsed time
duration=$SECONDS
echo "Process completed in $duration seconds."

[wizard]

@rcrsn51
@MochiMoppel

You are assuming that the "done" variable acts like a C global variable. You want it to be visible to the background process started by the &. So when the foreground process sets "done=1", the loop in the background would detect this and terminate.
But bash variables don't work that way.

That explains it, the explanation that the background process/subshell was created by the "&" and its behavior.

One work-around would be for the foreground to write the value of "done" into a file like "/tmp/done.txt". The background could read the contents of the file each time it goes around the loop and terminate when it detects the change

Had already thought of that and it works fine, just wanted to know why it didn't work with a plain bash variable.

Thanks to everyone who replied.
wizard