Subshells vs the Current Shell ( ( ) vs { } )

Variable Changes Inside a Subshell Don't Escape

Because ( ... ) forks a separate copy of the shell, any variable assignments made inside it are lost once the subshell finishes; the parent shell's variables are untouched.

Example: Variable Changes Inside a Subshell Don't Escape

#!/bin/bash
count=1
( count=99; echo "inside subshell: $count" )
echo "outside subshell: $count"

Curly Braces Share the Current Shell

{ commands; } groups commands but runs them in the current shell environment, so variable assignments and cd calls inside persist after the block finishes.

Note: Note the required space after { and the semicolon before } — these are syntax requirements, not style choices.

Example: Curly Braces Share the Current Shell

#!/bin/bash
count=1
{ count=99; echo "inside group: $count"; }
echo "outside group: $count"

Isolating a Directory Change

A subshell is a clean way to temporarily cd somewhere and run a command there without having to remember to cd back afterward; once the subshell exits, the parent's working directory is unaffected.

Example: Isolating a Directory Change

#!/bin/bash
mkdir -p sub_demo
echo "before: $(pwd | xargs basename)"
( cd sub_demo && echo "inside subshell: $(pwd | xargs basename)" )
echo "after: $(pwd | xargs basename)"
rmdir sub_demo

exit Behaves Differently in Each

Calling exit inside ( ... ) only terminates that subshell, letting the parent script continue; calling exit inside { ...; } terminates the entire script since it runs in the same process.

Example: exit Behaves Differently in Each

#!/bin/bash
( echo "in subshell before exit"; exit 1; echo "never printed" )
echo "script continues after the subshell exited, status was $?"