Return true if nothing has failed in cli

[logingood]

We should ensure that we return true,
otherwise changes introduced in #276 will cause stack_master to always
exit with code 1, which breaks integration tests.

Addresses issues opened in #277.

cc: @grosser

[mesge]

Add a test? Otherwise 👍

[grosser]

this looks wrong to me, the result of execute_stacks_command is not used anywhere afaik

[grosser]

#276 made the execute_stacks_command do an exit 1 when any command.perform failed ... so maybe the issue is that some command.perform do not return true ?
if commander fails when true is not returned, then the exit 1 added in #276 should be removed

[grosser]

yep ... lib/stack_master/commands/apply.rb does never return true

[logingood]

so just remove OS exit from #276 @grosser and maybe supply a message that stack is missing instead ?

[grosser]

The exit is correct in that position afaik, we want exit 1 if a command fails and should not ignore the issue. For some reason an empty apply thinks it's failed ... most likely coming from lib/stack_master/command.rb:26 ... so add a puts caller there to see who triggers it ...

[logingood]

it doesn't go in command.rb:26, will take further look

[logingood]

@grosser it comes from here https://github.com/envato/stack_master/blob/master/bin/stack_master#L14, thing does exit false, which returns status code 1 if there is no true returned by execute! from cli.rb

[logingood]

ruby /tmp/test.rb; echo $?
1
⌂145% [mmukhtarov:~] % cat /tmp/test.rb
#!/usr/bin/ruby

def foo
  puts
end

result = foo
exit !!result

[logingood]

well, I probably getting the problem wrong, anyway would appreciate any help, also will look into tests.

[grosser]

#280