cdist/docs/dev/logs/2013-07-25.source-error-does-not-stop-cdist

57 lines
1.6 KiB
Text
Raw Permalink Normal View History

Symptom:
running something in a manifest and that fails does not exist
the cdist run
Analysis:
Find out what the shell does:
[23:56] bento:testshell% cat a.sh
# source something that fails
. b.sh
[23:57] bento:testshell% cat b.sh
nosuchcommand
[23:57] bento:testshell% sh -e a.sh
a.sh: 2: .: b.sh: not found
[23:57] bento:testshell% echo $?
2
-> exit 2 -> looks good
Find out what the python does:
[23:57] bento:testshell% python3
Python 3.3.2 (default, May 21 2013, 15:40:45)
[GCC 4.8.0 20130502 (prerelease)] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import subprocess
>>> subprocess.check_call(["/bin/sh", "-e", "a.sh"])
a.sh: 2: .: b.sh: not found
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/usr/lib/python3.3/subprocess.py", line 544, in check_call
raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '['/bin/sh', '-e', 'a.sh']' returned non-zero exit status 2
>>>
Conclusion:
Manifests that execute (!) other shell scripts does
not necessarily give the -e flag to the other script
-> called script can have failures, but exit 0
if something the last thing executed does exit 0!
Solution:
Instead of doing stuff like
"$__manifest/special"
use
sh -e "$__manifest/special"
or source the script:
. "$__manifest/special"
(runs the script in the same namespace/process as everything in the
calling script)