Common advantage of Shell scripting over other languages like Go or Python for scripting, is that commands can be simply written in plaintext,
without any hassle. Well, cops-hq commands.Executor enables you to do the same, for example:
result, err := executor.Execute("echo test")With commands.Executor you get full native output and error parsing, in statically typed objects in Go. The result, which is always a
string, can simply be converted to a struct or any other object if parsable (like commands outputting json or yaml). Error checking is
also much simpler then in shell scripting, because both command errors and stderr outputs are considered as errors.
Make sure to check out the code docs on the commands.Executorinterface for available methods and settings!
Usually, you will use the HQ setup, in which you can call the hq.GetExecutor() method to get the currently configured
executor. Executor can be configured in one of two modes:
- chatty, which is similar to shell scripting - all commands and their outputs are shown on stdout / stderr
- quiet, in which no command output will be piped to stdout/stderr (but it is always written to log file), unless a verbose flag
is provided via viper (can be loaded from CLI arguments, more info about that in next sections of the documentation). Quiet
mode is interesting because it removes all the clutter of command output from the console, and you then only show semantic messages
like
logrus.Info("deploying something important...). Pattern with configuring the executor in quiet mode is keeping it quiet for local developers, but setting the--verbosein CI systems so that the full output is always shown.
You can instantiate the Executor yourself using one of the factory methods as well. Multiple instance of Executors are supported in your applications, in case you want multiple executors with different setups. Usually however, just ask your hq.HQ instance for one.