Script entry point
A program’s work lives in one function, marked as the entry point. It runs when the file is executed as a program, not when it is imported as a module.
Marking the entry point
@entry-point
def main []:
print "running as a program"
Save as tool.ki and run:
koskari tool.ki
running as a program
The launcher calls main with no arguments after the module has
finished loading. Its return value is the exit status: None exits
0, an int exits with that status, and any other value is printed to
stderr and exits 1. Importing tool from another module defines
main and prints nothing, and the importer can call main itself.
Arguments and status
The entry takes no arguments; read them from sys.argv, or let a
parameter’s default do it:
import sys
@entry-point
def main [args: (item-slice sys.argv 1)]:
if args:
print "hello," args[0]
else:
print "usage: tool NAME"
2
koskari tool.ki world # prints "hello, world", exits 0
koskari tool.ki # prints the usage line, exits 2
A function that calls sys.exit itself keeps working: the exit passes
through the launcher unchanged.
Compared with Python
Python runs a program’s work from a block guarded by
if __name__ == "__main__":, which executes while the module is still
loading. Koskari has the same guard as when-main, and it runs at the
same time, but a program’s work does not belong there: a body that runs
during the load cannot return a status, is not a function anything else
can call, and keeps the module’s load unfinished for as long as it runs.
The entry point runs after the load completes, so the module is sealed
and every definition is in place when the program starts.