Trawl 0.1

This is as straight of a clone of Rake that I could come up with.

Briefly:

# Trawlfile
@task
def bar():
    print "BAR!"

Does as you might expect:

$ trawl -T
bar
$ trawl bar
** Execute bar
BAR

And with dependencies:

# Trawlfile
@task
def bar():
    print "BAR!"

@task([bar])
def foo():
    print "FOO!"

Also does what you’d expect:

$ trawl -T
bar
foo
$ trawl bar
** Execute bar
BAR!
$ trawl foo
** Execute bar
BAR!
** Execute foo
FOO!

Which is awesome sauce, but what about files?

# Trawlfile
@build("myfile.txt", recreate=False)
def run(task):
    with open(task.name) as out:
        out.write("MUNCTIONAL!")

@build("yup.cfg", ["myfile.txt"])
def more(task):
    with open(task.name) as out:
        with open(task.source) as src:
            out.write(src.read() + "!!!1!")

Will cause “myfile.txt” to be created but not overwritten if it exists. “yup.cfg” will then only be rebuilt when “myfile.txt” changes as determined by the file’s mtime:

$ trawl yup.cfg
** Execute myfile.txt
** Execute yup.cfg
$ trawl yup.cfg
$

The last major bit is in defining rules:

@rule('.o', '.c'):
def compile(task):
    subprocess.check_call(["gcc", "-o", task.name, "-c", task.source])

objects = FileList("*.c").sub(".c$", ".o")
@build("appname", objects)
def link(task):
    subprocess.check_call(["gcc", "-o", task.name] + task.sources)

# Add a helpful handle for running
task("run", ["appname"])