muck is a general-purpose build tool, sort of like make, but with a different design philosophy:  make is "top-down" (you specify the high-level target you want to build, and make gathers and assembles the necessary pieces), muck is "bottom-up" (it looks at the available pieces, and figures out what can be built with them).  The Muck model enables automatic discovery of dependencies, and also enables you to write Muckfiles in any programming language.

muck excels in situations where dependencies are numerous, dynamic, or difficult-to-know.  (A good example of this situation is static websites.)

muck is based on Fabricate.  It uses strace to monitor execution and extract dependency information.

Example Muckfile

Here is a very simple exmple, so that you can easily compare Makefiles vs Muckfiles:

#!/usr/bin/env python3

# This Muckfile will do the same thing as this Makefile:
# (Notice that the Makefile needs to list the .h files,
#  while the Muckfile does not.)
#
# mycmd: a.c b.c a.h b.h c.h d.h
#     gcc -o mycmd a.c b.c -I.

import os

if os.environ['MUCK_REL_PATH']=='a.c': print('gcc -o mycmd a.c b.c -I.')

Here is a more realistic example, used for building static websites:

#!/usr/bin/env python

# Muckfiles can be written in any language.  They just need to print shell commands to stdout.
#
# Muck provides a few environment variables as inputs:
#
#     MUCK_IN_ROOT   -- The input root directory.  Example:  /path/to/proj/in
#     MUCK_OUT_ROOT  -- The output root directory. Example:  /path/to/proj/out
#     MUCK_REL_PATH  -- The file to process, relative to the MUCK_IN_ROOT.  Ex: a/b.c
#
# Muckfiles should be "functional";  For the same set of inputs (env vars),
# they should always produce the same set of outputs (shell commands).

import os, sys
from fnmatch import fnmatch

# Pre-calculate a few things to make our command recipes simpler:
relPath = os.environ['MUCK_REL_PATH']
filename = os.path.basename(relPath)
inRoot, outRoot = os.environ['MUCK_IN_ROOT'], os.environ['MUCK_OUT_ROOT']
inPath  = os.path.join(inRoot,  os.environ['MUCK_REL_PATH'])
outPath = os.path.join(outRoot, os.environ['MUCK_REL_PATH'])
inDir, outDir = os.path.dirname(inPath), os.path.dirname(outPath)

def end(string=None):
    if string: print(string)
    sys.exit()

def mk_outdir(): print('''
{ set +x; } 2>/dev/null  # Silently disable tracing
mkdir -p %(outDir)r
set -x                   # Re-enable tracing
'''%globals())

def SKIP(verbose=False):
    if verbose: print >>sys.stderr, 'Skipping:', relPath
    sys.exit()

def ENTER_SUBDIR(): end('''
{ set +x; } 2>/dev/null  # Silently disable tracing
true                     # The command doesn't actually matter, as long as it succeeds.
echo DIR DONE: $MUCK_REL_PATH/
find . >/dev/null        # So we re-visit this dir if any content changes.
''')

def COPY():
    mk_outdir()
    end('cp --preserve=all %(inPath)r %(outPath)r'%globals())

def MAKO():
    noext, ext = os.path.splitext(relPath)    # Chop off the .mako extension.
    assert ext == '.mako'
    values = {'noext':noext}; values.update(globals())
    mk_outdir()
    end('python -m pyhpy.cmd --template-dir=%(inRoot)r --module-dir=%(inRoot)r /%(relPath)r >%(noext)r'%values)

def MARKDOWN(): end('../PyHPy/bin/MARKDOWN')



if fnmatch(filename, '_*'):           SKIP()
if os.path.isdir(relPath):            ENTER_SUBDIR()


# A simple way to enable building of specific items (for improved development iteration time):
for pat in os.environ.get('MUCK_REL_PATH_PREFIXES', '').split('|'):   # If the env var is not set, the resulting 'pat' will be '', which matches everything.
    if relPath.startswith(pat): break
else: SKIP(True)


if fnmatch(filename, '.*'):           SKIP()
if fnmatch(filename, '*.meta'):       SKIP()
if fnmatch(filename, '*.pyc'):        SKIP()
if fnmatch(filename, '*.mako.py'):    SKIP()  # Mako-generated module
if fnmatch(filename, '*.mako'):       MAKO()
if fnmatch(filename, '*.md'):         MARKDOWN()
if fnmatch(filename, '*'):            COPY()  # Default processor

Usage

muck IN_DIR OUT_DIR