Contexter: A better contextlib
To keep it short: Contexter allows you to nest and stack context managers in an easy and intuitive way.
Enough talk, let's see an example:
''' Copy the content of one file to another and protect everything with a lock. ''' with Contexter(lock) as ctx: in_file = ctx << open('a.txt') out_file = ctx << open('b.txt', 'w') out_file.write(in_file.read())
Look at that. It's beautiful, isn't it? Let me explain: You call Contexter() with any number of context managers as arguments and later attach additional managers with the neat value = ctx << thing syntax. That's it. Only one level of indentation, no matter how many managers you need.
Just for comparison:
# Python 2.5 and 2.6 with lock: with open('a.txt') as in_file: with open('b.txt', 'w') as out_file: out_file.write(in_file.read()) # Starting with Python 2.7 and 3.2 with lock, open('a.txt') as in_file, open('b.txt', 'w') as out_file: out_file.write(in_file.read()) # Deprecated since Python 2.7 and 3.2 with contextlib.nested(lock, open('a.txt'), open('b.txt', 'w')) as values: in_file, out_file = values out_file.write(in_file.read()) # Since Python 3.3 (not backported to 2.7) with contextlib.ExitStack() as stack: stack.enter_context(lock) in_file = stack.enter_context(open('a.txt')) out_file = stack.enter_context(open('b.txt', 'w')) out_file.write(in_file.read())
Replacing contextlib.nested 
You can use Contexter(*managers) as a drop-in replacement for contextlib.nested(*managers), just without the confusing error prone quirks mentioned in the official documentation.
Replacing contextlib.closing 
Just forget about it. Contexter turns close-able objects into context managers automatically.
Replacing contextlib.ExitStack 
Contexter offeres everything contextlib.ExitStack does (and more). If you want a drop-in replacement that also works for Python 2.x and 3.2, you can use our backported ExitStack, a subclass of Contexter that is API compatible to the contextlib variant.
Replacing everything else from contextlib
If you really want to stick with the standard API, you can. Contexter implements all public APIs from contextlib and backports new features as soon as they are introduced.
Contexter keeps track of the results of all invoked context managers. You can access the results later and don't have to unpack them all at the beginning.
with Contexter(open('a.txt'), open('b.txt', 'w')) as ctx: in_file, out_file = ctx.values() assert ctx.value(0) is in_file assert ctx is in_file assert ctx[0:2] == [in_file, out_file] assert len(ctx) == 2
If you don't like the << syntax, there is a method that does the same.
with Contexter() as ctx: in_file = ctx << open('a.txt') out_file = ctx.append(open('b.txt', 'w'))
Contexter contexts are nestable. Each level of nesting maintains its own stack of context managers and result values. This allows you to control the lifetime of contexts very precisely.
with Contexter() as ctx: out_file = ctx << open('b.txt', 'w') with ctx: in_file = ctx << open('a.txt') copy_data(in_file, out_file) assert in_file.closed == True assert out_file.closed == False