On Fri, Jul 22, 2016 at 2:07 AM, Malcolm Greene <pyt...@bdurham.com> wrote: > The source file we're generating has one main function (go) with some > supporting functions and classes as well. Will there be any problems > referencing additional functions or classes defined in the source that > gets passed to exec ... as long as references to those functions and > classes happen within the generated module? > > I assume that one downside to the exec() approach is that there is no > persistent namespace for this code's functions and classes, eg. after > the exec() completes, its namespace disappears and is not available to > code that follows?
Functions keep references to their globals, so you should be safe. > The Python documentation also warns: "Be aware that the return and yield > statements may not be used outside of function definitions even within > the context of code passed to the exec() function. The return value is > None." Yeah, that's just emphasizing that exec() is looking for module-level code. Just as you can't put a flush-left "return" statement into a .py file, you can't use one in exec(). As long as your code builder is returning a string equivalent to what you'd find in an actual Python module (that is, exactly what it's already doing, given that you currently attempt to import it), you'll be fine. ChrisA -- https://mail.python.org/mailman/listinfo/python-list
