On Sun, Nov 10, 2013 at 2:37 AM, Steven D'Aprano <steve+comp.lang.pyt...@pearwood.info> wrote: > I think that Chris is wrong about Python "only" interning > strings if you explicitly ask for it. I recall that Python will (may?) > automatically intern strings which look like identifiers (e.g. "spam" but > not "Hello World" or "123abc").
I'm pretty sure it's simply that literals are interned, or at least shared across a module (and the interactive interpreter "counts" as a module). And it might still only be ones which look like identifiers, because: >>> foo = "lorem ipsum dolor sit amet" >>> bar = "lorem ipsum dolor sit amet" >>> foo is bar False My "only" was false because of the sharing/interning of (some) literals, which I'd forgotten about; however, there's still the distinction that I was trying to draw, that in Python _some strings_ are interned (a feature you can explicitly request), rather than _all strings_ being interned. And as is typical of python-list, it's this extremely minor point that became the new course of the thread - my main point was not whether all, some, or no strings get interned, but that string interning makes the storage space of duplicate strings immaterial :) ChrisA -- https://mail.python.org/mailman/listinfo/python-list
