[Python-Dev] If you shadow a module in the standard library that IDLE depends on, bad things happen (original) (raw)

Laura Creighton lac at openend.se
Thu Oct 29 15:23:42 EDT 2015

In a message of Thu, 29 Oct 2015 19:13:08 +0000, Paul Moore writes:

I am actually sick of the 'consenting adults' argument. I am dealing with '11 year old children trying to write their first, third and tenth python programs'. For the life of me I cannot see how convenience for the sort of person who has a legitimate reason to shadow the syslib should get a higher priority over these mites who are doing their damndest to write python despite natural language barriers and the fact that their peers and parents think they are nuts to want to do so. That's actually a very good point, and I agree totally. To my mind, the point about "consenting adults" (and when I referred to that I was anticipating others using that argument, not proposing it myself) is that we don't prevent people from doing weird and wonderful things. But conversely, it's not a reason for making it easy to do such things. Quite the opposite - a "consenting adult" should be assumed to be capable of writing an import hook, or manipulating sys.path, or whatever. Paul

Hmmm, I think the set of 'consenting adults who cannot write an import hook' is rather large.

But all I am asking for is a warning -- and it would be good if Idle noticed the warning and before it fell over dead with its message of firewalls it would mention the warning again, as a problem to look for.

It will bugger up doctests for people who legitimately shadow the stdlib and now get a new warning. Anybody else be seriously inconvenienced if we do this? I cannot think of any, but then legitimately shadowing the stdlib in not on the list of things I have done. Perhaps Dstufft has ideas on this line.

Laura

More information about the Python-Dev mailing list