Changes by Steve Dower steve.do...@microsoft.com:
--
nosy: +steve.dower
___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue23809
___
___
Terry J. Reedy added the comment:
This is a fairly common newbie bug. random.py seems to be favorite name.
--
nosy: +terry.reedy
___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue23809
___
Robert Collins added the comment:
Implementation wise: this is not part of the generic rendering-of-tracebacks;
I'd like to make the traceback new stuff be tastefully extensible - I'd be
inclined to do this with a per-frame-callback on render (so we don't pay
overhead on unrendered tb's) and
New submission from Nick Coghlan:
A colleague just ran into the issue where they created a json.py module in
their local directory and broke a previously working program. I picked up on
the mistake when I saw the following traceback snippet:
Traceback (most recent call last):
File
Nick Coghlan added the comment:
I had an idea for a possible importlib.util API to support this capability: an
ignore_entries=0 default arg to
https://docs.python.org/dev/library/importlib.html#importlib.util.find_spec
The new arg would say how many found entries to skip when looking for the
Robert Collins added the comment:
Why limit this to just stdlib shadowing?
A local module can shadow a top level module-or-package across the board. If we
don't limit it to stdlib names, it becomes a lot easier to implement.
--
___
Python tracker
Nick Coghlan added the comment:
On the implementation front, +1 for looking at a per-frame callback API rather
than hardcoding this directly into the existing traceback rendering code.
--
___
Python tracker rep...@bugs.python.org
Nick Coghlan added the comment:
I proposed limiting it to stdlib names as that's the case where we see the most
beginner confusion (experimenting with sockets in a file named socket.py,
etc), and the one where we can generate a comprehensive list of known module
names ahead of time (so in