[Python-Dev] Issue 19332: Guard against changing dict during iteration

[Ethan Furman]

http://bugs.python.org/issue19332

Summary:

--> d = {1: 'one'}
--> for k in d:
...   d[k+1] = d[k] * k
...
Traceback (most recent call last):
   File "<stdin>", line 1, in <module>
RuntimeError: dictionary changed size during iteration

--> for k in d:
...   del d[k]
...
Traceback (most recent call last):
   File "<stdin>", line 1, in <module>
RuntimeError: dictionary changed size during iteration


While the above two cases are covered, the following is not:

--> d = dict.fromkeys('abcd')
--> for i in d:
...     print(i)
...     d[i + 'x'] = None
...     del d[i]
...
d
a
dx
dxx
ax
c
b

Which can yield pretty bizarre results and does not raise an exception.

There is a patch to fix this.

Raymond's first comment was:

"The decision to not monitor adding or removing keys was intentional.  It is just not worth the cost in either time or 
space."

Considering that the first two scenarios do raise exceptions that statement does not seem correct, at least not with 
current code.

Later, Raymend rejected the patch with the comment:

"I disagree with adding such unimportant code to the critical path."

I understand we need to keep the critical path as fast as possible, and that it is a balancing act between completely 
correct code and warnings in the docs.

What I'm hoping for is either someone to shed some light on what is the "unimportant" part, or perhaps some performance 
measurements that would show there is no performance based reason to not have the patch added.  (I can try to do the 
performance part myself if necessary, I'm just not sure what all the steps are yet.)

--
~Ethan~