Ken Hilton writes:
> Can you spot the problem? The "with open(filename)" statement is supposed
> to ensure that the file object is disposed of properly. However, the "yield
> f.read()" statement suspends execution within the with block, so if this
> happened:
>
> for contents in read_multiple('chunk1', 'chunk2', 'chunk3'):
> if contents == 'hello':
> break
>
> and the contents of "chunk2" were "hello" then the loop would exit, and
> "chunk2" would never be closed! Yielding inside a with block, therefore,
> doesn't make sense and can only lead to obscure bugs.
You have one example of a bad outcome. That hardly shows that *no*
yield in a with block makes sense. On the contrary Chris A's
suggestion that a function might yield multiple times within one
context manager seems valid, although it might be bad practice.
There's another possibility. Context managers have general
initialize-operate-finalize semantics. True, they *often* have the
simple form "return the value of *expr* bound to *var* at entry, clean
up the contents of *var* at exit". But I find it easy to imagine that
there are situations where the __enter__ method is complex and the
__exit__ method is a no-op. In that case the problem you identify
can't happen, and yielding inside such a context manager would be
harmless. A third possibility is that the __exit__ method has
semantics like the "else" of "while ... else", where the "normal" use
case involves a break, but occasionally the input is exhausted, and
some "default" action needs to be taken.
I grant that I don't have examples of the latter (after all, writing
context managers is a relatively rare activity, and the resource
acquisition-and-release paradigm is prominent). But Chris's point
about line-oriented generators seems both plausible and common.
His analysis of why the risk is small also seems plausible, so I would
say this problematic case (and it is problematic) is covered by the
"consenting adults" principle.
Steve
_______________________________________________
Python-ideas mailing list
Python-ideas@python.org
https://mail.python.org/mailman/listinfo/python-ideas
Code of Conduct: http://python.org/psf/codeofconduct/
