[Python-ideas] Python-ideas Digest, Vol 138, Issue 95

[Angus Hollands]

Concerning my previous email,

Yes, my mistake. I'd forgotten (ironically) that the whole point is that
it's an expression itself.

So
> while (value:=get_next_pool_item()).in_use:
>    print(value.refcount())
would be the appropriate analogue.

Consequently, my example is invalid. A better example would have been where
one needs to access more than one attribute of the expression

while (node.x, node.y) > (5.0, 5.0) given node = get_neighbours(node):
    pass

This wouldn't be so easy to do without writing a lambda, or conditional
break.
Again, the example is completely arbitrary, simply there to indicate that
moving the assignment out of the expression itself gives more scope for use.
Hence, by superset, I meant that all cases using ":=" can be expressed with
given, but not visa-versa.

---

> Clearly the objectively best choice is "<-".
I'm convinced ;)

> while (cmd := get_command()).token !=
> CMD_QUIT:
I think this example is concise enough that it doesn't consider the
readability when get_command is nontrivial.

Examples:
https://gist.github.com/agoose77/7ac98e74c3b7f1892789575c8cd50536

If i'm refactoring this kind of code, or new to the module and attempting
to unpick what is happening, I would argue that the latter example presents
me with what I most care about up-front (the condition) and maybe care
about (the implementation) second. Furthermore, the assignment is very
obvious, which I think is quite important.
I don't think that the ':=' approach is entirely cryptic, of course, I can
read it. But I find the clarity in putting the expression before the
assignment to be worthwhile.
---

> A bit like how "strawman argument" is mostly used to mean "dammit, you
> just spotted an unwelcome consequence and/or flaw in my position which I
> have no counter too".
Not sure if this is directed at me (will assume so). I'm sure people misuse
the term, but here I
think it's appropriate. It doesn't quite fit the exact definition, I grant
you, but arguing about character cost seems premature in the question of
whether one should care about character cost (within reason).
---

>:= would prevent you from using assignment expressions inside f-strings, which
could be argued is a good thing.
Oh wow, that's rather interesting! I'll put it up there with "py2 f-strings
as codecs". :)

Cheers,
Angus
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-ideas/attachments/20180512/5710c214/attachment-0001.html>