An array that doesn’t automatically but intentionally resize is easily extended
into one, with the automatic/current behaviour, and it needs very little additional
code.
But given the current version, you can’t do it the other way around, you’re
forced to waste the resources while there could be a version where it’s optional
and you, as the developer, should be the one knowing which one better fits your
current needs.
Also besides the waste of memory it also can trigger a series of reallocs while you
initially fill your array.
And yes, this is one of the very few parts where I think the language could
and should be extended, as it’s such a fundamental part and in (of course
my personal experience) way more than 50% of the situations where an
array is the data struct of choice, it’s actual (max)size is/can be known upfront.
