Functions unnecessarily called in Python/pylifecycle.c:_Py_InitializeCore() ?

[Ned Batchelder]

On 3/1/18 7:40 AM, Thomas Nyberg wrote:
> On 03/01/2018 12:46 PM, bartc wrote:
>> If they're only called once, then it probably doesn't matter too much in
>> terms of harming performance.
> Oh yeah there's no way this has any affect on performance. A smart
> compiler might even be able optimize the call away entirely. Even if it
> couldn't, it's about as fast as an operation could possibly be.
>
>> As for leaving them in, there might be a number of reasons. One, if one
>> day some special initialisation does need to be done, then this gives a
>> place to put it.
>>
>> I quite often have an initialisation routine for a module, that
>> sometimes ends up empty, but I keep it in anyway as often things can get
>> added back.
> Yeah I figure this is the reasoning. Personally I don't like having
> noops for this sort of code structure reasoning, but I can see how rules
> like "everything gets init'ed" is also a fair strategy. Half of my
> question was simply whether that is the case or if possible it was just
> a mistake when older versions that did something could be removed (e.g.
> the code example you put int).
>
>

This sounds like it could make a good contribution to CPython :)

--Ned.