> On Mar 6, 2025, at 3:28 PM, Marcin Jaczewski via Std-Proposals <std-proposals_at_[hidden]> wrote:
>
> czw., 6 mar 2025 o 23:51 Thiago Macieira <thiago_at_[hidden]> napisał(a):
>>
>> On Thursday, 6 March 2025 14:16:47 Pacific Standard Time Marcin Jaczewski
>> wrote:
>>>> Please make your software reproducible: it should be bitwise exactly
>>>> identical if it is compiled by the same compiler.
>>>
>>> Isn't this even worse? could this cause ODR violations?
>>> As same inline function depend on time and two TU process it
>>> a couple of seconds apart, now both TU will have code generated
>>> differently.
>>
>> Is it an ODR violation? It's the same definition, only the compiler produced
>> two different bit streams for it. It would be no different than applying
>> different optimisations. Which one ends up being used is arbitrary.
>>
>
> But if based on that date we remove member from structure?
> Or something like this:
> ```
> auto foo()
> {
> if constexpr (std::chrono::system_clock::now() < is_in_10y_to_2038y)

This would be a bad approach: consider what happens if you want to ship an update to an old platform, now your code has different behavior because of the date you are building.

The kind of guard you want is something I would call a deployment target, but that’s a constant based on the deployment target not the current time of day. (Darwin also has something called a “linked on or after” guard, which sounds like a date based guard but is also based on the deployment target of the main process binary)

Basing the behavior change on the date of compilation is almost certainly the wrong approach.

—Oliver