|
Boost : |
From: Robert Ramey (ramey_at_[hidden])
Date: 2022-05-12 15:03:00
On 5/12/22 4:22 AM, John Maddock via Boost wrote:
>
>> If I recall correctly, the boost "rules" are that a library submitted
>> to boost must be conformant with the most recent version of the C++
>> standard. Assume the library gets accepted. Its then no required to be
>> constantly upgraded to subsequent standards. Soooo
>>
>> a) Because specifies backward compatibility as a requirement, the
>> library is guaranteed to compile and function in all subsequent
>> versions of C++
>>
>> b) Any subsequently added library is also guaranteed to compile and
>> function in all version of C++.
>
> Point of pedantry: that's not true at all, stuff get's deprecated and
> then removed.
I don't think it's pedantic. You're right.
I had forgotten about these. I can say that over 20 years, I've never
run across this with the serialization library which has a lot of lines
of code. Were it to happen, I'd have to decide what kind of
compatibility to maintain. Given that 03 isn't really testible these
days, that might be the end of that.
It's also very possible that one or more of these presented themselves
and it so trivial to update that I forgot about it.
All in all I don't think this should be a big issue.
>
> For example unary_function is deprecated in C++11 and removed in C++17,
> which has just hit for example:
> https://github.com/boostorg/container_hash/issues/22#issuecomment-1124205780
> as gcc-12 starts enforcing this (msvc has done so for a while).
>
> There are plenty of other examples.
>
> Best, John.
>
>
> _______________________________________________
> Unsubscribe & other changes:
> http://lists.boost.org/mailman/listinfo.cgi/boost
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk