Re: [boost] Formal Review of Proposed Boost.Histogram Library Starts TODAY

Subject: Re: [boost] Formal Review of Proposed Boost.Histogram Library Starts TODAY
From: Steven Watanabe (watanabesj_at_[hidden])
Date: 2018-09-18 19:13:04

AMDG

On 09/18/2018 12:55 PM, Hans Dembinski via Boost wrote:
>>
>> <snip>
>> I'm not 100% clear on the standards guarantees for the constructor
>> initialiser list (in particular if a base initialiser is guaranteed to
>> be fully constructed before the arguments to other initialisers are
>> evaluated or not), so it's possible that this is UB if that guarantee is
>> not met, but I would have thought it'd be more sensible to init as:
>>
>>      regular(...)
>>          : transform_type(std::move(trans)
>>          , min_(transform_type::forward(lower))
>>          , ...
>>
>> Such that the same instance is used in all cases.
> 
> 
> AFAIK, the order in which these statements are executed is not guaranteed,
> that's why the ctor uses trans instead of transform_type. I am happy to be
> proven wrong by a language expert. If that's the case, I agree that your
> suggestion makes more sense.
> 

Initializers are never interleaved.

"In a non-delegating constructor, initialization proceeds in the
following order:
- ...virtual base classes ...
- Then, direct base classes ...
- Then, non-static data members ..." [class.base.init]

"The initialization performed by each mem-initializer constitutes
a full-expression" [class.base.init]

"Every value computation and side effect associated
with a full-expression is sequenced before every
value computation and side effect associated with
the next full-expression to be evaluated" [intro.execution]

In Christ,
Steven Watanabe