Re: async-mqtt5 is CONDITIONALLY ACCEPTED

From: Thomas Fowlery (thomas.fowlery.yes_at_[hidden])
Date: 2024-11-05 05:39:11

On Mon, Nov 4, 2024, 23:11 Vinnie Falco via Boost <boost_at_[hidden]>
wrote:

> On Mon, Nov 4, 2024 at 1:40 PM Kostas Savvidis via Boost <
> boost_at_[hidden]> wrote:
>
> "The library is designed a certain way, and it works. However it would be
> better if the library was designed this different way."
>
> This may be true, and based on conversations with other contributors who
> have been here longer than me I don't think it is a strong reason (on its
> own) to reject a library.

Isn't criticizing the design of a library the whole point of the review
process? The design may be "good enough" and it might "work", but is this
the standard Boost is aspiring to? Should all comments about the
overarching design be automatically dismissed and focus shifted solely on
reviewing function signatures?

Given the rationale like this:

*This is the core of this review: a personal preference of how*
*libraries like this have to be structured.*
*It's not a review of the library submitted, but a rejection of the*
*kind of library it is (an asio based client).*

it seems like this is the case, and you cannot disagree with "the kind of
the library it is".

I completely agree with Kostas here.