On 2020-03-04 16:17, Ruben Perez via Boost wrote:
El mié., 4 mar. 2020 10:18, Andrey Semashev via Boost
escribió: I don't see the point of having a Boost wrapper for just one specific backend. People who want that could just use MySQL C or C++ API directly.
Just to clarify, this is *not* a wrapper around the MySQL C API, it is an implementation of the MySQL protocol based in ASIO.
I see this as a building block, that can be used by any other bigger, more abstract library providing multiple backends. This library takes the approach of doing a single thing well. It provides no abstraction over MySQL specifics. But it can be seen as a step towards something bigger.
The other value I see in this is that it is Asio-based. So it would contribute towards the creation of an Asio-based ecosystem. The aim is to integrate well with any project that uses Asio (e.g. using Beast), which I don't think the MySQL C++ API gives you.
I see, thanks for clarifying. A custom implementation of MySQL protocol could be useful to MySQL and ASIO users, so there is value. But I'm still not quite sure it would be universally useful. I mean, if I was implementing an application that needs integration with an SQL database, I'm not sure I would have picked your library over the other alternatives. I'd be more interested in a universal API that allows me to plug in whatever SQL database engine I need without having to rewrite my code. But that's just my wishful thinking.