Re: Any objections to making C++14 the current baseline for Boost support and testing?
Dominique Pellé via Boost-users <[email protected]>
| Newsgroups | gmane.comp.lib.boost.user |
|---|---|
| Message-ID | <CAON-T_hgqZ9D467sja+OWhTCSD-9eb4u6LH4FZRY7GPJrihGVg@mail.gmail.com> |
Christophe B via Boost-users <[email protected]> wrote: > > Hi, > > Would there be an intent to increase the support of new standard libraries by boost libraries? On top of my head, I think first to shared_ptr<> and > error_code. > Mixing std libraries and their equivalent in boost is most of the time tedious. Concretely, it becomes difficult to integrate components and > interoperate them when some adopt std::error_code while others must stick boost::system::error_code due to constraints (mainly asio and beast. We > thought to migrate to the standalone asio, but there is no equivalent for beast). Even if some efforts have been done to convert boost error_code to > std ones, this is far from ideal. > > And as mentioned by Martijn Otto, I would opt also for jumping directly to C++17. In my company at least, we're still stuck with C++14 because some customers use old toolchains and we use boost. So requiring C++17 would be annoying. Boost is nice for projects that can't use a recent C++ standard as it provides things like boost::optional, boost::variant, boost::filesystem (etc.) instead of the std ones which require newer C++ standards. Dominique