Re: [lmi] Dealing with deleted operator<<() overloads for wide char types in C++20

[Vadim Zeitlin]

On Tue, 2 Mar 2021 22:51:41 +0000 Greg Chicares <gchicares@sbcglobal.net> wrote:

[...discussion snipped...[
GC> Both points of view are valid, but if you call yours "more
GC> constructive", I can call mine "purer" as I would introduce
GC> no novel concept.

 I completely agree with everything above.

GC> Is there an extremely simple and robust lmi_test::to_string()
GC> implementation that you have in mind? If so, I wouldn't want
GC> to dismiss it sight unseen.

 I don't have time to submit a proper[ly tested] patch today, but the
implementation I had in mind would just use std::ostringstream, i.e. rely
on the same overloaded operator<<(std::ostream&, T) which is used right
now, by default. If you'd like, I could provide such patch later this week,
it should be relatively small even though not completely trivial (if only
because we need this to_string() in a couple of different places and I'm
not sure if we want to call it from both of them or try refactoring the
code slightly to avoid duplicating it).

 Please let me know if I should [not] do this,
VZ

pgpjfac6bt5Kd.pgp
Description: PGP signature