[PATCH 0/4] Add variant type

Andrew Burgess aburgess@redhat.com
Fri Feb 4 12:44:31 GMT 2022


* Simon Marchi via Gdb-patches <gdb-patches@sourceware.org> [2022-02-02 20:32:45 -0500]:

> On 2022-02-02 19:02, Andrew Burgess wrote:
> > * Simon Marchi via Gdb-patches <gdb-patches@sourceware.org> [2022-02-01 09:07:13 -0500]:
> > 
> >> I sometimes encounter situations where I think using std::variant would
> >> be useful.  The benefits would be to be able to use non-POD types in the
> >> alternatives, have them automatically constructed when making an
> >> alternative active, and automatically destructing them when switching to
> >> another alternative of destroying the variant.  Also, the validation of
> >> access to the alternatives, to prevent accessing an alternative that is
> >> not the current one.
> >>
> >> However, std::variant is only available in C++17.  I suggest that we
> >> import a variant implementation to allow us to use it right now, a bit
> >> like we did for optional and string_view.  We could maybe import the
> >> implementation from libstdc++, like we did for these last two types, but
> >> that sounds really tedious (we would also want to import any tests).
> >> Instead, we can take advantage of a popular implementation distributed
> >> as a single header file.  I chose this one:
> >>
> >>   https://github.com/martinmoene/variant-lite
> >>
> >> because it is close to the std implementation, and the license (Boost
> >> Software License) is clear and compatible with the GPL v3.
> > 
> > Is this compatible with the copyright assignment requirement of
> > contributing to GDB?
> > 
> > I'm assuming copyright of that file is not assigned to FSF.
> 
> Exactly, we're merely redistributing it (as its license allows).  Its
> header does not say "Copyright FSF" like the other files.
> 
> > My concern would be, what if code was copied from that file into the
> > rest of GDB?  Does that cause us problems?
> 
> Well, we should avoid that :).  I see it exactly like zlib, which is also bundled.

OK.  Thank you for the explanation.

I took a look through this whole series, and it all looks good to me.

Thanks,
Andrew



More information about the Gdb-patches mailing list