Re: [PATCH 2/5] gdbsupport: make gdb_abspath return an std::string

[Pedro Alves <pedro@palves.net>]

On 2022-04-18 20:41, Tom Tromey wrote:
>>>>>> "Simon" == Simon Marchi via Gdb-patches <gdb-patches@sourceware.org> writes:
> 
> Simon> From: Simon Marchi <simon.marchi@efficios.com>
> Simon> I'm trying to switch these functions to use std::string instead of char
> Simon> arrays, as much as possible.  Some callers benefit from it (can avoid
> Simon> doing a copy of the result), while others suffer (have to make one more
> Simon> copy).
> 
> I hate copies but the main thing is to make sure none of them are in hot
> paths.  This seems fine on that front.
> 
> Sometimes I wonder if gdb would be better off with its own string type
> -- say a combination of gdb::unique_xmalloc_ptr<char> and string_view.
> This would interact more nicely with legacy code without sacrificing C++
> string methods.  We could stick the string length in there as well if we
> thought that was worthwhile.

That'd be something like cstring_view, and a corresponding cstring type, I suppose:

 http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2019/p1402r0.pdf
 https://github.com/breyerml/cstring_view

I've seen other projects and public discussions on the interwebs adding or reporting
use of such types.