Re: [PATCH v2] gdb: make gdbpy_parse_command_name return a unique_xmalloc_ptr

[Marco Barisione <mbarisione@undo.io>]

On Tue, 23 Mar 2021 at 18:38, Simon Marchi via Gdb-patches
<gdb-patches@sourceware.org> wrote:
>
> This avoids some manual memory management.
>
> cmdpy_init correctly transfers ownership of the name to the
> cmd_list_element, as it sets the name_allocated flag.  However,
> cmdpy_init (and add_setshow_generic) doesn't, it looks like the name is
> just leaked.  This is a bit tricky, because it actually creates two
> commands (one set and one show), it would take a bit of refactoring of
> the command code to give each their own allocated copy.  For now, just
> keep doing what the current code does but in a more explicit fashion,
> with an explicit release.
>
> gdb/ChangeLog:
>
>         * python/python-internal.h (gdbpy_parse_command_name): Return
>         gdb::unique_xmalloc_ptr.
>         * python/py-cmd.c (gdbpy_parse_command_name): Likewise.
>         (cmdpy_init): Adjust.
>         * python/py-param.c (parmpy_init): Adjust.
>         (add_setshow_generic): Take gdb::unique_xmalloc_ptr, release it
>         when done.

Any news on this patch (which looks good to me)?
You wrote this in response to my command renaming work[1]. Sorry for
only looking at it now.
It would be helpful to merge it, so I can rebase my command renaming
patches on this. If it's not ready and needs more work, I'm happy to help.

[1] https://sourceware.org/pipermail/gdb-patches/2021-March/177165.html