Re: [PATCH 3/3] Use std::string for interpreter_p

[Tom Tromey <tromey@adacore.com>]

>>>>> "Andrew" == Andrew Burgess <aburgess@redhat.com> writes:

Andrew> Tom Tromey via Gdb-patches <gdb-patches@sourceware.org> writes:
>> The global interpreter_p is a manually-managed 'char *'.  This patch
>> changes it to be a std::string instead, and removes some erroneous
>> comments.

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

Thanks.

Andrew> I don't like the name interpreter_p - the _p says "predicate" to me, but
Andrew> I that's not your fault, so can be fixed some other day.

Yeah.  I think maybe it means "pointer"?  Not sure.

I looked at removing this entirely, and it's almost possible except for
the hacks in tui-interp.c:

  interp_factory_register (INTERP_TUI, tui_interp_factory);

  if (interpreter_p == INTERP_TUI)
    tui_start_enabled = true;

  if (interpreter_p == INTERP_CONSOLE)
    interpreter_p = INTERP_TUI;

I don't really understand why the TUI is considered a distinct
interpreter at all.  That seems weird to me.  For one thing, I found it
doesn't work properly with "new-ui" (I filed a bug about this).

Tom