Re: [PATCH 09/10] Class-ify lm_info_darwin

[Simon Marchi <simon dot marchi at polymtl dot ca>]

On 2017-04-28 12:00, Pedro Alves wrote:
> On 04/26/2017 11:51 PM, Simon Marchi wrote:
> > This patch makes lm_info_darwin a "real" class.  It initializes the
> > field and replaces XCNEW/xfree with new/delete.
> >
> > I believe the cleanup in darwin_current_sos can be removed, I don't 
> > see
> > anything that can throw after the allocation and the call to
> > discard_cleanups.
>
> Agreed.
>
> LGTM but ...
>
> > --- a/gdb/solib-darwin.c
> > +++ b/gdb/solib-darwin.c
> > @@ -41,6 +41,8 @@
> >  #include "mach-o.h"
> >  #include "mach-o/external.h"
> >
> > +#include <memory>
>
> ... this doesn't look necessary?  (it was not in the ChangeLog.)

Ah, the comment in the commit log and this include is a leftover.  While 
doing the preparatory patch for darwin (01/10), I had put an xfree 
cleanup for the lm_info_darwin structure, but I later removed it.

Should we expect that the "new" operator can throw if memory allocation 
fails?  In that case we can't get rid of the cleanup for newobj.

Simon