This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [RFA] Inform about new thread in a single place.
> From: Vladimir Prus <vladimir@codesourcery.com>
> Date: Fri, 14 Dec 2007 22:07:19 +0300
>
> On Friday 14 December 2007 21:47:45 Mark Kettenis wrote:
> > > From: Vladimir Prus <vladimir@codesourcery.com>
> > > Date: Fri, 14 Dec 2007 20:09:08 +0300
> > >
> > > We have lots of place where "[New thread XXX]" is printed.
> > > This patch makes add_thread responsible for that. OK?
> >
> > No, there are places in the code where add_thread() is called where we
> > should not print the "[New thread xxx]" message. For example in
> > bsd-uthread.c, where the main thread is added to the list of threads
> > to associate process ID and thread ID.
>
> Do you mean this code:
>
> /* HACK: Twiddle INFERIOR_PTID such that the initial thread of a
> process isn't recognized as a new thread. */
> if (ptid_get_tid (ptid) != 0 && !in_thread_list (ptid)
> && ptid_get_tid (inferior_ptid) == 0)
> {
> add_thread (ptid);
> inferior_ptid = ptid;
> }
>
> ? I've just re-checked, and it appears that:
Yes.
> 1. The modules that called add_thread without printing
> a message about new thread, prior to my patch are:
>
> - aix-thread.c
> - bsd-uthread.c
> - corelow.c
> - hpux-thread.c
> - nto-procfs.c
>
> Looking at those, it seems like the place you've pointed
> at is the only one where reporting of new thread is undesirable.
>
> What about me adding 'add_thread_silent' function that will
> not print any message, and using it in bds-uthread.c?
That'd be acceptable to me. I'm not absolutely true that none of the
modules you mention above don't need this, but we can fix those when
we notice the problem.
Mark