This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH 07/10] Add __pthread_set_abort_hook export
On Wed, Jan 16, 2013 at 12:52:48PM -0800, Andi Kleen wrote:
> Rich Felker <dalias@aerifal.cx> writes:
> >
> > I object to the existence of such a function entirely. It creates
> > global state where the desired behavior is local. If such a function
> > were to exist, it should be setting an attribute of the particular
> > mutex to store the function pointer, not setting a global.
>
> TSX transactions always flatten, there is no abort to a nested
> transaction. Any abort always has to be handled in the outer layer.
> TSX aborts are by definition global.
I mean it's unreasonable for library 1 to setup an about handler when
library 2 might have already set one. Especially since they may both
be running simultaneously in different threads..
> > I
> > agree with everybody else that external tools, not introspective
> > interfaces, are the right way to handle something like this.
>
> I don't think it makes sense to rely on external tools for assert()
> functionality.
How is this "assert functionality"?
Rich