Re: [PATCH 2/4] allow poisoning input_location in ranges it should not be used

[Trevor Saunders <tbsaunde@tbsaunde.org>]

On Fri, Jul 02, 2021 at 11:46:46AM -0400, Jason Merrill wrote:
> On 7/1/21 6:16 AM, Trevor Saunders wrote:
> > On Wed, Jun 30, 2021 at 11:13:23AM -0400, David Malcolm wrote:
> > > On Wed, 2021-06-30 at 01:35 -0400, Trevor Saunders wrote:
> > > > This makes it possible to assert if input_location is used during the
> > > > lifetime
> > > > of a scope.  This will allow us to find places that currently use it
> > > > within a
> > > > function and its callees, or prevent adding uses within the lifetime
> > > > of a
> > > > function after all existing uses are removed.
> > > > 
> > > > bootstrapped and regtested on x86_64-linux-gnu, ok?
> > > > 
> > > > Trev
> > > 
> > > [...snip...]
> > > 
> > > > diff --git a/gcc/diagnostic.c b/gcc/diagnostic.c
> > > > index d58586f2526..3f68d1d79eb 100644
> > > > --- a/gcc/diagnostic.c
> > > > +++ b/gcc/diagnostic.c
> > > > @@ -1835,7 +1835,7 @@ internal_error (const char *gmsgid, ...)
> > > >     auto_diagnostic_group d;
> > > >     va_list ap;
> > > >     va_start (ap, gmsgid);
> > > > -  rich_location richloc (line_table, input_location);
> > > > +  rich_location richloc (line_table, UNKNOWN_LOCATION);
> > > >     diagnostic_impl (&richloc, NULL, -1, gmsgid, &ap, DK_ICE);
> > > >     va_end (ap);
> > > 
> > > I actually make use of this in the analyzer: the analyzer sets
> > > input_location to stmt->location when analyzing a given stmt - that
> > > way, if the analyzer ICEs, the ICE is shown at the code construct that
> > > crashed the analyzer.
> > > 
> > > This behavior is useful to me, and would be lost with the proposed
> > > patch.
> > 
> > I made this change because otherwise if the compiler ICE's while access
> > to input_location is blocked we end up infinitely recursing complaining
> > we can't access it while trying to say where the last error was.  I was
> > nervous about the change before, and now I agree we need something
> > else.
> 
> How about allowing a scope (like internal_error) to explicitly un-poison the
> variable?

The only reason I didn't add such a thing was not having a need for it,
and not wanting to worry about the possible complexity of combining that
with allowing nested poisoning with a refcount.  However if people think
it would be useful to have this I'm happy to add it to a v2, otherwise
I'm happy to just use this locally for finding code using
input_location, or other globals.

> Separately, all the location_t (input_location) changes are pretty clunky,
> can we make that conversion part of the macro?

Currently its not a macro,  while we could change that, defining  it as
#define input_location ((location_t &) input_location_real)
seems a bit sneaky, but I think you have to define it that way if you
want to avoid updating all the assignments to input_location?

Thanks

Trev

> 
> Jason
>