On Thu, Jan 08, 2026 at 06:54:03PM -0800, Steve Kargl wrote:
> On Thu, Jan 08, 2026 at 09:42:53PM +0100, Thomas Koenig wrote:
> > Am 08.01.26 um 16:56 schrieb Hans-Peter Nilsson:
> > > > From: Hans-Peter Nilsson <[email protected]>
> > > > Date: Thu, 8 Jan 2026 09:01:56 +0100
> > > 
> > > > That appears to do it!  With your "p3a.diff" patch applied
> > > > to r16-6514 (i.e. in place of the faulty r16-6515), build is
> > > > restored for cris-elf.  Test-results are not yet ready.
> > > 
> > > Test results are clean at r16-6514 + p3a.diff (i.e. no
> > > regressions compared to r16-6514).  Thanks again for the
> > > fix, I hope it gets committed soon.
> > 
> > Committed as r16-6586-gd5964a270de33349ff137b1835046e65ec81fa1e
> > (this may be pushing obvious and simple for a bit, but I did want
> > to get this out of the way as fast as possible).
> > 
> > Sorry for breaking bootstrap, and thanks a lot for the help!
> > 
> 
> Unfortunately, it is still broken. (unit)->self has a type
> of __gthread_t, which appears to be mapped to pthread_t
> on FreeBSD, which is a struct pthread *.  
> 
> In file included from ../../../gcc/libgfortran/io/unit.c:30:
> ../../../gcc/libgfortran/io/unit.c: In function 'insert_unit':
> ../../../gcc/libgfortran/io/async.h:396:18: error: assignment to 
> '__gthread_t' {aka 'struct pthread *'} from 'int' makes pointer from integer 
> without a cast [-Wint-conversion]
>   396 |     (unit)->self = 1;                                           \

If I add a cast here and elsewhere, I can get bootstrap to complete.

(unit)->self = (__gthread_t)1;

-- 
Steve

Reply via email to