Re: port-sun2/56158

[Jan-Benedict Glaw <jbglaw%lug-owl.de@localhost>]

The following reply was made to PR port-sun2/56158; it has been noted by GNATS.

From: Jan-Benedict Glaw <jbglaw%lug-owl.de@localhost>
To: gnats-bugs%gnats.netbsd.org@localhost
Cc: 
Subject: Re: port-sun2/56158
Date: Thu, 20 May 2021 08:17:51 +0200

 --qcafsfxjcxbk4oxx
 Content-Type: text/plain; charset=utf-8
 Content-Disposition: inline
 Content-Transfer-Encoding: quoted-printable
 
 Hi!
 
 Seems I didn't send email to GNATS about this build issue, so here's a
 final update:
 
   * As it turns out, the (full path and) filenames of libc sources
     turn up in the ramdisk's multicall binary. The ramdisk is tight
     for sun2, so this broke with my setup (Linux host, CI builds not
     /usr/src, but similar to
     /var/lib/laminar/run/4/netbsd-sun2-m68000/NetBSD-src/ . The
     additional binary size made the ramdisk go out of free space.
 
   * Christos suggested to just build it as a reproducible build, as
     that should cut down my long pathes to /usr/src. This failed, the
     long pathes kept showing up, breaking ramdisk creation.
 
   * Turned out: GCC's path remapping code accidentally got merged out.
     Unfortunately, the reproducibility checking scripts didn't catch
     that, was both reference builds were done from the same directory.
 
   * Christos reinstated the GCC patch and sun2 now builds for me as
     well, with the long pathes properly substituted.
 
 `---> Bug is successfully fixed, report can be closed.
 
 Thanks a lot!
 
 Jan-Benedict
 
 --=20
 
 --qcafsfxjcxbk4oxx
 Content-Type: application/pgp-signature; name="signature.asc"
 
 -----BEGIN PGP SIGNATURE-----
 
 iF0EABECAB0WIQQlDTvPcScNjKREqWEdvV51g5nhuwUCYKX/CwAKCRAdvV51g5nh
 u64mAJ4mLkDRfX07bZfOVb2+9OyElca+YwCeM3bfnuRDhQHkAAQ5M05u9W0nAzU=
 =8vf5
 -----END PGP SIGNATURE-----
 
 --qcafsfxjcxbk4oxx--