Re: pkg/51534: qemu-2.7.0 fails to build: bad register name

[Kamil Rytarowski <n54%gmx.com@localhost>]

The following reply was made to PR pkg/51534; it has been noted by GNATS.

From: Kamil Rytarowski <n54%gmx.com@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc: 
Subject: Re: pkg/51534: qemu-2.7.0 fails to build: bad register name
Date: Fri, 7 Oct 2016 22:24:41 +0200

 This is an OpenPGP/MIME signed message (RFC 4880 and 3156)
 --jQqjliatJ4KqMe5MLf4CifjIpNvjsR0S6
 Content-Type: multipart/mixed; boundary="90JSrRx1o971x4D8OVav1kfmOpHI6OeEs"
 From: Kamil Rytarowski <n54%gmx.com@localhost>
 To: gnats-bugs%NetBSD.org@localhost
 Message-ID: <5b7cd034-9b91-0479-b781-5314ea8e297c%gmx.com@localhost>
 Subject: Re: pkg/51534: qemu-2.7.0 fails to build: bad register name
 References: <pr-pkg-51534%gnats.netbsd.org@localhost>
  <20161006191814.E3DE07A2A9%mollari.NetBSD.org@localhost>
  <20161007201001.AF4D57A2B5%mollari.NetBSD.org@localhost>
 In-Reply-To: <20161007201001.AF4D57A2B5%mollari.NetBSD.org@localhost>
 
 --90JSrRx1o971x4D8OVav1kfmOpHI6OeEs
 Content-Type: text/plain; charset=windows-1252
 Content-Transfer-Encoding: quoted-printable
 
 
 
 On 07.10.2016 22:10, Joerg Sonnenberger wrote:
 > The following reply was made to PR pkg/51534; it has been noted by GNAT=
 S.
 >=20
 > From: Joerg Sonnenberger <joerg%bec.de@localhost>
 > To: gnats-bugs%NetBSD.org@localhost
 > Cc: pkg-manager%netbsd.org@localhost, gnats-admin%netbsd.org@localhost, pkgsrc-bugs@netbsd.=
 org,
 > 	kuehro%posteo.de@localhost
 > Subject: Re: pkg/51534: qemu-2.7.0 fails to build: bad register name
 > Date: Fri, 7 Oct 2016 22:06:48 +0200
 >=20
 >  On Fri, Oct 07, 2016 at 07:20:00PM +0000, Kamil Rytarowski wrote:
 >  >  2. changing from append to prepend is no-operation for qemu
 > =20
 >  That's rather obviously wrong as the patch wouldn't work if there was =
 no
 >  difference.
 > =20
 >  >  3. changing from append to prepend might create new hard to debug i=
 ssues
 >  >  (like ABI mismatch) in fragile makefiles as the last -mABI wins and=
  it
 >  >  won't be corrected
 > =20
 >  This is exactly the intended goal. If a package explicitly specifies a=
 n
 >  ABI, it should win. That's the whole point of the proposal to move to
 >  prepend.
 > =20
 >  >  This is why I prefer to let some software to have specified/injecte=
 d
 >  >  ${_COMPILER_ABI_FLAG.${ABI}} where needed. And my '-1" patch must b=
 e
 >  >  extended to pass ${_COMPILER_ABI_FLAG.${ABI}} accordingly. Changing=
 
 >  >  things to prepend just hides the issue. For now just qemu is affect=
 ed
 >  >  and I wouldn't reinvent new way to inject -mABI for every package f=
 or
 >  >  singular exceptions.
 > =20
 >  qemu needs special help because it wants to build a BIOS image, which
 >  has special ABI requirements. The *normal* qemu binary is still suppos=
 ed
 >  to have the *normal* pkgsrc ABI.
 > =20
 >  Joerg
 > =20
 >=20
 
 This is correct. Prepending -mABI doesn't help, it will break on
 customized pkgsrc ABI, as configure will pickup its own ABI and enforce i=
 t.
 
 
 --90JSrRx1o971x4D8OVav1kfmOpHI6OeEs--
 
 --jQqjliatJ4KqMe5MLf4CifjIpNvjsR0S6
 Content-Type: application/pgp-signature; name="signature.asc"
 Content-Description: OpenPGP digital signature
 Content-Disposition: attachment; filename="signature.asc"
 
 -----BEGIN PGP SIGNATURE-----
 Version: GnuPG v2
 
 iQIcBAEBCAAGBQJX+ASQAAoJEEuzCOmwLnZs3RUP/2Us4BTIYHvLchnEoBsSUHrm
 3FCLM3lz0wG0xKZwTof1pwtwTU4Jd9pi2L9aUlktefQ9D+WkPNxUFkfbTf3MLcD9
 3zmGz84mVsMkzXUDm+NDF7EiTFtPPwWDIOQcVr2FXQRsxOtx7qRhAEQ6XIinB6Th
 1SbyMc/hcKl10ZY/kvy3T2tAi4LlINXDAbLjEXcuuUWzHXf0/MRWLjsEYSsH6Clx
 qZVkp1JsMgE6dR//87lhWxNwE+InD12wE00z+9o8Elh71cCnJ52LN4DzfCzLrURb
 8TZokJA2u7K2p9zMsdbOpKAHM56qce6NLj11NaPlsOsKjYKTyuj1RX6F6dQlgzig
 6MczFk0SzaLUmRZlxJ11yeXMTpI90+PA+1i8FnwKFJi/vyyS/a5/tHXbTEyrqa//
 AhW663L99RL0j89L6QHqQHuqFJQL43Oo12tTLa+MGbwUF9cZUpHy/xQXLn6Q95rp
 et+XIP1K+stLwJvRbpvuD2d+hm64291QKw5IpjD9unoW4mbnefmxIg+STgOqJpoZ
 OmG2pNIQ19u0yOypyEjdV7vnlhhQcjzYoDZYsLCyRKWWxv/J8M3r5hsvd86zwxP2
 0eHWKDFZ9sxqrjLYu8GxIqAHOQdmSnASBgPHQCiGichOulDwqJsPqWk+9H9F7K4k
 oIv6NNOmg+y1JqznJKRB
 =1Yok
 -----END PGP SIGNATURE-----
 
 --jQqjliatJ4KqMe5MLf4CifjIpNvjsR0S6--