Date: Thu, 19 Nov 2020 15:30:14 -0800 From: Bryan Drewery <bdrewery@FreeBSD.org> To: Jilles Tjoelker <jilles@stack.nl>, Mathieu Arnold <mat@FreeBSD.org>, svn-ports-all@freebsd.org Cc: ports-committers@freebsd.org, svn-ports-head@freebsd.org, =?UTF-8?Q?Dag-Erling_Sm=c3=b8rgrav?= <des@FreeBSD.org> Subject: Re: svn commit: r554893 - head/Mk/Scripts Message-ID: <3568316e-f84c-984d-7d67-cd09a46c5742@FreeBSD.org> In-Reply-To: <20201119230309.GA10938@stack.nl> References: <202011111329.0ABDTqUD035770@repo.freebsd.org> <dac04210-4580-d5ab-49e0-c1b501ea7ee7@FreeBSD.org> <20201119230309.GA10938@stack.nl>
next in thread | previous in thread | raw e-mail | index | archive | help
This is an OpenPGP/MIME signed message (RFC 4880 and 3156) --RC1lmUz7EBdPmReWQJi2J0fUx1AyVD6Wx Content-Type: multipart/mixed; boundary="dhSY4123DpLhUsaLrSVFYadNEoRbRx72e"; protected-headers="v1" From: Bryan Drewery <bdrewery@FreeBSD.org> To: Jilles Tjoelker <jilles@stack.nl>, Mathieu Arnold <mat@FreeBSD.org>, svn-ports-all@freebsd.org Cc: ports-committers@freebsd.org, svn-ports-head@freebsd.org, =?UTF-8?Q?Dag-Erling_Sm=c3=b8rgrav?= <des@FreeBSD.org> Message-ID: <3568316e-f84c-984d-7d67-cd09a46c5742@FreeBSD.org> Subject: Re: svn commit: r554893 - head/Mk/Scripts References: <202011111329.0ABDTqUD035770@repo.freebsd.org> <dac04210-4580-d5ab-49e0-c1b501ea7ee7@FreeBSD.org> <20201119230309.GA10938@stack.nl> In-Reply-To: <20201119230309.GA10938@stack.nl> --dhSY4123DpLhUsaLrSVFYadNEoRbRx72e Content-Type: multipart/mixed; boundary="------------A8C5A05EC0328B6279B2CFCD" Content-Language: en-US This is a multi-part message in MIME format. --------------A8C5A05EC0328B6279B2CFCD Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable On 11/19/2020 3:03 PM, Jilles Tjoelker wrote: > On Thu, Nov 19, 2020 at 11:41:57AM -0800, Bryan Drewery wrote: >> On 11/11/2020 5:29 AM, Mathieu Arnold wrote: >>> Author: mat >>> Date: Wed Nov 11 13:29:52 2020 >>> New Revision: 554893 >>> URL: https://svnweb.freebsd.org/changeset/ports/554893 >>> Log: >>> Add set pipefail in most framework scripts. >>> set pipefail changes the pipeline return status from being the retu= rn >>> status of the last command to the last non 0 exit status of any com= mand >>> in the pipeline. This is needed to make sure all the commands in a= >>> pipeline did actually return a non 0 status and not only the last o= ne. >> [snip] >>> Modified: head/Mk/Scripts/check-desktop-entries.sh >>> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D >>> --- head/Mk/Scripts/check-desktop-entries.sh Wed Nov 11 13:24:31 2020= (r554892) >>> +++ head/Mk/Scripts/check-desktop-entries.sh Wed Nov 11 13:29:52 2020= (r554893) >>> @@ -4,6 +4,7 @@ >>> # MAINTAINER: portmgr@FreeBSD.org >>> =20 >>> set -e >>> +set -o pipefail >>> =20 >>> . "${dp_SCRIPTSDIR}/functions.sh" >>> =20 >>> >> This can prevent someone from upgrading from an unsupported release. T= he >> workaround is simple enough so I think we should take it rather than >> create burdens for people. >> command set -o pipefail 2>/dev/null || : > Hmm, does an upgrade require building ports on an old release? I meant more about upgrading specific ports while having an unsupported FreeBSD release. As we move towards pkgbase it [ports compat policy] really could block full system upgrades. >=20 > When scripts are written for use with 'set -o pipefail', allowing them > to run without it seems unwise. Very good point. Thank you for the rest of your response too. directed in general: Pipefail came in 12.1 which released just 1 year ago. I choose to use a 12.0 jail to ensure compatibility for users. If we are going to make it harder on users it should be worth it and for something like pipefail there is often an workaround. The reality is not everyone is on a supported FreeBSD version but if they want to upgrade their ports why stop them? I despise us making draconian decisions about what our users are allowed to do. I think it is fair to warn them about unsupported FreeBSD version, not block, before a build and then yell at them on build failure. I think it is also fair to limit patches to supported releases but rude to drop them on EOL day. Many ports would still work fine. Breaking the framework is different. --=20 Regards, Bryan Drewery --------------A8C5A05EC0328B6279B2CFCD-- --dhSY4123DpLhUsaLrSVFYadNEoRbRx72e-- --RC1lmUz7EBdPmReWQJi2J0fUx1AyVD6Wx Content-Type: application/pgp-signature; name="OpenPGP_signature.asc" Content-Description: OpenPGP digital signature Content-Disposition: attachment; filename="OpenPGP_signature" -----BEGIN PGP SIGNATURE----- wsB5BAABCAAjFiEE+Rc8ssOq6npcih8JNddxu25Gl88FAl+3AAYFAwAAAAAACgkQNddxu25Gl88t bgf/emwB+QH0XGVryr9MumMVoJusWHqDnwCvKGFk5Q1EIm93Rm00TdF7WnR4hZsvSapfoMiutcRa MpJwMbafWcR5VGBlfAnqQ9g0y+eTH8QbtwUlyhwAxKm1ktjNWvqQAZ4SVR/hhUtbkwIw0yjzMsZF ihesGd1EK0ZIdoy734xa6SD4nd75KpFuSk8KhuIBdW28KZbj7pGXC0FU99jgbrpLOfGUyAAuf1/k HR2lcI40l2hVIwoUiDrkPwACBK/pD5c5Ssqz2HjCoTGt4k37QHrsY/D5huSFHHzt8qcJW8Uz1sXN 9+cqePYZDDkSTM5alfG9/mreWt5lxtvMK0/oz+YMIA== =z3rD -----END PGP SIGNATURE----- --RC1lmUz7EBdPmReWQJi2J0fUx1AyVD6Wx--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3568316e-f84c-984d-7d67-cd09a46c5742>