Date: Sat, 17 Jul 2021 10:28:24 +0200 From: Guido Falsi <madpilot@FreeBSD.org> To: Alexey Dokuchaev <danfe@freebsd.org> Cc: ports-committers@freebsd.org, dev-commits-ports-all@freebsd.org, dev-commits-ports-main@freebsd.org Subject: Re: git: dc9bf7d64926 - main - net/asterisk*: Add aarch64 support Message-ID: <d96049d2-15e9-1088-1449-0ef05d645a53@FreeBSD.org> In-Reply-To: <YPBpeWTJvGwMv1FL@FreeBSD.org> References: <202107141033.16EAX60T044972@gitrepo.freebsd.org> <YO%2BYP1TDjq6MruvC@FreeBSD.org> <81bc6b76-7b74-9990-d7dc-54ca14b0ee4f@FreeBSD.org> <YPBpeWTJvGwMv1FL@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On 15/07/21 18:59, Alexey Dokuchaev wrote: > On Thu, Jul 15, 2021 at 04:38:53PM +0200, Guido Falsi wrote: >> On 15/07/21 04:06, Alexey Dokuchaev wrote: >>> ... >>> I think ONLY_FOR_ARCHS here clearly outlived its purpose and can just >>> be removed. It arguably misses MIPS, but we aren't promising much for >>> that tier anyways, and this harness just gets in the way of those who >>> know that they're doing. After all, there shouldn't be anything >>> inherently architecture-specific about the telephony toolkit, but even >>> if there is, individual BROKEN_$arch would be more appropriate. >> >> The fact that mips is not included is exactly the reason I did not >> remove ONLY_FOR_ARCHS. >> >> If what you say is the general consensus I'll be quite happy to get rid >> of it. > > You'd likely not get general consensus as most people don't care. :-) Maybe I did not choose my words wisely. More than consensus I was looking for a "no objection" consensus. > >> I guess this compiler mix situation is not an issue anymore. > > Right. I think it's generally bad idea to prematurely restrict software > to certain arches unless it's clearly arch-specific (e.g. comes only in > binary precompiled form or uses asm code). New arches appear frequently > (e.g. powerpc64le, riscv64) and some go away as well (ia64, sparc64), it > just does not look feasible to maintain those ONLY_FOR_ARCHS lists so > they'd always reflect the reality. > > Even if one's port does break on MIPS, one should get a notice from the > build cluster and either fix it, or mark is as BROKEN_$arch, rather than > keep ONLY_FOR_ARCHS which would only get amended, causing needless repo > churn, upon every next time someone discovers that it builds and works > on their architecture of interest. > I fundamentally agree with what you say so I've followed your suggestion and removed the restriction. -- Guido Falsi <madpilot@FreeBSD.org>
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?d96049d2-15e9-1088-1449-0ef05d645a53>