From nobody Thu Sep 21 13:27:28 2023 X-Original-To: freebsd-ports@mlmmj.nyi.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2610:1c1:1:606c::19:1]) by mlmmj.nyi.freebsd.org (Postfix) with ESMTP id 4Rrx5t0P6hz4v8Ph for ; Thu, 21 Sep 2023 13:27:30 +0000 (UTC) (envelope-from brooks@spindle.one-eyed-alien.net) Received: from spindle.one-eyed-alien.net (spindle.one-eyed-alien.net [199.48.129.229]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (4096 bits) server-digest SHA256) (Client did not present a certificate) by mx1.freebsd.org (Postfix) with ESMTPS id 4Rrx5s5vRQz3TGm; Thu, 21 Sep 2023 13:27:29 +0000 (UTC) (envelope-from brooks@spindle.one-eyed-alien.net) Authentication-Results: mx1.freebsd.org; none Received: by spindle.one-eyed-alien.net (Postfix, from userid 3001) id DF7E63C019A; Thu, 21 Sep 2023 13:27:28 +0000 (UTC) Date: Thu, 21 Sep 2023 13:27:28 +0000 From: Brooks Davis To: Gleb Popov Cc: Mark Millard , FreeBSD Mailing List Subject: Re: devel/llvm10: deprecate, set short expiration ( lang/ghc810 has LLVM_VERSION= 10 ) Message-ID: References: <3925E646-9214-46FF-A01E-1598473FB547@yahoo.com> List-Id: Porting software to FreeBSD List-Archive: https://lists.freebsd.org/archives/freebsd-ports List-Help: List-Post: List-Subscribe: List-Unsubscribe: Sender: owner-freebsd-ports@freebsd.org X-BeenThere: freebsd-ports@freebsd.org MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: X-Spamd-Bar: ---- X-Rspamd-Pre-Result: action=no action; module=replies; Message is reply to one we originated X-Spamd-Result: default: False [-4.00 / 15.00]; REPLY(-4.00)[]; ASN(0.00)[asn:36236, ipnet:199.48.128.0/22, country:US] X-Rspamd-Queue-Id: 4Rrx5s5vRQz3TGm On Thu, Sep 21, 2023 at 03:08:43PM +0300, Gleb Popov wrote: > On Thu, Sep 21, 2023 at 2:28???PM Brooks Davis wrote: > > > > This seems to be an arm-only thing. What's going on here? Do we have to > > keep llvm10 around forever just for ghc on arm? > > > > -- Brooks > > > > GHC has a native code generator and a LLVM-based one. On ARM the > native generator doesn't work with old GHC's, so the only way to have > Haskell working is via LLVM. At the same time the LLVM backend can > still be used even with new GHC and on amd64. > > Removing llvm10 would remove lang/ghc810 and its dependents from the > ARM package list and would also prevent bootstrapping newer GHC's on > this platform. Luckily, we already have GHC 8.10 bootstrapped for ARM, > which still makes it possible to build lang/ghc92 and proceed from > there. GHC 9.2 uses LLVM 12, which I hope will stay for a while. > > That being said, I don't really have an opinion on LLVM 10 removal. I > don't have time to keep up with Haskell stuff anymore so I don't want > to drag other into maintaining more old software. One compromise position would be to dial back the default options in llvm10 to only those required by ghc (or just llvm+clang+lld). It probably doesn't make sense to confine it to arm only as it would be missed in exp-runs. -- Brooks