Date: Thu, 14 Mar 2019 23:02:06 +0000 From: Mark Raynsford <list+org.freebsd.java@io7m.com> To: Greg Lewis <glewis@eyesbeyond.com> Cc: freebsd-java@freebsd.org Subject: Re: JDK 11 update Message-ID: <20190314230206.6ea05f51@almond.int.arc7.info> In-Reply-To: <20190314214648.GA38669@misty.eyesbeyond.com> References: <20190308180402.GA61500@misty.eyesbeyond.com> <20190314105052.56b1c6c3@almond.int.arc7.info> <20190314214648.GA38669@misty.eyesbeyond.com>
next in thread | previous in thread | raw e-mail | index | archive | help
--Sig_/zq9J976NOlKA792yjvP/vtc Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: quoted-printable On 2019-03-14T14:46:48 -0700 Greg Lewis <glewis@eyesbeyond.com> wrote: > > * Talk to AdoptOpenJDK folk to get more details of what the need in terms > of machines/VMs. What sort of resources do their machines have > (CPU/memory/disk)? What sort of access does the CI/build infrastructure > need? Is this a case of supplying machines/VMs or would it make more > sense to fund them in their existing build farm? I've pointed them at this thread. Hopefully I'll get more information soon. > * Based on that come up with a rough costing and put it to the FreeBSD > Foundation. You might even want to talk to them earlier and see if > it was something they were even interested in at all. They have > sponsored Java work in the past, but that doesn't mean it fits into > their current priorities. > > FWIW, changes to the existing openjdk8 and openjdk11 code bases don't > appear to be all that frequent. So dedicated machines would likely just > sit idle most of the time. Using EC2 on demand instances, assuming > we have a suitable FreeBSD AMI, and storing the releases on S3 would > make sense to me, and is likely fairly inexpensive. Right. I'd guess costs would be pretty minimal. > My biggest question would be what value does this add versus the existing > FreeBSD package infrastructure? That will also feature binary packages > for all supported FreeBSD versions and costs neither resources nor time > to set up other maintaining the FreeBSD port (which has to be done anyway= ). The main advantage for me personally is that I'd be able to fetch runtime images supporting FreeBSD for use in jlink. My ultimate intention is to put together a library/command-line tool/Maven plugin that can automatically fetch JVM runtime images for a range of architectures and operating systems and produce a jlinked application image for every specified platform. "Build once, run on a selection of platforms", sort of thing. The first part of this work is already done: https://blog.io7m.com/2018/11/18/choosing-coffee.xhtml https://github.com/io7m/coffeepick As you can probably imagine... None of the supported providers supplies FreeBSD JVM runtime images. I could, with some hacking, grab the FreeBSD binary packages and then unpack and reorganize the contents... But it'd be a lot nicer if FreeBSD was a platform that worked just like all of the others supported by AdoptOpenJDK and friends. More people building and running the test suite can't hurt. --=20 Mark Raynsford | http://www.io7m.com --Sig_/zq9J976NOlKA792yjvP/vtc Content-Type: application/pgp-signature Content-Description: OpenPGP digital signature -----BEGIN PGP SIGNATURE----- iHUEARYKAB0WIQTjeVlXFhMdtOerh91nsYzB+azjRQUCXIrdbgAKCRBnsYzB+azj RZAYAQCaoB9+JqOygP1/p5eFpjEr7mSZtx+uoA3MS8vhNV93tQD+J5bb9uTnaVWv XPkUJLGraZOzbtPQCtvfLuDRjZ8ItQ4= =9GF/ -----END PGP SIGNATURE----- --Sig_/zq9J976NOlKA792yjvP/vtc--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20190314230206.6ea05f51>