Date: Tue, 21 Jan 2020 23:48:34 -0500 From: "Phillip R. Jaenke" <prj@rootwyrm.com> To: ports <freebsd-ports@freebsd.org>, mono@freebsd.org Subject: lang/mono: testers wanted for 6.8.0.96 update! Message-ID: <99cfb7f3-f600-1dbd-bcb5-41be270b706b@rootwyrm.com>
next in thread | raw e-mail | index | archive | help
All; Boy, it sure feels like forever ago that I said "hey, folks, what do you think about updating past mono 5.10" doesn't it? Oh. It was. But I hope that you'll all find the wait was worth it. I've been hard at work along with a lot of assistance from the Mono team themselves to get things not just into a good place, but an actually great place. What do I define as a great place? Actual CI testing, no patches required to build[1], no missing features, no actually failing tests, and full functionality. Well, here is the official Mono community-supported CI. https://jenkins.mono-project.com/job/test-mono-mainline-community/label=freebsd-12-amd64/ - ISSUE: System.Web (affects XNS consumers) sometimes experiences kevent failures on slow systems. Please let me know directly if you have a reproducible case on a fast amd64 system, rather than opening an upstream bug. - ISSUE: System.Web.Extensions has a known crash but has not been investigated yet. The crash isn't consistently reproducible. If you have an application that can consistently reproduce a crash, please do open an issue at github.com/mono/mono directly. - ISSUE: Mono.Posix fails when testing multiple control messages to a single socket. Since this is a 'fails because it succeeds' it likely actually isn't an error. (Please let me know if you know otherwise!) - NON-ISSUE: The CI reports runtime instability; this is actually a reporting error. The runtime tests all pass. - NON-ISSUE: corlib fails certain FileSystemInfoTests; these are expected failures and can safely be ignored. The tests will be fixed in the future. - NON-ISSUE: Mono.Posix IPv6 tests will register a failure because getsockname() works on FreeBSD. This is a bug in the test. - NON-ISSUE: Mono.Profiler.Log-xunit indicates failures in the CI; profiling should work as expected. This is a CI issue. - NON-ISSUE: All known network code failures in CI are OS-specific tests for other OSes. So, you ask now, how do I help test this lang/mono? I would like to know more, you say! I have important feedback! Just go here: https://reviews.freebsd.org/D2330 Your feedback as a Mono consumer (either as maintainer of a port that uses Mono or as a user of a Mono-based application) is VERY important to me, because I only know a VERY small set of Mono applications. And it likely doesn't include yours. I want to make sure that this update goes great for everyone, even if you just use an application. I mentioned improvements, and we really have made some tremendous strides that I want to go over briefly so that you're not sitting there going "wait what is going on here?" OPTIONS for lang/mono: MONOLITE INOTIFY* NINJA NLS ODBC SQLITE X11 DEVELOPER DEFAULTS for lang/mono: MONOLITE SQLITE X11 (Note: INOTIFY is currently broken and being worked on upstream.) - NINJA is only used to use Ninja for builds. May help on slow systems. - NLS is, well, it's NLS - ODBC provides full native unixODBC support - SQLITE provides full native SQlite3 support - X11 is of course graphics support - DEVELOPER is intended for Mono, Roslyn, and .NET developers and disables certain privacy functions to allow examination of crash traces. * INOTIFY is a fully optional package and is not required for most users; actual kqueue has been implemented in Mono. (Please direct your applause to the Xamarin and Microsoft teams.) Thank you all for your patience, thank you to everyone on the Mono and Dotnet Runtime team for putting up with my 'I really don't know C#, I just wanna run precompiled bins' questions, and a massive thank you to the Mono Gitter crew (we know who we are) for all their invaluable help. And last but not least, thank YOU for any help YOU can provide. Even if it's just a "my application works faster and doesn't crash," that can be very helpful information. Thanks and happy... uh... well... running binaries compiled on Windows intended for Linux on your god-less FreeBSD infrastrcutre! \o/ -Phillip R. Jaenke | prj@rootwyrm.com http://www.rootwyrm.com/ p.s. please email me directly as I am not subbed to the mono@ list [1] Okay, ONE patch due to an autotools issue resulting from crossing up a fix we made in CoreFX for libinotify. The full fix will be upstreamed Very Soon(TM) I promise.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?99cfb7f3-f600-1dbd-bcb5-41be270b706b>