OK, I have a status update on this. * I can't get a minimal patch set against mono-2018-06 branch (Mono 5.16) which builds the entire release to completion on s390x on versions of gcc with PIC/PIE by default (Debian gcc-6 or later) * I picked the version I uploaded, on the basis that it is part of the `vs` packaging branch on mono-project.com, which mirrors what last went into a stable release of Visual Studio for Mac (and hopefully should have something resembling a vendor support story) * I _can_ get a minimal patch set against mono-2018-08 (Mono 5.18) * 5.18 Should not cause any major changes at a packaging level, compared to 5.16 (i.e. it should not go into binary NEW) * 5.18 is considered "stable" by the Mono runtime team, but not yet by the Visual Studio for Mac team (the standalone runtime doesn't get QA before releases, only as part of a larger VSMac release cycle) * 5.18 might totally break on other architectures! There is no MIPS testing upstream, so mipsel might be broken on any given release (upstream builds for i386, armel, armhf, arm64, amd64, and ppc64el)
I don't want to preempt any product announcements, but 5.18 may ship significantly faster than originally planned, meaning investing too much time in fixing up 5.16 this week may be wasted, compared to 5.18 closer to *mumble mumble*. I'll discuss with relevant stakeholders tomorrow. On 15/01/2019 07:37, Jo Shields wrote: > Thanks for the tracking bug. I spotted this last night and am working on > zelenka to convince the damn thing to build - it's something to do with PIE, > and only shows up as a problem with the Runtime crashing in the literal last > 30 seconds of the build step > > Sent from my iPhone > >> On 15 Jan 2019, at 05:45, Matthias Klose <d...@debian.org> wrote: >> >> Package: src:mono >> Version: 5.16.0.220+dfsg3-2 >> Severity: serious >> Tags: sid buster >> >> mono won't migrate to testing, still b-d on gcc-5 on s390x. >> >
