[libcxx-commits] [PATCH] D81602: [libc++] Provide SEM_VALUE_MAX fallback on Solaris

Mon Sep 6 02:59:51 PDT 2021

ro added a comment.

In D81602#2982911 <https://reviews.llvm.org/D81602#2982911>, @ldionne wrote:

> Is there interest for pushing this forward? We have instructions for setting up workers for your platform here: https://libcxx.llvm.org/AddingNewCIJobs.html#addingnewcijobs
>
> Once this is done, I'd be happy to help you with reviewing patches that expand support to Solaris and make sure it doesn't bit rot like the previous port.

I'm pretty certain I won't move this (and my subsequent patches I hadn't submitted) forward any time soon for several reasons:

- I've spent way too much time in LLVM in recent months, with limited gain, meanwhile massively neglecting my Solaris GCC maintainer duties.
- buildkite hasn't even been ported to Solaris or Illumos AFAICT and I have very little inclination of doing this and adding yet another buildbot system to my mix.
- I can't tell for certain how libcxx CI works, but if LLVM's buildbot is any indication, the agent is useless unless you manage to not only make libcxx buildable on your target (which I did for Solaris), but also have no failures in the testsuite.  When I stopped working on my port more than a year ago, I was at 20 failures for libcxx, plus 6 for libunwind and 1 for libc++abi.
- Last but not least, I would only be able to handle the Solaris side of things.  Given that Solaris and Illumos have diverged considerably especially in areas that affect libcxx (like locale support), any patch I would add is likely to break Illumos (even further).  When I reached out to handle the same problem for clang and compiler-rt, there has been almost no feedback from the Illumos community.

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D81602/new/

https://reviews.llvm.org/D81602