[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: test results for 32-bit platforms
From: |
Jose E. Marchesi |
Subject: |
Re: test results for 32-bit platforms |
Date: |
Sun, 07 Mar 2021 23:59:56 +0100 |
User-agent: |
Gnus/5.13 (Gnus v5.13) Emacs/28.0.50 (gnu/linux) |
>> I was waiting for having all the tests passing in 32-bit hosts ;)
>
> Cool! So, here are test results for various 32-bit platforms (assuming
> the patches that I sent this week-end).
>
> Builds fine and runs fine on:
>
> Linux/i386
> Linux/arm
> Linux/mips
> Linux/powerpc
> Linux/sparc
> Linux/s390
>
> Hurd 2019/i386
>
> macOS 10.5/i386
> FreeBSD 12/i386
> NetBSD 8/i386
>
> Solaris 10/i386
> Solaris 10/sparc
> Solaris 11/i386
> Solaris OpenIndiana/i386
Ueh! :D
> Now the trouble:
Uoh :(
> 1) On Linux/arm (real hardware, not emulated) I once saw this test failure:
>
> Setting LD_LIBRARY_PATH to :/usr/arm-linux-gnueabihf/lib
> spawn [open ...]^M
> Foo {a=struct {a1=0x0UH,a2=0x0UH}}
> WARNING: program timed out.
> FAIL: poke.pkl/scons-int-struct-1.pk execution test
>
> Not reproducible, but still worrisome if some test just hangs.
Yes, it is indeed worrisome. I wonder how can we hunt this.
> 2) Boehm GC not working on several platforms:
>
> OpenBSD 6.0/i386 GC test hangs
> Solaris 11/sparc GC tests fail (both with cc and gcc)
> AIX/powerpc GC has build failures
> Minix 3.3 GC (--disable-threads) not yet ported
> Haiku GC test crashes
We really gotta get rid of the boehm GC.
Luca, how is yours doing? ;)