On Wed, Dec 19, 2018 at 12:37:41AM +0000, Ramsay Jones wrote:
>
>
> On 17/12/2018 23:57, Luc Van Oostenryck wrote:
> > Sparse pre-release v0.6.0-rc1 is now pushed out and contains
> > 850 patches made by 10 people.
> >
> > The source code can be found at its usual repository:
> > git://git.kernel.org/pub/scm/devel/sparse/sparse.git v0.6.0-rc1
>
> I have tested v0.6.0-rc1, as usual, on LM-19 64-bit, LM-18.3 32-bit,
> cygwin 64-bit and fedora-28 64-bit. (Both sparse and its use with git).
> No errors or regressions found!
Great!
> Also, I have looked at the 'dump macros' output:
>
> $ gcc -m64 -dM -E - </dev/null | sort >ggg-64
> $ gcc -m32 -dM -E - </dev/null | sort >ggg-32
> $ gcc -mx32 -dM -E - </dev/null | sort >ggg-x32
> $ gcc -m16 -dM -E - </dev/null | sort >ggg-16
> $ gcc -dM -E - </dev/null | sort >ggg
> $ ./sparse -m64 -dM -E - </dev/null | sort >sss-64
> $ ./sparse -m32 -dM -E - </dev/null | sort >sss-32
> $ ./sparse -mx32 -dM -E - </dev/null | sort >sss-x32
> $ ./sparse -m16 -dM -E - </dev/null | sort >sss-16
> $ ./sparse -dM -E - </dev/null | sort >sss
> $
>
> on each 64-bit platform:
>
> $ diff ggg ggg-64
> $ diff sss sss-64
> $ meld ggg sss
> $
>
> ... has shown no issues.
>
> on the only 32-bit platform:
>
> $ diff ggg ggg-32
> $ diff sss sss-32
> $ meld ggg sss
> $
>
> ... has also shown no issues.
Excellent, thanks you for that.
I've also done essentially the same on some recent Ubuntu 32 & 64-bit,
an older Debian 64-bit, some Ubuntu on ARM64, CentOS on ppc64 & ppc64el,
and some tests on Debian with sparc64 and Solaris 10 & 11 on sparc32.
I've also checked it with just a cross-compiler for the different ABI
on mips32 & mips64, RISC-V 32 & 64-bit and S390X but these may depends
on how the compiler was configured when built).
I'll give a try later on some {Net,Free,Open}BSD and ARM32.
> Yes, I have seen some issues in some pairs not compared above, but
> I don't think that should hold up the release. (I haven't studied
> _all_ of the above pairs yet - I wanted to report what I had found
> so far tonight ...).
Thank you very much.
I'm happy to see that cygwin 64-bit is OK (as it's one I don't test).
> [Note: I tested specifically commit 105e081 (ie the tag v0.6.0-rc1)
> because I had already started testing, before seeing commit 1cf02bb]
Yes, sure, 1cf02bb is just a small fix with no functional effect.
> > The tarballs are found at:
> > https://www.kernel.org/pub/software/devel/sparse/dist/
>
> Ah, yes, I never think about testing the tarballs! (I guess that is
> what prompted commit 1cf02bb?).
I just add them because not everyone is using git but well ...
It's more *because* of them that I forgot to update the version in
the Makefile.
Best regards thanks again
-- Luc
