ghc 9.4.7 needs some manual bootstrap on hppa

[Adrian Bunk]

FYI:

- haskell-quickcheck/hppa did FTBFS with ghc 9.4.6 due to #1052306
- ghc >= 9.4.7-1~exp1 uses the the Hadrian build system (due to #1051493)
- haskell-hadrian/hppa is uncompiled, blocked by haskell-quickcheck

cu
Adrian

[Adrian Bunk]

On Tue, Oct 24, 2023 at 04:01:57PM +0200, John Paul Adrian Glaubitz wrote:

> There have been a number of regressions in GHC upstream unfortunately and despite
> me spending hours over hours into getting those sorted out, I have only been able
> to get GHC to work on sparc64 again.
>...

9.4.6 did build on hppa (although not at first attempt),
breaking the dependency cycle might be enough to get 9.4.7.

> Adrian

cu
Adrian

[John David Anglin]

How did i386 work around this issue?

I can try building haskell-quickcheck with tests disabled.  But if it is seriously broken,
then this isn't going to work.

Dave

--
John David Anglin dave....@bell.net

[Adrian Bunk]

On Tue, Oct 24, 2023 at 10:21:05AM -0400, John David Anglin wrote:
> On 2023-10-24 9:27 a.m., Adrian Bunk wrote:
> > FYI:
> >
> > - haskell-quickcheck/hppa did FTBFS with ghc 9.4.6 due to #1052306
> > - ghc >= 9.4.7-1~exp1 uses the the Hadrian build system (due to #1051493)
> > - haskell-hadrian/hppa is uncompiled, blocked by haskell-quickcheck
> How did i386 work around this issue?

https://tracker.debian.org/news/1471018/accepted-haskell-hadrian-947-3-source-i386-into-unstable/
Architecture: source i386

(I don't know how that binary was built)

> I can try building haskell-quickcheck with tests disabled.  But if it is seriously broken,
> then this isn't going to work.

The issue fixed by [1] did not break everything on i386, but I cannot
really judge it.

> Dave

cu
Adrian

BTW: Please Cc me on replies.

[1] https://gitlab.haskell.org/ghc/ghc/-/merge_requests/9778