Re: Bug#839840: ghci segfaults on armel, related to doctest failure

[Niels Thykier <niels@thykier.net>]

On Fri, 7 Oct 2016 11:10:50 +0200 Sven Bartscher
<sven.bartscher@weltraumschlangen.de> wrote:
> On Wed, 5 Oct 2016 17:12:45 +0000
> Clint Adams <clint@debian.org> wrote:
> 
> > Source: ghc
> > Version: 7.10.3-9
> > Severity: serious
> > 
> > clint@abel ~/haskell-http-api-data-0.2.4
> >  % ghci -isrc -XCPP -D"MIN_VERSION_base(X,Y,Z)=1"
> > GHCi, version 7.10.3: http://www.haskell.org/ghc/  :? for help
> > Prelude> :l src/Web/HttpApiData.hs  
> > [1 of 2] Compiling Web.HttpApiData.Internal ( src/Web/HttpApiData/Internal.hs, interpreted )
> > [2 of 2] Compiling Web.HttpApiData  ( src/Web/HttpApiData.hs, interpreted )
> > Ok, modules loaded: Web.HttpApiData, Web.HttpApiData.Internal.
> > *Web.HttpApiData> :set -XOverloadedStrings
> > *Web.HttpApiData> import Control.Applicative  
> > *Web.HttpApiData Control.Applicative> import Data.Time
> > *Web.HttpApiData Control.Applicative Data.Time> import Data.Int
> > *Web.HttpApiData Control.Applicative Data.Time Data.Int> import Data.Text (Text)
> > *Web.HttpApiData Control.Applicative Data.Time Data.Int Data.Text> import Data.Time (Day)
> > *Web.HttpApiData Control.Applicative Data.Time Data.Int Data.Text Data.Time> import Data.Version
> > *Web.HttpApiData Control.Applicative Data.Time Data.Int Data.Text Data.Time Data.Version> toUrlPiece True
> > "true"
> > *Web.HttpApiData Control.Applicative Data.Time Data.Int Data.Text Data.Time Data.Version> parseUrlPiece "false" :: Either Text Bool
> > zsh: segmentation fault  ghci -isrc -XCPP -D"MIN_VERSION_base(X,Y,Z)=1"
> > 
> > The backtrace is useless.
> 
> This bug I reported upstream[1] might be related to that.
> 
> Regards
> Sven
> 
> [1]: https://ghc.haskell.org/trac/ghc/ticket/11831

Hi,

We got a double RC buggy ghc on armel (this bug plus #851373, which also
affects armhf).  Unfortunately, I cannot see any progress on these bugs
(this one has been without updates since Oct 2016 and the other one
since Jan 2017).

It would be very unfortunate to release with packages we cannot compile
(any longer), so I would prefer this being fixed or as a last resort,
drop of ghc and all rdeps on armel+armhf.

Ideally, this resolution should happen fairly soon, so it does not end
up stalling the release.

Thanks,
~Niels