[Debian-med-packaging] [Python-modules-team] Any reason why python-numpy version 1.9.2 is in experimental?

[Kevin Murray]

Andreas,
Andreas,

On 16:20 04/08, Andreas Tille wrote:
> Hi Kevin,
> 
> it seems I created more trouble in yesterdays update than just ignoring
> your push.  I was not aware of #794511 and that it might cause some
> delay on the python-numpy 1.9.2 migration to unstable.  We now have
> several options:
> 
>   1. Do nothing and wait until #794511 will be fixed.  This will put an
>      RC bug on python-skbio 0.4.0 since it will not build from source.
>   2. Upload to experimental and remove from unstable.
>   3. Use an epoch and upload 1:0.2.3-1 to unstable
> 
> I'm not happy about either option (shame on my again for doing things
> when beeing tired).  Since bug #794511 is a libc++6 transition issue
> this might be resolved not that far in time and we choose option 1.
> 
> What do you think
> 
>       Andreas.
> 

I reckon that either option 1 or 3 or some combination of the two would be the
best option. Is it a bad idea to have skbio 0.2.3 in unstable now (for QIIME's
sake at least), and deal with 0.4.0 when numpy becomes installable? I am
running sid here on my laptop, and there's a whole heap of uninstallable
updates due to the libc++6/gcc5 issues, so I don't think we'll be alone
whatever happens.

Cheers,
K

---
Kevin Murray

GPG pubkey: http://www.kdmurray.id.au/static/A4B4EE6A.asc
FPR: 656C 0632 1EAB 2C3F 3837  9767 17C2 8EB1 A4B4 EE6A