Re: [gentoo-science] GSoC Proposal: Improvements to the BLAS / LAPACK and their reverse-dependencies

[Benda Xu]

Hi Michael,

Michael Everitt <m.j.everitt@×××.org> writes:

> On 08/04/19 05:17, Mo Zhou wrote [as excerpted]:
>> I'm going to participate the gsoc project "BLAS and LAPACK runtime"[1].
>> However it seems to pose a significant change on Gentoo Science Team's
>> packages, so I'd better ask for approval first before really working
>> on this, in case anybody disagree with the proposal.
>>
>> <snip>
>> The proposed solution involes no usage of USE flag, or eselect feature.
>>
>> Objective of this project:
>>
>> 1. Integrate update-alternatives mechanism into Gentoo's reference blas,
>> i.e. netlib blas' packaging.
>
> Whilst I'm, by all means, no expert on this; I think when you look into it,
> you'll find that the eselect and update-alternatives functions are highly
> equivalent, so you should find it feasible to do the port the other way
> around, and translate the 'update-alternatives' to 'eselect'. Ideally if
> you can keep within the frameworks Gentoo already provides, this should
> make integration and maintenance a little easier!
>
> Otherwise, it sounds like you've got the size of the problem in hand, and I
> wish you all the best with the project!
> Best regards,

I agree with your points.  I think porting the logic of u-a to eselect
will make the project go smoother than porting vanilla u-a to Gentoo.

The functionality of the 2 frameworks largely overlap. I could imagine
the resistance of Gentoo community to use u-a along with eselect or
adopt u-a over eselect.

Yours,
Benda