[Hamara-devel] is there a case for multiarch in hamara ?

[Vikas Tara]

On 19/05/15 16:25, shirish wrote:
> Hi all,
> I was researching something and suddenly realized that there has never 
> been a discussion or even talk about multiarch. While we have just 3 
> targets for now (32-bit, 64-bit and ARM) don't remember whether it's 
> just armel or armhf (probably armel) but if the number of 
> architectures goes up, the need for multiarch goes up as well.
We have both armel and armhf
>
> See https://wiki.debian.org/Multiarch/TheCaseForMultiarch for 
> arguments in favour of multiarch. While they have listed only a couple 
> of examples, they have been able to convince and got multiarch support 
> added by huge number of upstreams. If we go the same way then there 
> would be that less of a diff. between us and debian.
>
> There is a cost to this route though, proper multiarch support does 
> add anywhere between 5-10% to added code as correct paths need to be 
> figured in every instance. Debian does also spend quite a bit of 
> significant time in trouble-shooting multiarch issues but then it has 
> enough people who have a buy-in and do support the idea of multiarch 
> and do try to make sure that each new package going in is multi-archable.
>
> It is possible that it may seem superfluous/overkill at this point of 
> time but needed to share that as a matter of record in case we feel 
> the need to revisit this issue sometime in future.
I would say overkill right now - it's too early. But good to have in the 
backlog?

-- 
Founder - Hamara Linux
www.hamaralinux.org
www.twitter.com/hamaralinux