[kde-community] [Kde-pim] A new home for Mozilla Thunderbird at KDE?

[Boudhayan Gupta]

Hi,

On 27 April 2016 at 17:05, Frederik Schwarzer <schwarzer at kde.org> wrote:
> Hi,
>
> I share Eike's concerns about the extra workload for our sysadmins and I
> think they should definitely have a word in this.
> While a welcoming atmosphere is a great goal and should be strived for in
> general, technical issues should be handled as such.
> So I guess we need to ask concrete questions like:
> - what do the Thunderbird developers expect from their future host and can
> we deliver that?
> - what are the costs in terms of manpower on our side?
> - are our sysadmins willing to let a bunch of contractors paid by Mozilla
> run through our infrastructure and tell them what to do and how?
> - do the possible changes increase daily maintenance workload after the
> contractors left?
> - how much of our infrastructure can Thunderbird actually benefit from?
>  ... or rather, are they willing to benefit from it at all? I have
> Translations in mind. Will they switch to our way of doing things or will
> they be an encapsulated project within KDE?
>
> In short: this should not solely be a community decision but a technical one
> as well.

Speaking on behalf of KDE Sysadmin, here's our deal now.

The Phabricator migration is stretching us to our limits. It turns out
the Phabricator migration isn't a simple "install this and import
data" thing, but we have to re-architect significant parts of our
server infrastructure, upgrade and/or spin up new servers/containers,
kill old services and preserve data, and wait around on Phabricator
upstream to fix things to better accommodate our use-cases.

We're understaffed as it is. Tickets and the general churn of things
take up a lot of our time. Scarlett single-handedly has to battle
against the CI demons and there are days when CI wins. Me and my GSoC
student plan to overhaul Git and get it ready for Phabricator over the
summer. I'll also have a day job in addition to SoC.

That said, we're not going to throw up our hands and say "too much
work, can't do more". Given that Thunderbird uses absolutely zero KDE
libraries and shares no development processes with KDE, I suspect
T-bird can make use of exactly nothing of our existing infra (except
project management and perhaps CI). If manpower and funds are
available, we're more than willing to co-operate with and accommodate
a dedicated team who'll work on the Thunderbird bits of our infra.

My personal opinion that T-bird will go the way of the dodo
notwithstanding, I think the board needs to take a look at whether
this is feasible in terms of funds and manpower. If the Kommunity
wants it, and the board can fund it, we'd love to accommodate T-bird,
but we'll need a lot of help.

-- Boudhayan Gupta