GSoC Proposal

[Mojca Miklavec]

On Sat, 6 Apr 2019 at 13:54, KARAN SHETH wrote:
>
> I would also like to ask if anything else needs to be added or updated
> in the proposal as the deadline is nearing.

There's nothing specific that I would want to comment on.
Well, to me it's not entirely clear how the automatic updates would be
supposed to work, but ...
And the stretch goals are somewhat short :)

The only thing (mentioned yesterday) is that you either try to support
Haskell, or you don't even try. If we want to try (and spend the exact
same amount of time as you initially planned), then it makes sense to
spend a bit of time earlier than in week 11. If you only start
exploring the options in week 11, I would say that it's nearly
guaranteed that you would run out of time while potentially waiting
for someone else to do something, or waiting for some good ideas to
pop up.

If you would want to do Haskell, you would probably need to plant a
seed early (not spend more than half a day or a day in any case), let
it ripe while working on other important features, and then see if
this is feasible / worth finishing or not. I still don't know what the
major blocks are, but just because something is not written in json
should not be the reason to give up. The format still seems parsable
to me (not sure how many complications would arise).

http://hackage.haskell.org/package/cabal2arch
http://hackage.haskell.org/package/cabal2nix
http://hackage.haskell.org/package/cabal2spec
https://github.com/ddssff/cabal-debian

Please note that I'm not implying that this is of vital importance to
the project. A well written support for automatic updates of
python/ruby/perl/npm ports would be worth a lot more than a bunch of
semi-broken haskell/cabal packages.

Regarding updates: it would probably help to have two modes: one that
actually changes the Portfile, and one that just tells you what to
change (warning if there are new or dropped dependencies).

Mojca