GSoC Proposal

[KARAN SHETH]

Hey,

I have completed most of the changes that you suggested.
Haskell frontend has been now added to technical details and I guess I got
why you were saying that why it would be tricky to add it as a frontend.
Web UI has been removed seeing that it won't be a good add-on as suggested
by Renne.

Please point out anything that I may have missed or needs to added/updated.
Also sorry for late reply, just being project submission week in college, I
have been busy with it.

Thanks,
Karan Sheth
On Mon, Apr 1, 2019 at 1:16 AM Cyril Roelandt <cyril.roelandt at aquilenet.fr>
wrote:

> Hello,
>
> On 2019-04-01 00:30, KARAN SHETH wrote:
> > I have started drafting my proposal for GSoC 2019. It would be great to
> > have a review on it before I submit the same.
>
> A few comments.
>
> First of all, "upt" is written in lower case, except when found at the
> beginning of a sentence :)
>
> To someone who is new to the subject, it is unclear what cpan2port and
> pypi2port do. Maybe you could write that they automatically write a
> Portile for a piece of software available on CPAN or PyPI, and that it
> is really useful for packagers.
>
> Then it is not really clear why upt would be better suited for this
> task. You could use a few lines to explain that, thanks to its modular
> design, it is easier to maintain, and will allow MacPorts to get
> something to work with minimal effort.
>
> I would probably have a "upt update" subcommand rather than a new option
> to "upt package", but it's nitpicking, and definitely not the most
> challenging part of this feature.
>
> About the "Adding MacPorts Backend to UPT" section: calling it very
> basic may not sound really nice to Mojca :) I think you could also
> mention that the package needs unit tests, which you will happily write!
>
> About the "Adding npm frontend" section. I do not really expect you to
> add support for NPM to all the backends: we should probably stay focused
> on MacPorts. This could be a "bonus" if you find yourself a bit bored at
> the end of the summer, though. You could write "default.nix" instead of
> "nix" in the "Package definitions" column.
>
> About the "Input / Ouptut of UPT" section. The "-o/--output" option
> already exists in upt. It is not always used by the backends though. The
> "-i/--input" option does not exist, and I would rather use the following
> syntax:
>
>         $ upt package -f pypi -b macports pkg1,pkg2,pkg3
>
> What do you think?
>
> About the "Support for Package Updates in UPT" section. Some questions
> will need to be answered first. Will all the frontends be able to
> support this? How hard will it be for other backends to implement this?
>
> About the schedule. You mention the support of npm/cpan/pypi for
> Macports in both the pre-GSoC and the Week 1. Just do it in Week 1:)
>
> Week 2: you mention the colored output, but it is not part of the
> "Technical details" section.
>
> Week 3.4: why would you add two methods of updating packages? Shouldn't
> we spend more time thinking about the right way of doing things, rather
> than coding 2 different solutions?
>
> Week 10,11: if you want to add the Haskell frontend to the "Technical
> Details", I can tell you more about why it is going to be a bit more
> difficult than other package managers:)
>
> About the additional questions. Linux is written with a capital "L".
>
>
> I hope this helps!
> Cyril.
>

-- 

 <https://www.somaiya.edu>        <http://www.somaiya-ayurvihar.org>  
<http://nareshwadi.org>  <http://somaiya.com>  <http://www.helpachild.in>  
<http://nareshwadi.org>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.macports.org/pipermail/macports-dev/attachments/20190402/bc68721a/attachment.html>