automake
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [GSOC] Proposal for "Modularize Automake to improve the test-suite p


From: Mathieu Lirzin
Subject: Re: [GSOC] Proposal for "Modularize Automake to improve the test-suite performance"
Date: Sat, 24 Mar 2018 04:08:25 +0100
User-agent: Gnus/5.13 (Gnus v5.13) Emacs/25.3 (gnu/linux)

Hello,

Matthias Paulmier <address@hidden> writes:

> I am candidating for the project "Modularize Automake to improve the
> test-suite performance"
>
> As you may have noticed, this is my second proposal for the GSoC to the
> Automake project. You may have also noticed that 2 other persons are
> interested in this other project. This is why I have decided, after
> speaking with Mathieu Lirzin, that it would be a wise idea to candidate
> to other projects or project ideas. The goal of the summer of code is
> not to have a race with other candidates. As Mathieu told me, there may
> be multiple students working for Automake under his mentorship.

To be precise I will mentor at most 2 students. :-)

> You will find my proposal for this subject here :
> matthias.paulmier.emi.u-bordeaux.fr/proposal_am_2.pdf. Please give me
> feedbacks on this proposal. I am conscious it is bit lackey as I have
> thought more about the previous one (for example I don't know what to
> add to the deliverables).

This project is highly iterative, meaning that you will have to define
the modules you will be writing based on what you analyse while
refactoring.  So for the deliverables a "set" of modules is fine.

The idea is that the modules should be written so that they can be
easily testable (without requiring convoluted mocks for
example). Ideally you can define the qualities (like low coupling) that
the written modules should have.

The most important part in this proposal should be the description of
the various testing strategies that you intend to use and the generic
software architectures that could help with the modularity goal.  If you
are not familiar with either Testing or Architecture concepts for
example with terms like blackbox/whitebox testing, you should consider
studying those concepts in the community bounding phase.

Regarding the plan I think the beginning should contain at least the
following two things:

* A rewrite of the few existing unit tests with Test::More to get
  familiar with the framework.

* An analysis phase with the production of a small report containing
  some informal UML diagrams representing the current architecture of
  'automake'.

> I have been advised to look at other projects but as I said, I didn't
> find much interest in other projects either because of technical or
> philosophical aspects. I will however take a second look at other
> projects from GNU as there has been new idea submissions since I
> candidated here and if I find interesting projects I will likely make a
> proposal tonight (CET) or tomorrow and keep you informed.

IMHO with 2 proposals you should be fine in particular with this
"Modularize" project that doesn't get much traction.  But feel free to
look around. :-)

Thanks.

-- 
Mathieu Lirzin
GPG: F2A3 8D7E EB2B 6640 5761  070D 0ADE E100 9460 4D37



reply via email to

[Prev in Thread] Current Thread [Next in Thread]