Ideas for Summer of Code 2018

[Mathieu Lirzin]

Hello,

Google is accepting organization applications for the next Summer of
Code [1] and as usual GNU is going to apply for it.  Let's start thinking
about a list of ideas for the next Summer of Code and potential mentors.

I plan to propose the following project:

    Modularize Automake to improve the test-suite performance

   Currently automake is implemented as a monolithic Perl script with some
   modules for code shared with aclocal which is another script
   distributed with Automake. To ensure its practical correctness Automake
   provides a huge test-suite and requires every non-trivial bug-fix to be
   covered by an additional non-regression test. Unfortunately this whole
   test-suite is quite slow to run (~20-30 min) which makes it hard for
   maintainers to effectively use it, and is long enough to refrain users
   to run it. The main reason for this slowness is that almost all tests
   are integration/validation tests that touch the file system. A more
   effective approach would be to replace those with unit tests when
   possible. The objective of this project is to incrementally refactor
   the current implementation by decomposing it into modules and unit test
   those with [1]Test::More.

   Skills: Good understanding of the different testing strategies, Basic
   knowledge of Perl
   Mentor: Mathieu Lirzin

References

   1. https://perldoc.perl.org/Test/More.html

Comments and suggestions welcome.

If you want to propose another project that you are willing to mentor,
feel free to share your ideas on this list too.  The deadline for
submitting a project is the beginning of March.

[1] https://summerofcode.withgoogle.com/

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