Add induceJust

[Avasil]

Fixes #202

[snowleopard]

@Avasil Looks great, many thanks! I've added a few minor comments.

[snowleopard]

By the way, we could also add induceJust1 to NonEmpty.* modules, which has a cute type signature:

induceJust1 :: Graph (Maybe a) -> Maybe (Graph a)

Feel free to ignore this if you don't have enough time.

[Avasil]

@snowleopard Thanks for the review! I implemented NonEmpty versions and other comments
It seems like the build is failing now https://api.travis-ci.org/v3/job/541821370/log.txt

Mostly in RewriteRules - is it nondeterministic test, or could it be because of my changes?

[snowleopard]

Hmm, I'm not sure what's going on with rewrite rules. @nobrakal Could you please have a look?

[nobrakal]

That is really strange, I just ran the test suite, and it compiled normally on my machine (with stack test). The failing from Travis seems pretty strange too... I will investigate further.

Speaking of rewrite rules, at first glance, it seems that induceJust could benefit from the same kind of rule than for induce (namely the "buildR/induce" rule). Even if this is a not-so-beautiful rule, it feels strange to have one for induce and not for induceJust.

Ps: I just noticed that the bench suite was broken (again...), I really need to fix this once and for all !

[snowleopard]

I'll try to fix the CI here: #210

[snowleopard]

@Avasil Would you like to be listed in https://github.com/snowleopard/alga/blob/master/AUTHORS.md? If yes, please add yourself to the list.

[Avasil]

@snowleopard Thanks, I added myself!

@nobrakal Good idea, I can look into that but I'd rather do it in separate PR

[snowleopard]

@Avasil Great! All looks good now. I'll merge once CI passes.