scripted-diff: stop using the gArgs wrappers

[benma]

Following #9494, where they were introduced to ease the transition.

The first commit removes the uses, the second commit removes the wrappers themselves.

[practicalswift]

utACK 39faab7

[sipa]

utACK 39faab7

[fanquake]

Concept ACK. Needs rebase.

[jnewbery]

Tested ACK 39faab7

[benma]

Rebased.

[TheBlueMatt]

Wait, is this the direction we want to go in, or the opposite? eg i figured we'd want to slowly move towards storing references to gArgs in things that need them, eg as we move towards classes which contain things (CConnman, CChainState, etc) we'd either pass in the options relevant to those objects (ala CConnman's constructor) or pass them the gArgs to avoid the global.

[benma]

@TheBlueMatt what you mention makes sense to me, but it seems unrelated to this PR, which is just about removing the wrappers (which use the same global). This PR is also good prep-work for storing references to gArgs in classes in that sense.

[TheBlueMatt]

Hmm, I figured it'd be easier to just sed/gArgs/locallyReferencedArgs/ in files, but I guess it doesnt matter much either way.

[benma]

One can easily do this after this PR, whereas before, gArgs didn't even appear in all the places you'd expect it to and you'd have to use the same kind of complex regexp to catch them all as in the scripted diff script. I agree it doesn't matter much, but it does make it easier I think.

It could be a long time too until someone starts doing this work, and in the meantime, we can avoid the confusion of how to use it, saving review cycles/time.

[practicalswift]

Needs rebase :-)

[benma]

Rebased 😆

[benma]

Rebased.

@sipa I'd be glad if this could be merged soon, as it conflicts with other changes very often.

[benma]

How can I retrigger CI?

[sipa]

@benma I've restarted Travis.