I've been reading the chapter on continuations, I am starting to get them, possibly I just need to give it a couple more reads.
First thing was mitigate the terrible data structure mutation I have to do when passing source code into a mutation, I want to copy the source and pass then into the mutation. It appears to be extremely difficult to do.
I wrote a script to get repos from Github that depend on mocha, which is the only test runner I currently support.
Now for the next bit, clone some of those and then
- run Cerebro
- see if fail
- fix it
- get the next
- adapt Cerebro
- repeat
I also managed to return both the mutated and original code from the mutation, so now I can clean a bit core.cljs, which was starting to become a real mess. It seems like a small thing but took me half today to figure out how to do it, it basically involved understanding a bit more about how JavaScript interop works:
(defn loop-nodes
"applies a mutation to the ASTs,
returns the mutated, or non mutated, ASTs"
[candidate]
(let [original-code (.parse js/JSON (.stringify js/JSON (candidate :code)))]
(.visit esrecurse (candidate :code) (clj->js {:BinaryExpression mutate}))
(hash-map :original
(hash-map :path (candidate :path) :code original-code)
:mutated
(hash-map :path (candidate :path) :code (candidate :code)))))this is the code that applies a mutation, (candidate :code) (ah BTW, so nice that in ClojureScript hashmaps are functions!)
containes the AST, and it is passed through esrecurse and the mutated in mutate. What I had to do is basically
a copy of (candidate :code), via the stringify and parse horrible bit. I should update that to use a less hacky
system and put it where it could be used by other mutations.
I've tried going through chapter 6 for a couple hours. Failed. Coulnd't really focus.