fixes #2172: allow escape sequences in JS FFI exported using strings

[michaelficarra]

Fixes #2172.

[coveralls]

Coverage increased (+0.009%) to 56.033% when pulling feb3e0a on michaelficarra:GH-2172 into 3944c6e on purescript:0.9.

[michaelficarra]

Some of the Travis builds passed and some failed 😕. Not sure what to do.

[hdgarrood]

Is examples/passing/2172.js being included in the source distribution? If not, the purescript.cabal file may need amending. It looks like the OSX builds may be failing because of a corrupted cache or a bug in the build script anyway though, which I can investigate this weekend if nobody else gets there first.

[michaelficarra]

Thanks, @hdgarrood. I've amended the purescript.cabal file to include examples/passing/*.js.

[coveralls]

Coverage increased (+0.009%) to 56.033% when pulling 90745e9 on michaelficarra:GH-2172 into 3944c6e on purescript:0.9.

[paf31]

So language-javascript gives the raw escaped string, straight out of the JS source?

[michaelficarra]

Yeah, unfortunately.

[paf31]

Thanks for this! I've cleared out the build cache in Travis, so we'll see what happens.

[paf31]

CI passes with the cache cleared out.

Is this ready to merge then? If so, I'll merge this and 0.9 into master, and then we can release 0.9.1 tomorrow.

[michaelficarra]

@paf31 Should be good to go once CI finishes.

[paf31]

Failing due to -Werror and

In an equation for `go': Patterns not matched: []

[michaelficarra]

@paf31 Fixed (even though it's unreachable 😓).

[coveralls]

Coverage increased (+0.008%) to 56.032% when pulling 17c486d on michaelficarra:GH-2172 into 3944c6e on purescript:0.9.

[paf31]

Thanks!