Wrong character encoding when using tagged templates

[HrX03]

Tagged templates seem to format special characters erroneously, for example “ gets compiled into \xe2\x80\x9c which gets parsed/printed/shown as �

Example playground file:
https://rescript-lang.org/try?version=v11.1.0&code=DYUwLgBCAeELwQBQGcwCcCWA7A5sgNBAPoCUAXBKprvAHwQDeAUBBAFLIB0AgmmgIYBPTgCsA9tgDqGMAAtEAIgWEq2PCSYBfJkwDCYrMjGhOwMTkQMYAA0A4BNc0kgA

Rescript version: v11.1.0

[cometkim]

there are three noticeable bugs in the result

1. Shouldn't the curly brace ({ ... }) be syntax error?
2. ex`...` impl shouldn't be inlined
3. and as you reported, wrong encoding

[zth]

1. Shouldn't the curly brace ({ ... }) be syntax error?

No, that's just another block that returns at the end. It's fine.

2. ex`...` impl shouldn't be inlined
3. and as you reported, wrong encoding

You can get this to work for external tagged template functions: https://rescript-lang.org/try?version=v11.1.0&code=AIFwhg5hCmAmAq0C2AHANmE0AE0AeWATgHZhq54Bc2AFGIYWAJ4A8AziIQJbEQB8AGmz1GrAPp8AlNgC8fbB269Z2AET5VAKE0BhAPbE2etNAB0aPRBoBvfAANAOAR2AvpM1A

However, right now there's a difference between external (bindings) and internal (written in ReScript) tagged template functions, where externals emit the tagged template syntax, but the internal ones are treated like regular functions. cc @tsnobip

[tsnobip]

hmm shoot, this is quite surprising given all strings in rescript should now be in unicode. Given the original PR was quite old, maybe it reused some out-of-date logic around strings.