proposal: desugar all ops (not just %)

[Buck Evan]

This would allow built-in operations to become valid arguments to the various higher-order functions in `std`. Since there are so many (23), I'd put them in an `std.op` subsection.

* `std.op.add(x, y)` implements `+`

* `std.op.neg(x)` implements unary `-`

This has a side-benefit of making the core language smaller and simpler, with no need for operator precedence. (BTW I don't see how precedence enters into the formal semantics, but I've no practice reading that kind of thing. :)

Prior art in https://docs.python.org/3/library/operator.html

[Sebastien Diot]

I was planing to create a small library for myself, which does that by simply having functions use the operators, but having it built-in would be even better.

[Dave Cunningham]

It's not a bad idea to have named versions of the operators.  Essentially it means you can to `std.add` instead of `function(x,y) x + y`, which can be more readable.

You don't need to desugar them to get that, though.  Desugaring means you have both the operator and the standard function, but the operator is implemented in terms of the standard function.  If all you want is to have both the operator and the standard function, it is also possible to implement the standard function in terms of the operator.

--
You received this message because you are subscribed to the Google Groups "Jsonnet" group.
To unsubscribe from this group and stop receiving emails from it, send an email to jsonnet+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/jsonnet/3b43ab69-2764-437b-abe5-36a4c8a90f1en%40googlegroups.com.

[Buck Evan]

Agreed. I just thought it a nice side benefit that the order of operations can be handled in the desugaring, and the core language no longer needs any notion of operator precedence nor "left-binding", just function calls. I don't think those are formally specified at the moment, so it also makes the formal definition more comprehensive.

I completely agree it's not essential though.

You received this message because you are subscribed to a topic in the Google Groups "Jsonnet" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/jsonnet/Qk6ILtk-kxg/unsubscribe.
To unsubscribe from this group and all its topics, send an email to jsonnet+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/jsonnet/CANarjTiBfHouCeGRgvKx%3DrAMR6Se5w1-W2rhb5Lo6QKHwkDCuQ%40mail.gmail.com.

[Buck Evan]

What was the cause for desugaring modulus but not division? In my mind they're nearly identical.

[Dave Cunningham]

Modulus is also used for string formatting (intended to be Python compatible)