Xag

Try it

The editor is here once the page has loaded. It runs on a small interpreter in your browser, not on the compiler.

What this is not

This is not the compiler. It is a few hundred lines written for this website, and where it disagrees with xagc, it is wrong and xagc is right.

It exists because trying a syntax and trusting a language are different questions, and the first should not cost a checkout and a build of LLVM. It answers the first one only.

What it knows:

• var and var.mut inside START, and const above it, of type int64
• set, on a name whose chain said mut
• arithmetic: +, -, x, / and mod, with mathematics' own precedence
• print.stdout, with str: text, names, sums and \n
• loop.range.int64, over a first and last value
• one START block, and comments

What it does not:

• ownership — nothing is moved, borrowed or checked. This is the half of Xag that matters most and the playground has none of it
• every type but int64: no str variables, no bin, no deci, no bool
• functions, struct, many, if, when, give, read.stdin
• the compiler's diagnostics — errors here are this interpreter's own words, and much poorer
• anything about how fast a real Xag program runs
• being right. Where it and xagc disagree, xagc is right

Every program it ships with is run through both it and the real compiler, and they agree — that check is src/verify_play.rs and somebody has to run it. It is not protection against the language moving underneath, which Xag does, weekly. Build the compiler for the language itself.