assert¶

Assertions for writing candela tests.

import "std/assert";

The functions are reached through the module's name, assert::eq(a, b). A test file that calls a few of them often can name them in the import instead, import "std/assert" { eq, ne };, and call eq(a, b).

Each assertion raises an error when its check fails, so a test file runs its checks in sequence and the first failure stops the run and prints the message. A thrown message doubles as the error's code, so a harness reading the run's output can match on it. Wrapping a check in try catches the failure instead of ending the run, which is how a harness reports the failing check and carries on. See error handling.

import "std/assert";

fn main() {
    try {
        assert::eq(2, 3);
    } catch e {
        print("check failed: " + e);
    }
}

The module is pure candela, so it compiles into a .cdlb artifact and runs under candela-vm with no dynamic library.

that¶

assert::that(cond)

• cond: a bool.
• Returns: nothing.

Raises assertion failed when cond is false.

msg¶

assert::msg(cond, message)

• cond: a bool.
• message: the string to raise.
• Returns: nothing.

Raises message when cond is false. Use it when the check alone does not say what went wrong.

import "std/assert";

fn main() {
    let users = ["ada", "grace"];
    assert::msg(users.len() == 2, "expected two users");
    print("ok");
}

is_true¶

assert::is_true(cond)

• cond: a bool.
• Returns: nothing.

Raises expected true when cond is false.

is_false¶

assert::is_false(cond)

• cond: a bool.
• Returns: nothing.

Raises expected false when cond is true.

eq¶

assert::eq(a, b)

• a, b: two values of the same type.
• Returns: nothing.

Raises assert::eq failed: <a> != <b> when the two differ, rendering both sides with string. Comparison is !=, so it works on any type the operator accepts, including lists and maps.

ne¶

assert::ne(a, b)

• a, b: two values of the same type.
• Returns: nothing.

Raises assert::ne failed: both sides are <a> when the two are equal.