Summary

EZ doesn't have a way to convert strings to primitive types with error handling. The only option is cast(s, int) which panics on bad input. There's no fallible conversion, no string -> uint, and no string -> bool.

A strconv module fixes this with proper fallible returns and fills the gaps.

Proposed API

strconv.to_int(s string) -> (int, Error)
strconv.to_uint(s string) -> (uint, Error)
strconv.to_float(s string) -> (float, Error)
strconv.to_bool(s string) -> (bool, Error)

strconv.from_int(n int) -> string
strconv.from_uint(n uint) -> string
strconv.from_float(f float) -> string
strconv.from_bool(b bool) -> string

strconv.is_numeric(s string) -> bool
strconv.is_integer(s string) -> bool

Example

import @strconv

do main() {
    mut val, err = strconv.to_int("42")
    if err != nil {
        println("bad input")
    }

    mut s = strconv.from_int(100)

    if strconv.is_integer(userInput) {
        mut n, _ = strconv.to_int(userInput)
    }
}

Impact on cast()

cast(s, int) currently uses internal C helpers (ez_builtin_string_to_int, ez_builtin_string_to_float) that panic on bad input. These are not user-facing builtins — they're codegen internals for cast(). Once strconv lands, cast() string-to-numeric behavior can optionally be rewired to use the strconv C implementations, but cast() itself stays as-is since its semantics are "convert or panic".

Notes

• to_int / to_uint are base-10 for now. Optional base parameter can come later
• Follows the to_X / from_X convention used by bytes, encoding, and other stdlib modules
• All to_X functions return (T, Error) consistent with EZ's fallible pattern