slovo/compiler/tests/standard_string_search_trim_beta20.rs

use std::{
    env,
    ffi::OsStr,
    fs,
    path::{Path, PathBuf},
    process::{Command, Output},
    sync::atomic::{AtomicUsize, Ordering},
};

static NEXT_TEMP_ID: AtomicUsize = AtomicUsize::new(0);

const EXPECTED_TEST_OUTPUT: &str = concat!(
    "test \"explicit std string contains\" ... ok\n",
    "test \"explicit std string index_of_option\" ... ok\n",
    "test \"explicit std string last_index_of_option\" ... ok\n",
    "test \"explicit std string ascii trim\" ... ok\n",
    "test \"explicit std string search trim composition\" ... ok\n",
    "test \"explicit std string search trim all\" ... ok\n",
    "6 test(s) passed\n",
);

const STANDARD_STRING_SEARCH_TRIM_BETA20: &[&str] = &[
    "contains",
    "index_of_option",
    "last_index_of_option",
    "trim_ascii_start",
    "trim_ascii_end",
    "trim_ascii",
];

const ALLOWED_STD_REFERENCES: &[&str] = &[
    "std.result",
    "std.string.parse_bool_result",
    "std.string.parse_f64_result",
    "std.string.parse_i64_result",
    "std.string.parse_u64_result",
    "std.string.parse_i32_result",
    "std.string.parse_u32_result",
    "std.string.byte_at_result",
    "std.string.slice_result",
    "std.string.starts_with",
    "std.string.ends_with",
    "std.string.concat",
    "std.string.len",
];

#[test]
fn explicit_std_string_search_and_ascii_trim_helpers_check_and_test() {
    let project = write_project(
        "std-string-search-trim-beta20",
        r#"
(module main)

(import std.string (contains index_of_option last_index_of_option trim_ascii_start trim_ascii_end trim_ascii))

(fn option_i32_eq ((maybe (option i32)) (expected i32)) -> bool
  (match maybe
    ((some value)
      (= value expected))
    ((none)
      false)))

(fn option_i32_none ((maybe (option i32))) -> bool
  (match maybe
    ((some value)
      false)
    ((none)
      true)))

(fn imported_string_contains_ok () -> bool
  (if (contains "slovo compiler" "slo")
    (if (contains "slovo compiler" "compiler")
      (= (contains "slovo compiler" "missing") false)
      false)
    false))

(fn imported_string_index_of_ok () -> bool
  (if (option_i32_eq (index_of_option "bananana" "ana") 1)
    (if (option_i32_eq (index_of_option "slovo" "s") 0)
      (if (option_i32_eq (index_of_option "slovo" "vo") 3)
        (option_i32_none (index_of_option "slovo" "compiler"))
        false)
      false)
    false))

(fn imported_string_last_index_of_ok () -> bool
  (if (option_i32_eq (last_index_of_option "bananana" "ana") 5)
    (if (option_i32_eq (last_index_of_option "slovo" "o") 4)
      (if (option_i32_eq (last_index_of_option "slovo" "s") 0)
        (option_i32_none (last_index_of_option "slovo" "compiler"))
        false)
      false)
    false))

(fn imported_string_ascii_trim_ok () -> bool
  (if (= (trim_ascii_start "\n\t  slovo  \t") "slovo  \t")
    (if (= (trim_ascii_end "\n\t  slovo  \t") "\n\t  slovo")
      (if (= (trim_ascii "\n\t  slovo  \t") "slovo")
        (if (= (trim_ascii "slovo") "slovo")
          (= (trim_ascii "   ") "")
          false)
        false)
      false)
    false))

(fn imported_string_search_trim_composes_ok () -> bool
  (if (= (trim_ascii "  slovo compiler  ") "slovo compiler")
    (if (contains (trim_ascii "  slovo compiler  ") "compiler")
      (if (option_i32_eq (index_of_option (trim_ascii_start "\t\tprefix-core") "core") 7)
        (option_i32_eq (last_index_of_option (trim_ascii_end "core-core\n") "core") 5)
        false)
      false)
    false))

(fn imported_string_search_trim_all_ok () -> bool
  (if (imported_string_contains_ok)
    (if (imported_string_index_of_ok)
      (if (imported_string_last_index_of_ok)
        (if (imported_string_ascii_trim_ok)
          (imported_string_search_trim_composes_ok)
          false)
        false)
      false)
    false))

(fn main () -> i32
  (if (imported_string_search_trim_all_ok)
    42
    1))

(test "explicit std string contains"
  (imported_string_contains_ok))

(test "explicit std string index_of_option"
  (imported_string_index_of_ok))

(test "explicit std string last_index_of_option"
  (imported_string_last_index_of_ok))

(test "explicit std string ascii trim"
  (imported_string_ascii_trim_ok))

(test "explicit std string search trim composition"
  (imported_string_search_trim_composes_ok))

(test "explicit std string search trim all"
  (= (main) 42))
"#,
    );

    let source = read(&project.join("src/main.slo"));
    let std_string = read(&std_string_path());

    assert!(
        !project.join("src/string.slo").exists(),
        "beta20 fixture must exercise repo-root std.string, not a local module copy"
    );
    assert!(
        source.starts_with("(module main)\n\n(import std.string ("),
        "beta20 fixture must use an explicit std.string import"
    );
    assert_std_string_search_trim_facades(&std_string);

    let fmt = run_glagol([
        OsStr::new("fmt"),
        OsStr::new("--check"),
        project.as_os_str(),
    ]);
    assert_success("std string search trim fmt --check", &fmt);

    let check = run_glagol([OsStr::new("check"), project.as_os_str()]);
    assert_success_stdout(check, "", "std string search trim check");

    let test = run_glagol([OsStr::new("test"), project.as_os_str()]);
    assert_success_stdout(test, EXPECTED_TEST_OUTPUT, "std string search trim test");
}

#[test]
fn string_search_and_ascii_trim_helpers_are_not_compiler_known_runtime_calls() {
    let std_string = read(&std_string_path());
    assert_std_string_search_trim_facades(&std_string);

    for helper in STANDARD_STRING_SEARCH_TRIM_BETA20 {
        assert!(
            !std_string.contains(&format!("std.string.{}", helper)),
            "std.string.{} must remain source-authored, not a compiler-known runtime call",
            helper
        );
        assert!(
            !std_string.contains(&format!("__glagol_string_{}", helper)),
            "std.string.{} must not introduce a private runtime symbol",
            helper
        );
    }

    let cases = [
        UnsupportedRuntimeCase {
            name: "contains",
            symbol: "std.string.contains",
            source: r#"
(module main)

(fn main () -> i32
  (if (std.string.contains "slovo" "ovo")
    0
    1))
"#,
        },
        UnsupportedRuntimeCase {
            name: "index-of-option",
            symbol: "std.string.index_of_option",
            source: r#"
(module main)

(fn main () -> i32
  (match (std.string.index_of_option "slovo" "o")
    ((some value)
      value)
    ((none)
      0)))
"#,
        },
        UnsupportedRuntimeCase {
            name: "last-index-of-option",
            symbol: "std.string.last_index_of_option",
            source: r#"
(module main)

(fn main () -> i32
  (match (std.string.last_index_of_option "slovo" "o")
    ((some value)
      value)
    ((none)
      0)))
"#,
        },
        UnsupportedRuntimeCase {
            name: "trim-ascii-start",
            symbol: "std.string.trim_ascii_start",
            source: r#"
(module main)

(fn main () -> i32
  (std.string.len (std.string.trim_ascii_start " slovo")))
"#,
        },
        UnsupportedRuntimeCase {
            name: "trim-ascii-end",
            symbol: "std.string.trim_ascii_end",
            source: r#"
(module main)

(fn main () -> i32
  (std.string.len (std.string.trim_ascii_end "slovo ")))
"#,
        },
        UnsupportedRuntimeCase {
            name: "trim-ascii",
            symbol: "std.string.trim_ascii",
            source: r#"
(module main)

(fn main () -> i32
  (std.string.len (std.string.trim_ascii " slovo ")))
"#,
        },
    ];

    for case in cases {
        let fixture = write_fixture(case.name, case.source);
        let output = run_glagol([fixture.as_os_str()]);
        assert_failure_stderr_contains(
            &format!("direct {} runtime call", case.symbol),
            &output,
            &format!("standard library call `{}` is not supported", case.symbol),
        );
    }
}

fn assert_std_string_search_trim_facades(std_string: &str) {
    assert!(
        std_string.starts_with("(module string (export "),
        "lib/std/string.slo must stay a source-authored module export"
    );

    let mut non_allowed_std = std_string.to_owned();
    for allowed in ALLOWED_STD_REFERENCES {
        non_allowed_std = non_allowed_std.replace(allowed, "");
    }
    assert!(
        !non_allowed_std.contains("std."),
        "std.string beta20 helpers must use only existing std.result bridges and promoted beta16-or-earlier std.string primitives"
    );

    for helper in STANDARD_STRING_SEARCH_TRIM_BETA20 {
        assert!(
            std_string.contains(&format!("(fn {} ", helper)),
            "lib/std/string.slo is missing source facade `{}`",
            helper
        );
    }

    let search_trim_source = search_trim_source_region(std_string);
    for primitive in [
        ("len", ["std.string.len", "(len "]),
        (
            "byte_at_result",
            ["std.string.byte_at_result", "(byte_at_result "],
        ),
        (
            "slice_result",
            ["std.string.slice_result", "(slice_result "],
        ),
        ("starts_with", ["std.string.starts_with", "(starts_with "]),
    ] {
        assert!(
            primitive
                .1
                .iter()
                .any(|needle| search_trim_source.contains(needle)),
            "beta20 search/trim facades must compose over existing beta16 string primitive `{}`",
            primitive.0
        );
    }

    assert!(
        !std_string.contains("unicode")
            && !std_string.contains("grapheme")
            && !std_string.contains("locale")
            && !std_string.contains("case_insensitive")
            && !std_string.contains("regex"),
        "beta20 string helpers must not claim deferred Unicode, locale, case-folding, or regex APIs"
    );
}

fn search_trim_source_region(source: &str) -> &str {
    let ends_with_end = function_range(source, "ends_with").1;
    let parse_start = function_range(source, "parse_i32_result").0;
    &source[ends_with_end..parse_start]
}

fn function_range(source: &str, name: &str) -> (usize, usize) {
    let needle = format!("(fn {} ", name);
    let start = source
        .find(&needle)
        .unwrap_or_else(|| panic!("missing function `{}`", name));
    let mut depth = 0usize;

    for (offset, byte) in source.as_bytes()[start..].iter().enumerate() {
        match byte {
            b'(' => depth += 1,
            b')' => {
                depth = depth
                    .checked_sub(1)
                    .unwrap_or_else(|| panic!("unbalanced function `{}`", name));
                if depth == 0 {
                    return (start, start + offset + 1);
                }
            }
            _ => {}
        }
    }

    panic!("unterminated function `{}`", name);
}

fn run_glagol<I, S>(args: I) -> Output
where
    I: IntoIterator<Item = S>,
    S: AsRef<OsStr>,
{
    Command::new(env!("CARGO_BIN_EXE_glagol"))
        .args(args)
        .current_dir(Path::new(env!("CARGO_MANIFEST_DIR")))
        .output()
        .expect("run glagol")
}

fn write_project(name: &str, source: &str) -> PathBuf {
    let root = temp_root(name);
    let src = root.join("src");
    fs::create_dir_all(&src).unwrap_or_else(|err| panic!("create `{}`: {}", src.display(), err));
    fs::write(
        root.join("slovo.toml"),
        format!(
            "[project]\nname = \"{}\"\nsource_root = \"src\"\nentry = \"main\"\n",
            name
        ),
    )
    .unwrap_or_else(|err| panic!("write project manifest: {}", err));
    fs::write(src.join("main.slo"), source.trim_start())
        .unwrap_or_else(|err| panic!("write project main.slo: {}", err));
    root
}

fn write_fixture(name: &str, source: &str) -> PathBuf {
    let mut path = env::temp_dir();
    path.push(format!(
        "glagol-standard-string-search-trim-beta20-{}-{}-{}.slo",
        name,
        std::process::id(),
        NEXT_TEMP_ID.fetch_add(1, Ordering::Relaxed)
    ));
    fs::write(&path, source.trim_start())
        .unwrap_or_else(|err| panic!("write `{}`: {}", path.display(), err));
    path
}

fn temp_root(name: &str) -> PathBuf {
    let root = env::temp_dir().join(format!(
        "glagol-standard-string-search-trim-beta20-{}-{}-{}",
        name,
        std::process::id(),
        NEXT_TEMP_ID.fetch_add(1, Ordering::Relaxed)
    ));
    let _ = fs::remove_dir_all(&root);
    fs::create_dir_all(&root).unwrap_or_else(|err| panic!("create `{}`: {}", root.display(), err));
    root
}

fn std_string_path() -> PathBuf {
    Path::new(env!("CARGO_MANIFEST_DIR")).join("../lib/std/string.slo")
}

fn read(path: &Path) -> String {
    fs::read_to_string(path).unwrap_or_else(|err| panic!("read `{}`: {}", path.display(), err))
}

fn assert_success(context: &str, output: &Output) {
    let stdout = String::from_utf8_lossy(&output.stdout);
    let stderr = String::from_utf8_lossy(&output.stderr);
    assert!(
        output.status.success(),
        "{} failed\nstatus: {:?}\nstdout:\n{}\nstderr:\n{}",
        context,
        output.status.code(),
        stdout,
        stderr
    );
    assert!(stderr.is_empty(), "{} wrote stderr:\n{}", context, stderr);
}

fn assert_success_stdout(output: Output, expected: &str, context: &str) {
    assert_success(context, &output);
    let stdout = String::from_utf8_lossy(&output.stdout);
    assert_eq!(stdout, expected, "{}", context);
}

fn assert_failure_stderr_contains(context: &str, output: &Output, needle: &str) {
    let stdout = String::from_utf8_lossy(&output.stdout);
    let stderr = String::from_utf8_lossy(&output.stderr);
    assert!(
        !output.status.success(),
        "{} unexpectedly passed\nstdout:\n{}\nstderr:\n{}",
        context,
        stdout,
        stderr
    );
    assert!(
        stdout.is_empty(),
        "{} rejected compile wrote stdout:\n{}",
        context,
        stdout
    );
    assert!(
        stderr.contains(needle),
        "{} stderr did not contain `{}`:\n{}",
        context,
        needle,
        stderr
    );
}

struct UnsupportedRuntimeCase {
    name: &'static str,
    symbol: &'static str,
    source: &'static str,
}