Hojun-Cho
d1310a03ad
Replace magic ASCII decimals with char literals in ascii/fnmatch/shlex predicates (e.g. `c < 48` → `c < '0'`). Byte-id-neutral: ascii params are rune, so rune<rune emission is unchanged; fnmatch/shlex compare u8 against value-preserving (<=126) rune constants. Range bounds (0/31/127), the ±32 case offset, the 128 high-bit mask, and fnmatch 0u8 sentinels stay decimal. Regenerate the three combined.ww that embed ascii (w6c, wwdump, smoke). Add functional rows pinning predicates reachable only via fnmatch ctype classes / shlex split: [[:space:]]/[[:print:]]/[[:graph:]] + the '\t' arm of [[:blank:]] (fnmatchtest), '\t'/'\n' split separators + issafe's special-char set (shlextest) — so a wrong substitution would be caught.
276 lines
9.2 KiB
Plaintext
276 lines
9.2 KiB
Plaintext
// fnmatchtest — exercises lib/fnmatch. Run with
|
|
// `out/bin/ww run lib/fnmatch/fnmatchtest.ww`.
|
|
//
|
|
// Table-driven via the [[check]] helper: every row is the uniform
|
|
// 4-tuple `(pattern, string, expected, flags)`. Cases follow Hare's
|
|
// ref/hare/fnmatch/+test.ha (the de-facto spec for this port);
|
|
// multibyte UTF-8 rows (わたし, ε) are skipped — our matcher is
|
|
// byte-wise and those would match byte-identically but obscure the
|
|
// ASCII semantics the test is asserting.
|
|
//
|
|
// Failure path: each @test fn bumps `signalled` to its slot index,
|
|
// `check` does `exit(signalled + 10)` on miscompare so the harness
|
|
// reports `WEXITSTATUS = 11..N` pointing at the failing scenario.
|
|
// Same convention as lib/log/logtest.
|
|
|
|
package fnmatch;
|
|
|
|
import fnmatch;
|
|
|
|
// Direct rt_syscall binding rather than `use os;` — os exports
|
|
// read/write/close, which collide with io.read/write/close under the
|
|
// driver's flat-scope concat. Mirrors logtest / fmttest / bufiotest.
|
|
@symbol("rt_syscall") fn syscall1ww(num: i64, a: i64) i64;
|
|
fn doexit(code: i32) void = {
|
|
syscall1ww(60i64, code: i64);
|
|
};
|
|
|
|
let signalled: i32 = 0;
|
|
|
|
fn fail() void = { doexit(signalled + 10); };
|
|
|
|
// check — pin one row. `flags` is taken as i32 and cast to
|
|
// fnmatch.flag so the test rows can read as integer bitmasks
|
|
// without dragging enum literals into every line.
|
|
fn check(pat: str, s: str, expected: bool, flags: i32) void = {
|
|
let f: fnmatch.flag = flags as fnmatch.flag;
|
|
if (fnmatch.fnmatch(pat, s, f) != expected) { fail(); };
|
|
};
|
|
|
|
// ---- basic literal / wildcard cases ---------------------------------
|
|
|
|
@test fn basic() void = {
|
|
check("a", "a", true, 0);
|
|
check("b", "b", true, 0);
|
|
check("\0", "\0", true, 0);
|
|
check("abcde", "abcde", true, 0);
|
|
check("aaa", "bbb", false, 0);
|
|
|
|
check("ab*cde", "abcde", true, 0);
|
|
check("g*a", "gordana", true, 0);
|
|
check("ab*cde*foo*bar", "abcdefooba", false, 0);
|
|
check("*", "foo", true, 0);
|
|
check("aa*", "aafoo", true, 0);
|
|
check("bb*", "foo", false, 0);
|
|
check("*cc", "foocc", true, 0);
|
|
check("*dd", "foo", false, 0);
|
|
check("ra**ra", "rarara", true, 0);
|
|
check("x*yy*x", "xxyyyyyxxx", true, 0);
|
|
check("*", "*", true, 0);
|
|
check("*", "", true, 0);
|
|
check("****", "a", true, 0);
|
|
check("**a**", "a", true, 0);
|
|
check("****", "", true, 0);
|
|
|
|
check("?", "*", true, 0);
|
|
check("?", "", false, 0);
|
|
check("??", "a", false, 0);
|
|
check("??", "abc", false, 0);
|
|
check("?aa", "bbb", false, 0);
|
|
|
|
check("**?**", "", false, 0);
|
|
check("*?*?*?*?", "abcd", true, 0);
|
|
check("*?*?*?*?", "abc", false, 0);
|
|
};
|
|
|
|
// ---- bracket expressions: literal / range / negation ----------------
|
|
|
|
@test fn brackets() void = {
|
|
check("[b]", "b", true, 0);
|
|
check("a[b]c", "abc", true, 0);
|
|
check("a[b]c", "axc", false, 0);
|
|
check("[a-c]", "b", true, 0);
|
|
check("[a-z]", "a", true, 0);
|
|
check("[c-a]", "b", false, 0);
|
|
check("x[a-c]y", "xay", true, 0);
|
|
check("x[a-c]y", "xby", true, 0);
|
|
check("x[a-c]y", "xcy", true, 0);
|
|
check("x[a-c]y", "xzy", false, 0);
|
|
check("x[a-c]y", "xy", false, 0);
|
|
check("[a-c]*[a-c]", "axxxb", true, 0);
|
|
|
|
// Special members: '-' at start/end, ']' as first member.
|
|
check("[-]", "-", true, 0);
|
|
check("[.]", ".", true, 0);
|
|
check("[:ias]", ":", true, 0);
|
|
check("[-]", "a", false, 0);
|
|
check("[-ac]", "a", true, 0);
|
|
check("[-ac]", "-", true, 0);
|
|
check("[-ac]", "b", false, 0);
|
|
check("[ac-]", "a", true, 0);
|
|
check("[ac-]", "-", true, 0);
|
|
check("[ac-]", "b", false, 0);
|
|
|
|
// Equivalence-class-like syntax: invalid in Hare/POSIX.
|
|
check("[.a.]", "a", false, 0);
|
|
|
|
// Negation with '!'.
|
|
check("[!b]", "b", false, 0);
|
|
check("a[!b]c", "abc", false, 0);
|
|
check("a[!b]c", "axc", true, 0);
|
|
check("[!a-c]", "b", false, 0);
|
|
check("[!c-a]", "b", true, 0);
|
|
check("x[!a-c]y", "xay", false, 0);
|
|
check("x[!a-c]y", "xby", false, 0);
|
|
check("x[!a-c]y", "xcy", false, 0);
|
|
check("x[!a-c]y", "xzy", true, 0);
|
|
check("x[!a-c]y", "xy", false, 0);
|
|
check("[!a-c]*[!a-c]", "axxxb", false, 0);
|
|
|
|
check("[!-]", "-", false, 0);
|
|
check("[!-]", "a", true, 0);
|
|
check("[!-ac]", "a", false, 0);
|
|
check("[!-ac]", "-", false, 0);
|
|
check("[!-ac]", "b", true, 0);
|
|
};
|
|
|
|
// ---- POSIX character classes: [[:alnum:]] / [[:alpha:]] / ... -------
|
|
|
|
@test fn ctype() void = {
|
|
check("[[:alnum:]]", "7", true, 0);
|
|
check("[[:alpha:]]", "[", false, 0);
|
|
check("[[:alpha:]]", "[[", false, 0);
|
|
check("[[alpha:]]", "a]", true, 0);
|
|
check("[[alpha:]]", ":]", true, 0);
|
|
check("[[:alpha:]]", "a", true, 0);
|
|
check("[[:blank:]]", " ", true, 0);
|
|
// pin the ascii.is{blank,space,print,graph} predicates: their
|
|
// char-literal bounds are reachable only through these classes.
|
|
check("[[:blank:]]", "\t", true, 0);
|
|
check("[[:space:]]", " ", true, 0);
|
|
check("[[:space:]]", "\t", true, 0);
|
|
check("[[:space:]]", "x", false, 0);
|
|
check("[[:print:]]", "~", true, 0);
|
|
check("[[:print:]]", "\t", false, 0);
|
|
check("[[:graph:]]", "a", true, 0);
|
|
check("[[:graph:]]", " ", false, 0);
|
|
check("[[:alnum:]]a", "a]a", false, 0);
|
|
check("[[:alnum:][[:digit:]]", "a", true, 0);
|
|
|
|
check("[![:alnum:]]", "a", false, 0);
|
|
check("[![:alpha:]]", "[", true, 0);
|
|
check("[![:alnum:]]a", "a]a", false, 0);
|
|
check("[![:alpha:]]a", "[a", true, 0);
|
|
check("[![:alnum:][:digit:]]", "a", false, 0);
|
|
};
|
|
|
|
// ---- flag.PERIOD: leading '.' must be literal in the pattern --------
|
|
|
|
@test fn period() void = {
|
|
let fp: i32 = 4; // flag.PERIOD
|
|
check(".", ".", true, fp);
|
|
check("*", ".", false, fp);
|
|
check("?", ".", false, fp);
|
|
check("[.]", ".", false, fp);
|
|
check(".*", ".asdf", true, fp);
|
|
check(".*", "asdf", false, fp);
|
|
};
|
|
|
|
// ---- flag.NOESCAPE: '\\' loses its escape meaning -------------------
|
|
|
|
@test fn noescape() void = {
|
|
let nesc: i32 = 2; // flag.NOESCAPE
|
|
check("\\", "\\", true, nesc);
|
|
check("\\*", "\\asdf", true, nesc);
|
|
};
|
|
|
|
// ---- musl-adapted cases (no flags) ---------------------------------
|
|
|
|
@test fn musl_basic() void = {
|
|
check("*.c", "foo.c", true, 0);
|
|
check("*.c", ".c", true, 0);
|
|
check("*.a", "foo.c", false, 0);
|
|
check("*.c", ".foo.c", true, 0);
|
|
check("a\\*.c", "ax.c", false, 0);
|
|
check("a[xy].c", "ax.c", true, 0);
|
|
check("a[!y].c", "ax.c", true, 0);
|
|
check("-O[01]", "-O1", true, 0);
|
|
check("[[?*\\]", "\\", true, 0);
|
|
check("[]?*\\]", "]", true, 0);
|
|
check("[!]a-]", "b", true, 0);
|
|
check("[]-_]", "^", true, 0);
|
|
check("[!]-_]", "X", true, 0);
|
|
check("??", "-", false, 0);
|
|
|
|
// Without PATHNAME, `[...]` and `?` happily eat '/'.
|
|
check("[*]/b", "a/b", false, 0);
|
|
check("[*]/b", "*/b", true, 0);
|
|
check("[?]/b", "a/b", false, 0);
|
|
check("[?]/b", "?/b", true, 0);
|
|
check("[[a]/b", "a/b", true, 0);
|
|
check("[[a]/b", "[/b", true, 0);
|
|
check("\\*/b", "a/b", false, 0);
|
|
check("\\*/b", "*/b", true, 0);
|
|
check("\\?/b", "a/b", false, 0);
|
|
check("\\?/b", "?/b", true, 0);
|
|
check("[/b", "[/b", false, 0);
|
|
check("\\[/b", "[/b", true, 0);
|
|
check("a[/]b", "a/b", true, 0);
|
|
|
|
// "[![:d-d]" — bracket carrying a class-name start without
|
|
// proper close-bracket is treated as a non-class literal set.
|
|
// Hare's table asserts false in all three rows.
|
|
check("[![:d-d]", "b", false, 0);
|
|
check("[[:d-d]", "[", false, 0);
|
|
check("[![:d-d]", "[", false, 0);
|
|
};
|
|
|
|
// ---- flag.PATHNAME: '/' must be matched by literal '/' --------------
|
|
|
|
@test fn pathname() void = {
|
|
let fp: i32 = 1; // flag.PATHNAME
|
|
check("[a-z]/[a-z]", "a/b", true, fp);
|
|
check("a[/]b", "a/b", false, fp);
|
|
check("*", "a/b", false, fp);
|
|
check("*[/]b", "a/b", false, fp);
|
|
check("*[b]", "a/b", false, fp);
|
|
check("a[a/z]*.c", "a/x.c", false, fp);
|
|
check("a/*.c", "a/x.c", true, fp);
|
|
check("a*.c", "a/x.c", false, fp);
|
|
check("*/foo", "/foo", true, fp);
|
|
check("???b", "aa/b", false, fp);
|
|
};
|
|
|
|
// ---- flag.PERIOD + flag.PATHNAME combined ---------------------------
|
|
|
|
@test fn combined() void = {
|
|
let pp: i32 = 1; // PATHNAME
|
|
let fp: i32 = 4; // PERIOD
|
|
let pf: i32 = pp | fp;
|
|
check("?a/b", ".a/b", false, pf);
|
|
check("a/?b", "a/.b", false, pf);
|
|
check("*a/b", ".a/b", false, pf);
|
|
check("a/*b", "a/.b", false, pf);
|
|
check("[.]a/b", ".a/b", false, pf);
|
|
check("a/[.]b", "a/.b", false, pf);
|
|
check("*/?", "a/b", true, pf);
|
|
check("?/*", "a/b", true, pf);
|
|
check(".*/?", ".a/b", true, pf);
|
|
check("*/.?", "a/.b", true, pf);
|
|
check("*/*", "a/.b", false, pf);
|
|
check("*?*/*", "a/.b", true, fp);
|
|
check("*[.]/b", "a./b", true, pf);
|
|
check("*[[:alpha:]]/*[[:alnum:]]", "a/b", true, pp);
|
|
check("a?b", "a.b", true, pf);
|
|
check("a*b", "a.b", true, pf);
|
|
check("a[.]b", "a.b", true, pf);
|
|
|
|
check("*.c", ".foo.c", false, fp);
|
|
check("*.c", "foo.c", true, fp);
|
|
|
|
let nesc: i32 = 2; // NOESCAPE
|
|
check("a\\*.c", "a*.c", false, nesc);
|
|
};
|
|
|
|
export fn main() i32 = {
|
|
signalled = 1; basic();
|
|
signalled = 2; brackets();
|
|
signalled = 3; ctype();
|
|
signalled = 4; period();
|
|
signalled = 5; noescape();
|
|
signalled = 6; musl_basic();
|
|
signalled = 7; pathname();
|
|
signalled = 8; combined();
|
|
return 0;
|
|
};
|