| 1 | package www_test |
| 2 | |
| 3 | // The _redirects grammar and matching of CARLOS's static edge, mirrored |
| 4 | // from internal/edge/redirects.go in carlos/platform (origin/main, |
| 5 | // a8b52609) so the test server answers as the edge does and the tests |
| 6 | // can refuse a file `carlos ship` would refuse. Only the edge's own |
| 7 | // policy around the rules (its probe guard) is left out. When the edge |
| 8 | // changes its grammar, this changes with it. |
| 9 | |
| 10 | import ( |
| 11 | "errors" |
| 12 | "fmt" |
| 13 | "net" |
| 14 | "net/url" |
| 15 | "path" |
| 16 | "strconv" |
| 17 | "strings" |
| 18 | ) |
| 19 | |
| 20 | // The edge's limits on one _redirects file. |
| 21 | const ( |
| 22 | redirectsMaxBytes = 256 << 10 |
| 23 | redirectsMaxRules = 2000 |
| 24 | redirectsMaxLine = 2048 |
| 25 | ) |
| 26 | |
| 27 | // rule is one parsed line of a _redirects file. |
| 28 | type rule struct { |
| 29 | line int // 1-based line number |
| 30 | host string // lowercased; "" matches any host |
| 31 | segs []string // source path segments; ":name" captures one |
| 32 | splat bool // the source ends in "*" |
| 33 | prefix string // a URL destination's "scheme://host", never substituted |
| 34 | to string // the destination template after prefix |
| 35 | status int |
| 36 | force bool // "!": applies even when a file exists at the request path |
| 37 | } |
| 38 | |
| 39 | // isRedirect reports whether status is one a redirect rule may answer with. |
| 40 | func isRedirect(status int) bool { |
| 41 | switch status { |
| 42 | case 301, 302, 303, 307, 308: |
| 43 | return true |
| 44 | } |
| 45 | return false |
| 46 | } |
| 47 | |
| 48 | // parseRedirects parses a _redirects file as `carlos ship` does: every |
| 49 | // line it cannot use is an error, named by number. |
| 50 | func parseRedirects(data []byte) ([]rule, error) { |
| 51 | if len(data) > redirectsMaxBytes { |
| 52 | return nil, fmt.Errorf("_redirects is %d bytes; the limit is %d", len(data), redirectsMaxBytes) |
| 53 | } |
| 54 | var rules []rule |
| 55 | var errs []error |
| 56 | text := strings.TrimPrefix(string(data), "\ufeff") |
| 57 | for i, line := range strings.Split(text, "\n") { |
| 58 | n := i + 1 |
| 59 | line = strings.TrimSpace(strings.TrimSuffix(line, "\r")) |
| 60 | if line == "" || strings.HasPrefix(line, "#") { |
| 61 | continue |
| 62 | } |
| 63 | if len(line) > redirectsMaxLine { |
| 64 | errs = append(errs, fmt.Errorf("line %d: longer than %d characters", n, redirectsMaxLine)) |
| 65 | continue |
| 66 | } |
| 67 | r, err := parseRedirectLine(line) |
| 68 | if err != nil { |
| 69 | errs = append(errs, fmt.Errorf("line %d: %v", n, err)) |
| 70 | continue |
| 71 | } |
| 72 | if len(rules) == redirectsMaxRules { |
| 73 | errs = append(errs, fmt.Errorf("line %d: more than %d rules", n, redirectsMaxRules)) |
| 74 | continue |
| 75 | } |
| 76 | r.line = n |
| 77 | rules = append(rules, r) |
| 78 | } |
| 79 | return rules, errors.Join(errs...) |
| 80 | } |
| 81 | |
| 82 | func parseRedirectLine(line string) (rule, error) { |
| 83 | var fields []string |
| 84 | for _, f := range strings.Fields(line) { |
| 85 | if strings.HasPrefix(f, "#") { |
| 86 | break |
| 87 | } |
| 88 | fields = append(fields, f) |
| 89 | } |
| 90 | var r rule |
| 91 | switch { |
| 92 | case len(fields) < 2: |
| 93 | return r, errors.New("want a source and a destination") |
| 94 | case len(fields) > 3: |
| 95 | if strings.Contains(fields[3], "=") { |
| 96 | return r, fmt.Errorf("conditions such as %q are not supported", fields[3]) |
| 97 | } |
| 98 | return r, fmt.Errorf("unexpected %q after the status", fields[3]) |
| 99 | } |
| 100 | |
| 101 | r.status = 301 |
| 102 | if len(fields) == 3 { |
| 103 | st := fields[2] |
| 104 | if s, ok := strings.CutSuffix(st, "!"); ok { |
| 105 | r.force, st = true, s |
| 106 | } |
| 107 | code, err := strconv.Atoi(st) |
| 108 | if err != nil || !(isRedirect(code) || code == 200 || code == 404) { |
| 109 | return r, fmt.Errorf("status %q: want 301, 302, 303, 307, 308, 200 or 404, optionally followed by !", fields[2]) |
| 110 | } |
| 111 | r.status = code |
| 112 | } |
| 113 | |
| 114 | src := fields[0] |
| 115 | if strings.HasPrefix(src, "http://") || strings.HasPrefix(src, "https://") { |
| 116 | u, err := url.Parse(src) |
| 117 | if err != nil || u.Host == "" || u.User != nil || u.Port() != "" || u.RawQuery != "" || u.Fragment != "" { |
| 118 | return r, fmt.Errorf("source %q: want a URL with only a host and a path", src) |
| 119 | } |
| 120 | r.host = hostName(u.Hostname()) |
| 121 | src = u.Path |
| 122 | if src == "" { |
| 123 | src = "/" |
| 124 | } |
| 125 | } |
| 126 | if !strings.HasPrefix(src, "/") { |
| 127 | return r, fmt.Errorf("source %q: want a path starting with / or an http(s) URL", fields[0]) |
| 128 | } |
| 129 | if strings.ContainsAny(src, "?#") { |
| 130 | return r, fmt.Errorf("source %q: query strings and fragments are not matched", fields[0]) |
| 131 | } |
| 132 | names := map[string]bool{} |
| 133 | trimmed := strings.Trim(src, "/") |
| 134 | for _, seg := range strings.Split(trimmed, "/") { |
| 135 | switch { |
| 136 | case seg == "": |
| 137 | if trimmed != "" { |
| 138 | return r, fmt.Errorf("source %q: empty path segment", fields[0]) |
| 139 | } |
| 140 | case r.splat: |
| 141 | return r, fmt.Errorf("source %q: * must be the last segment", fields[0]) |
| 142 | case seg == "*": |
| 143 | r.splat = true |
| 144 | case strings.Contains(seg, "*"): |
| 145 | return r, fmt.Errorf("source %q: * must be a whole segment", fields[0]) |
| 146 | case strings.HasPrefix(seg, ":"): |
| 147 | name := seg[1:] |
| 148 | if !validPlaceholder(name) || name == "splat" { |
| 149 | return r, fmt.Errorf("source %q: bad placeholder %q", fields[0], seg) |
| 150 | } |
| 151 | if names[name] { |
| 152 | return r, fmt.Errorf("source %q: placeholder %q appears twice", fields[0], seg) |
| 153 | } |
| 154 | names[name] = true |
| 155 | r.segs = append(r.segs, seg) |
| 156 | default: |
| 157 | dec, err := url.PathUnescape(seg) |
| 158 | if err != nil || strings.Contains(dec, "/") { |
| 159 | return r, fmt.Errorf("source %q: bad percent-encoding in %q", fields[0], seg) |
| 160 | } |
| 161 | if strings.HasPrefix(dec, ":") { |
| 162 | return r, fmt.Errorf("source %q: %q would read as a placeholder", fields[0], seg) |
| 163 | } |
| 164 | if dec == "." || dec == ".." { |
| 165 | return r, fmt.Errorf("source %q: dot segments never match", fields[0]) |
| 166 | } |
| 167 | r.segs = append(r.segs, dec) |
| 168 | } |
| 169 | } |
| 170 | |
| 171 | dst := fields[1] |
| 172 | switch { |
| 173 | case strings.HasPrefix(dst, "http://") || strings.HasPrefix(dst, "https://"): |
| 174 | if !isRedirect(r.status) { |
| 175 | return r, fmt.Errorf("destination %q: a %d serves a file from this site, so it takes a path, not a URL", dst, r.status) |
| 176 | } |
| 177 | u, err := url.Parse(dst) |
| 178 | if err != nil || u.Host == "" || u.User != nil { |
| 179 | return r, fmt.Errorf("destination %q: not a usable URL", dst) |
| 180 | } |
| 181 | i := len(u.Scheme) + len("://") + len(u.Host) |
| 182 | if !strings.EqualFold(dst[:i], u.Scheme+"://"+u.Host) { |
| 183 | return r, fmt.Errorf("destination %q: not a usable URL", dst) |
| 184 | } |
| 185 | r.prefix, r.to = dst[:i], dst[i:] |
| 186 | if r.to != "" && !strings.ContainsAny(r.to[:1], "/?#") { |
| 187 | return r, fmt.Errorf("destination %q: not a usable URL", dst) |
| 188 | } |
| 189 | case strings.HasPrefix(dst, "//") || strings.HasPrefix(dst, "/\\"): |
| 190 | return r, fmt.Errorf("destination %q: name another host with https://", dst) |
| 191 | case strings.HasPrefix(dst, "/"): |
| 192 | r.to = dst |
| 193 | default: |
| 194 | return r, fmt.Errorf("destination %q: want a path starting with / or an http(s) URL", dst) |
| 195 | } |
| 196 | return r, nil |
| 197 | } |
| 198 | |
| 199 | // validPlaceholder: a placeholder name is letters, digits and underscores. |
| 200 | func validPlaceholder(name string) bool { |
| 201 | if name == "" { |
| 202 | return false |
| 203 | } |
| 204 | for _, c := range name { |
| 205 | if !(c == '_' || c >= '0' && c <= '9' || c >= 'a' && c <= 'z' || c >= 'A' && c <= 'Z') { |
| 206 | return false |
| 207 | } |
| 208 | } |
| 209 | return true |
| 210 | } |
| 211 | |
| 212 | // hostName is a request's host as the edge compares it: lowercased, |
| 213 | // with no port and no trailing dot. The edge also folds an |
| 214 | // internationalised name to ASCII, which no host here needs. |
| 215 | func hostName(h string) string { |
| 216 | if host, _, err := net.SplitHostPort(h); err == nil { |
| 217 | h = host |
| 218 | } |
| 219 | return strings.TrimSuffix(strings.ToLower(h), ".") |
| 220 | } |
| 221 | |
| 222 | // segments splits a cleaned, rooted path into its segments; "/" has none. |
| 223 | func segments(p string) []string { |
| 224 | if p == "/" { |
| 225 | return nil |
| 226 | } |
| 227 | return strings.Split(p[1:], "/") |
| 228 | } |
| 229 | |
| 230 | // match reports whether the rule applies to a request for host whose |
| 231 | // cleaned path splits into reqSegs, and returns what the source captured. |
| 232 | func (r *rule) match(host string, reqSegs []string) (map[string]string, bool) { |
| 233 | if r.host != "" && r.host != host { |
| 234 | return nil, false |
| 235 | } |
| 236 | if len(reqSegs) < len(r.segs) || !r.splat && len(reqSegs) != len(r.segs) { |
| 237 | return nil, false |
| 238 | } |
| 239 | caps := map[string]string{} |
| 240 | for i, seg := range r.segs { |
| 241 | if strings.HasPrefix(seg, ":") { |
| 242 | caps[seg[1:]] = reqSegs[i] |
| 243 | } else if seg != reqSegs[i] { |
| 244 | return nil, false |
| 245 | } |
| 246 | } |
| 247 | if r.splat { |
| 248 | caps["splat"] = strings.Join(reqSegs[len(r.segs):], "/") |
| 249 | } |
| 250 | return caps, true |
| 251 | } |
| 252 | |
| 253 | // expand substitutes the captures into the destination template, each |
| 254 | // percent-encoded for a Location when escape is set, raw for a |
| 255 | // rewrite's file lookup. A ":name" the source did not capture stays. |
| 256 | func expand(t string, caps map[string]string, escape bool) string { |
| 257 | var b strings.Builder |
| 258 | inQuery := false |
| 259 | for i := 0; i < len(t); { |
| 260 | c := t[i] |
| 261 | if c == '?' || c == '#' { |
| 262 | inQuery = true |
| 263 | } |
| 264 | if c != ':' { |
| 265 | b.WriteByte(c) |
| 266 | i++ |
| 267 | continue |
| 268 | } |
| 269 | j := i + 1 |
| 270 | for j < len(t) && validPlaceholder(t[j:j+1]) { |
| 271 | j++ |
| 272 | } |
| 273 | v, ok := caps[t[i+1:j]] |
| 274 | if !ok { |
| 275 | b.WriteString(t[i:j]) |
| 276 | i = j |
| 277 | continue |
| 278 | } |
| 279 | switch { |
| 280 | case !escape: |
| 281 | b.WriteString(v) |
| 282 | case inQuery: |
| 283 | b.WriteString(url.QueryEscape(v)) |
| 284 | default: |
| 285 | parts := strings.Split(v, "/") |
| 286 | for k, p := range parts { |
| 287 | parts[k] = url.PathEscape(p) |
| 288 | } |
| 289 | b.WriteString(strings.Join(parts, "/")) |
| 290 | } |
| 291 | i = j |
| 292 | } |
| 293 | return b.String() |
| 294 | } |
| 295 | |
| 296 | // action is what a matched rule asks for: a redirect's status and |
| 297 | // Location, or a rewrite's status and file. |
| 298 | type action struct { |
| 299 | status int |
| 300 | location string |
| 301 | file string |
| 302 | } |
| 303 | |
| 304 | // resolve finds the first rule that applies to host and the cleaned |
| 305 | // path p. exists is whether the tree has a file at p, which shadows |
| 306 | // every rule not forced. rawPath and rawQuery are the request's own, so |
| 307 | // a redirect to the URL asked for is skipped rather than looped. |
| 308 | func resolve(rules []rule, host, p, rawPath, rawQuery string, exists bool) (action, bool) { |
| 309 | reqSegs := segments(p) |
| 310 | for i := range rules { |
| 311 | r := &rules[i] |
| 312 | if exists && !r.force { |
| 313 | continue |
| 314 | } |
| 315 | caps, ok := r.match(host, reqSegs) |
| 316 | if !ok { |
| 317 | continue |
| 318 | } |
| 319 | if sp := caps["splat"]; sp != "" && strings.HasSuffix(rawPath, "/") { |
| 320 | caps["splat"] = sp + "/" |
| 321 | } |
| 322 | if !isRedirect(r.status) { |
| 323 | to := r.to |
| 324 | if k := strings.IndexAny(to, "?#"); k >= 0 { |
| 325 | to = to[:k] |
| 326 | } |
| 327 | return action{status: r.status, file: path.Clean("/" + expand(to, caps, false))}, true |
| 328 | } |
| 329 | loc := expand(r.to, caps, true) |
| 330 | frag := "" |
| 331 | if k := strings.IndexByte(loc, '#'); k >= 0 { |
| 332 | loc, frag = loc[:k], loc[k:] |
| 333 | } |
| 334 | if r.prefix == "" { |
| 335 | pth, q, hasQ := strings.Cut(loc, "?") |
| 336 | trailing := strings.HasSuffix(pth, "/") |
| 337 | pth = path.Clean("/" + pth) |
| 338 | if trailing && pth != "/" { |
| 339 | pth += "/" |
| 340 | } |
| 341 | loc = pth |
| 342 | if hasQ { |
| 343 | loc += "?" + q |
| 344 | } |
| 345 | } |
| 346 | if !strings.Contains(loc, "?") && rawQuery != "" { |
| 347 | loc += "?" + rawQuery |
| 348 | } |
| 349 | if r.prefix == "" || hostOf(r.prefix) == host { |
| 350 | pth, q, _ := strings.Cut(loc, "?") |
| 351 | if pth == "" { |
| 352 | pth = "/" |
| 353 | } |
| 354 | if u, err := url.Parse(pth); err == nil && u.Path == rawPath && q == rawQuery { |
| 355 | continue |
| 356 | } |
| 357 | } |
| 358 | return action{status: r.status, location: r.prefix + loc + frag}, true |
| 359 | } |
| 360 | return action{}, false |
| 361 | } |
| 362 | |
| 363 | // names reports whether some rule names the cleaned path p outright: |
| 364 | // no splat and no ":name", its segments p's. |
| 365 | func names(rules []rule, host, p string) bool { |
| 366 | reqSegs := segments(p) |
| 367 | outer: |
| 368 | for i := range rules { |
| 369 | r := &rules[i] |
| 370 | if r.splat || len(r.segs) != len(reqSegs) || r.host != "" && r.host != host { |
| 371 | continue |
| 372 | } |
| 373 | for j, seg := range r.segs { |
| 374 | if strings.HasPrefix(seg, ":") || seg != reqSegs[j] { |
| 375 | continue outer |
| 376 | } |
| 377 | } |
| 378 | return true |
| 379 | } |
| 380 | return false |
| 381 | } |
| 382 | |
| 383 | // redirectsSlashForm reports whether the rules would redirect the slash |
| 384 | // form of rawPath, given that the directory there exists. |
| 385 | func redirectsSlashForm(rules []rule, host, p, rawPath, rawQuery string) bool { |
| 386 | act, ok := resolve(rules, host, p, rawPath+"/", rawQuery, true) |
| 387 | return ok && act.location != "" |
| 388 | } |
| 389 | |
| 390 | // hostOf is the host of a "scheme://host[:port]" prefix. |
| 391 | func hostOf(prefix string) string { |
| 392 | _, h, _ := strings.Cut(prefix, "://") |
| 393 | return hostName(h) |
| 394 | } |
| 395 | |
amadan