pr_01m47d15m3e54sn21z27rpy5n9/services/repos/src/git_http.rs
| 1 | //! Git over HTTPS: the smart HTTP remote at `/<namespace>/<repo>.git`, |
| 2 | //! proxied to the git store with a short-lived token. |
| 3 | |
| 4 | use g1t_contracts::identity::GitCredentialsArgs; |
| 5 | use g1t_contracts::repos::{GitAccess, GitService, RepoPath}; |
| 6 | use g1t_contracts::{FailureCode, Outcome, Viewer}; |
| 7 | use worker::js_sys::Uint8Array; |
| 8 | use worker::{Fetch, Fetcher, Headers, Method, Request, RequestInit, Response, Result, Url}; |
| 9 | |
| 10 | const ENDPOINTS: [&str; 3] = ["info/refs", "git-upload-pack", "git-receive-pack"]; |
| 11 | const FORWARDED_HEADERS: [&str; 5] = [ |
| 12 | "accept", |
| 13 | "content-encoding", |
| 14 | "content-type", |
| 15 | "git-protocol", |
| 16 | "user-agent", |
| 17 | ]; |
| 18 | |
| 19 | /// A git request, parsed from its URL. |
| 20 | pub struct GitRequest { |
| 21 | pub path: RepoPath, |
| 22 | pub endpoint: &'static str, |
| 23 | pub service: GitService, |
| 24 | } |
| 25 | |
| 26 | /// Parses `/<namespace>/<name>[.git]/<endpoint>`, or returns `None` if the |
| 27 | /// request is not git's. |
| 28 | pub fn parse(url: &Url) -> Option<GitRequest> { |
| 29 | let path = url.path().strip_prefix('/')?; |
| 30 | let endpoint = ENDPOINTS |
| 31 | .into_iter() |
| 32 | .find(|endpoint| path.ends_with(&format!("/{endpoint}")))?; |
| 33 | let repo = &path[..path.len() - endpoint.len() - 1]; |
| 34 | let (namespace, name) = repo.split_once('/')?; |
| 35 | let name = name.strip_suffix(".git").unwrap_or(name); |
| 36 | if namespace.is_empty() || name.is_empty() || name.contains('/') { |
| 37 | return None; |
| 38 | } |
| 39 | let service = if endpoint == "info/refs" { |
| 40 | url.query_pairs() |
| 41 | .find(|(key, _)| key == "service") |
| 42 | .map(|(_, value)| value.into_owned())? |
| 43 | } else { |
| 44 | endpoint.to_owned() |
| 45 | }; |
| 46 | let service = match service.as_str() { |
| 47 | "git-upload-pack" => GitService::UploadPack, |
| 48 | "git-receive-pack" => GitService::ReceivePack, |
| 49 | _ => return None, |
| 50 | }; |
| 51 | Some(GitRequest { |
| 52 | path: RepoPath { |
| 53 | namespace: namespace.to_owned(), |
| 54 | name: name.to_owned(), |
| 55 | }, |
| 56 | endpoint, |
| 57 | service, |
| 58 | }) |
| 59 | } |
| 60 | |
| 61 | /// The user named by an HTTP Basic `Authorization` header, as git sends it. |
| 62 | pub async fn viewer(request: &Request, identity: &Fetcher) -> Result<Viewer> { |
| 63 | let Some(header) = request.headers().get("authorization")? else { |
| 64 | return Ok(None); |
| 65 | }; |
| 66 | let Some((scheme, encoded)) = header.split_once(' ') else { |
| 67 | return Ok(None); |
| 68 | }; |
| 69 | if !scheme.eq_ignore_ascii_case("basic") { |
| 70 | return Ok(None); |
| 71 | } |
| 72 | let Some(decoded) = decode_base64(encoded.trim()) else { |
| 73 | return Ok(None); |
| 74 | }; |
| 75 | let Some((username, secret)) = decoded.split_once(':') else { |
| 76 | return Ok(None); |
| 77 | }; |
| 78 | g1t_kit::call( |
| 79 | identity, |
| 80 | "user_for_git_credentials", |
| 81 | &GitCredentialsArgs { |
| 82 | username: username.to_owned(), |
| 83 | secret: secret.to_owned(), |
| 84 | }, |
| 85 | ) |
| 86 | .await |
| 87 | } |
| 88 | |
| 89 | /// Standard base64 to a UTF-8 string, or `None` if either step fails. |
| 90 | fn decode_base64(input: &str) -> Option<String> { |
| 91 | let mut bytes = Vec::with_capacity(input.len() * 3 / 4); |
| 92 | let mut buffer = 0u32; |
| 93 | let mut bits = 0; |
| 94 | for byte in input.bytes().filter(|byte| *byte != b'=') { |
| 95 | let value = match byte { |
| 96 | b'A'..=b'Z' => byte - b'A', |
| 97 | b'a'..=b'z' => byte - b'a' + 26, |
| 98 | b'0'..=b'9' => byte - b'0' + 52, |
| 99 | b'+' => 62, |
| 100 | b'/' => 63, |
| 101 | _ => return None, |
| 102 | }; |
| 103 | buffer = (buffer << 6) | u32::from(value); |
| 104 | bits += 6; |
| 105 | if bits >= 8 { |
| 106 | bits -= 8; |
| 107 | bytes.push((buffer >> bits) as u8); |
| 108 | } |
| 109 | } |
| 110 | String::from_utf8(bytes).ok() |
| 111 | } |
| 112 | |
| 113 | /// The response for a refused git request. Anonymous callers are asked to |
| 114 | /// authenticate, which is what makes git prompt for credentials. |
| 115 | pub fn refuse<T>(outcome: Outcome<T>) -> Result<Response> { |
| 116 | let Outcome::Fail(failure) = outcome else { |
| 117 | return Response::error("Not found", 404); |
| 118 | }; |
| 119 | let mut response = Response::error(failure.message, failure.code.http_status())?; |
| 120 | if failure.code == FailureCode::Unauthenticated { |
| 121 | response |
| 122 | .headers_mut() |
| 123 | .set("www-authenticate", "Basic realm=\"g1t\"")?; |
| 124 | } |
| 125 | Ok(response) |
| 126 | } |
| 127 | |
| 128 | const ZERO_ID: &str = "0000000000000000000000000000000000000000"; |
| 129 | const HEADS: &str = "refs/heads/"; |
| 130 | const TAGS: &str = "refs/tags/"; |
| 131 | |
| 132 | /// One ref a push asks to change. |
| 133 | struct Command { |
| 134 | old: String, |
| 135 | new: String, |
| 136 | name: String, |
| 137 | } |
| 138 | |
| 139 | /// The commands at the start of a receive-pack request, and the |
| 140 | /// capabilities the client sent with the first of them. |
| 141 | fn commands(body: &[u8]) -> (Vec<Command>, String) { |
| 142 | let mut commands = Vec::new(); |
| 143 | let mut capabilities = String::new(); |
| 144 | let mut position = 0; |
| 145 | // Commands are pkt-lines; a flush packet ends them and the pack follows. |
| 146 | while let Some(length) = body |
| 147 | .get(position..position + 4) |
| 148 | .and_then(|hex| std::str::from_utf8(hex).ok()) |
| 149 | .and_then(|hex| usize::from_str_radix(hex, 16).ok()) |
| 150 | { |
| 151 | if length < 4 || position + length > body.len() { |
| 152 | break; |
| 153 | } |
| 154 | let line = &body[position + 4..position + length]; |
| 155 | position += length; |
| 156 | // `<old> <new> <ref>`, and on the first command a NUL then capabilities. |
| 157 | let mut halves = line.splitn(2, |byte| *byte == 0); |
| 158 | let command = halves.next().unwrap_or_default(); |
| 159 | if let Some(rest) = halves.next() { |
| 160 | capabilities = String::from_utf8_lossy(rest).trim().to_owned(); |
| 161 | } |
| 162 | let Ok(command) = std::str::from_utf8(command) else { |
| 163 | continue; |
| 164 | }; |
| 165 | let mut parts = command.trim_end().splitn(3, ' '); |
| 166 | if let (Some(old), Some(new), Some(name)) = (parts.next(), parts.next(), parts.next()) { |
| 167 | commands.push(Command { |
| 168 | old: old.to_owned(), |
| 169 | new: new.to_owned(), |
| 170 | name: name.to_owned(), |
| 171 | }); |
| 172 | } |
| 173 | } |
| 174 | (commands, capabilities) |
| 175 | } |
| 176 | |
| 177 | fn pkt_line(payload: &[u8]) -> Vec<u8> { |
| 178 | let mut line = format!("{:04x}", payload.len() + 4).into_bytes(); |
| 179 | line.extend_from_slice(payload); |
| 180 | line |
| 181 | } |
| 182 | |
| 183 | /// What git is told when a push would change a protected branch: every ref |
| 184 | /// in it is declined, with the reason against the protected one, so that |
| 185 | /// git prints it beside the branch. `None` if the push leaves the branch |
| 186 | /// alone, or creates it in a repository that does not have it yet. |
| 187 | fn refusal(body: &[u8], protected: &str) -> Option<Vec<u8>> { |
| 188 | let (commands, capabilities) = commands(body); |
| 189 | let reference = format!("{HEADS}{protected}"); |
| 190 | if !commands |
| 191 | .iter() |
| 192 | .any(|command| command.name == reference && command.old != ZERO_ID) |
| 193 | { |
| 194 | return None; |
| 195 | } |
| 196 | let mut report = pkt_line(b"unpack ok\n"); |
| 197 | for command in &commands { |
| 198 | let reason = if command.name == reference { |
| 199 | format!("{protected} is protected: push a branch and open a pull request") |
| 200 | } else { |
| 201 | format!("not pushed, because the same push would change {protected}") |
| 202 | }; |
| 203 | report.extend(pkt_line( |
| 204 | format!("ng {} {reason}\n", command.name).as_bytes(), |
| 205 | )); |
| 206 | } |
| 207 | report.extend_from_slice(b"0000"); |
| 208 | // With side-band the report travels inside channel 1. |
| 209 | let sideband = capabilities |
| 210 | .split(' ') |
| 211 | .any(|capability| capability.starts_with("side-band")); |
| 212 | Some(if sideband { |
| 213 | let mut framed = vec![1u8]; |
| 214 | framed.extend(report); |
| 215 | let mut body = pkt_line(&framed); |
| 216 | body.extend_from_slice(b"0000"); |
| 217 | body |
| 218 | } else { |
| 219 | report |
| 220 | }) |
| 221 | } |
| 222 | |
| 223 | /// A branch or tag a push asks to move. |
| 224 | #[derive(Debug, PartialEq, Eq)] |
| 225 | pub struct Pushed { |
| 226 | /// The full ref: `refs/heads/main`, `refs/tags/v1`. |
| 227 | pub git_ref: String, |
| 228 | /// Where it pointed before; `None` for a new ref. |
| 229 | pub before: Option<String>, |
| 230 | pub after: String, |
| 231 | } |
| 232 | |
| 233 | impl Pushed { |
| 234 | pub fn branch(&self) -> Option<&str> { |
| 235 | self.git_ref.strip_prefix(HEADS) |
| 236 | } |
| 237 | } |
| 238 | |
| 239 | /// The branches and tags a push asks to move, read from the commands at the |
| 240 | /// start of a receive-pack request. Deletions and other refs are left out. |
| 241 | fn pushed_branches(body: &[u8]) -> Vec<Pushed> { |
| 242 | commands(body) |
| 243 | .0 |
| 244 | .into_iter() |
| 245 | .filter(|command| command.new != ZERO_ID) |
| 246 | .filter(|command| command.name.starts_with(HEADS) || command.name.starts_with(TAGS)) |
| 247 | .map(|Command { old, new, name }| Pushed { |
| 248 | git_ref: name, |
| 249 | before: (old != ZERO_ID).then_some(old), |
| 250 | after: new, |
| 251 | }) |
| 252 | .collect() |
| 253 | } |
| 254 | |
| 255 | /// The git store's answer, and what the request asked it to change. |
| 256 | pub struct Forwarded { |
| 257 | pub response: Response, |
| 258 | /// For a push: the branches and tags it asks to move, and the commits |
| 259 | /// to move them to. Whether each moved is for the caller to confirm. |
| 260 | pub pushed: Vec<Pushed>, |
| 261 | } |
| 262 | |
| 263 | /// What became of a git request. |
| 264 | pub enum Push { |
| 265 | Forwarded(Forwarded), |
| 266 | /// A push to a protected branch, answered here without reaching the store. |
| 267 | Refused(Response), |
| 268 | } |
| 269 | |
| 270 | /// Sends the request on to the git store and returns its response as is, |
| 271 | /// unless it is a push that would change the `protected` branch. |
| 272 | pub async fn forward( |
| 273 | mut request: Request, |
| 274 | git: &GitRequest, |
| 275 | access: &GitAccess, |
| 276 | protected: Option<&str>, |
| 277 | ) -> Result<Push> { |
| 278 | let headers = Headers::new(); |
| 279 | headers.set("authorization", &format!("Bearer {}", access.token))?; |
| 280 | for name in FORWARDED_HEADERS { |
| 281 | if let Some(value) = request.headers().get(name)? { |
| 282 | headers.set(name, &value)?; |
| 283 | } |
| 284 | } |
| 285 | let query = request |
| 286 | .url()? |
| 287 | .query() |
| 288 | .map(|query| format!("?{query}")) |
| 289 | .unwrap_or_default(); |
| 290 | let mut init = RequestInit::new(); |
| 291 | init.with_method(request.method()).with_headers(headers); |
| 292 | let mut pushed = Vec::new(); |
| 293 | if request.method() == Method::Post { |
| 294 | // Pushes are capped at 100 MB by the platform, so buffering is safe. |
| 295 | let body = request.bytes().await?; |
| 296 | if git.endpoint == "git-receive-pack" { |
| 297 | if let Some(report) = protected.and_then(|branch| refusal(&body, branch)) { |
| 298 | let headers = Headers::new(); |
| 299 | headers.set("content-type", "application/x-git-receive-pack-result")?; |
| 300 | headers.set("cache-control", "no-cache")?; |
| 301 | return Ok(Push::Refused( |
| 302 | Response::from_bytes(report)?.with_headers(headers), |
| 303 | )); |
| 304 | } |
| 305 | pushed = pushed_branches(&body); |
| 306 | } |
| 307 | init.with_body(Some(Uint8Array::from(body.as_slice()).into())); |
| 308 | } |
| 309 | let upstream = |
| 310 | Request::new_with_init(&format!("{}/{}{query}", access.remote, git.endpoint), &init)?; |
| 311 | Ok(Push::Forwarded(Forwarded { |
| 312 | response: Fetch::Request(upstream).send().await?, |
| 313 | pushed, |
| 314 | })) |
| 315 | } |
| 316 | |
| 317 | #[cfg(test)] |
| 318 | mod tests { |
| 319 | use super::{Pushed, ZERO_ID, pushed_branches, refusal}; |
| 320 | |
| 321 | fn pkt(payload: &str) -> Vec<u8> { |
| 322 | format!("{:04x}{payload}", payload.len() + 4).into_bytes() |
| 323 | } |
| 324 | |
| 325 | #[test] |
| 326 | fn pushed_branches_are_read_from_the_commands() { |
| 327 | let old = "c71546fcd893ef8b0f57388b65e620d759705dda"; |
| 328 | let new = "4807077b296e6edbf410d55e72749d3e1170c291"; |
| 329 | let body = [ |
| 330 | pkt(&format!( |
| 331 | "{old} {new} refs/heads/main\0 report-status side-band-64k\n" |
| 332 | )), |
| 333 | pkt(&format!("{ZERO_ID} {new} refs/heads/feature/x\n")), |
| 334 | pkt(&format!("{old} {ZERO_ID} refs/heads/gone\n")), |
| 335 | pkt(&format!("{ZERO_ID} {new} refs/tags/v1\n")), |
| 336 | b"0000".to_vec(), |
| 337 | b"PACK\0\0\0\x02\0\0\0\0".to_vec(), |
| 338 | ] |
| 339 | .concat(); |
| 340 | assert_eq!( |
| 341 | pushed_branches(&body), |
| 342 | [ |
| 343 | Pushed { |
| 344 | git_ref: "refs/heads/main".to_owned(), |
| 345 | before: Some(old.to_owned()), |
| 346 | after: new.to_owned() |
| 347 | }, |
| 348 | Pushed { |
| 349 | git_ref: "refs/heads/feature/x".to_owned(), |
| 350 | before: None, |
| 351 | after: new.to_owned() |
| 352 | }, |
| 353 | Pushed { |
| 354 | git_ref: "refs/tags/v1".to_owned(), |
| 355 | before: None, |
| 356 | after: new.to_owned() |
| 357 | }, |
| 358 | ] |
| 359 | ); |
| 360 | } |
| 361 | |
| 362 | #[test] |
| 363 | fn a_push_to_a_protected_branch_is_declined_with_the_reason() { |
| 364 | let old = "c71546fcd893ef8b0f57388b65e620d759705dda"; |
| 365 | let new = "4807077b296e6edbf410d55e72749d3e1170c291"; |
| 366 | let body = [ |
| 367 | pkt(&format!("{old} {new} refs/heads/main\0 report-status\n")), |
| 368 | pkt(&format!("{ZERO_ID} {new} refs/heads/feature\n")), |
| 369 | b"0000".to_vec(), |
| 370 | ] |
| 371 | .concat(); |
| 372 | let report = String::from_utf8(refusal(&body, "main").unwrap()).unwrap(); |
| 373 | assert!(report.starts_with("000eunpack ok\n")); |
| 374 | assert!(report.contains("ng refs/heads/main main is protected")); |
| 375 | assert!(report.contains("ng refs/heads/feature not pushed")); |
| 376 | assert!(report.ends_with("0000")); |
| 377 | } |
| 378 | |
| 379 | #[test] |
| 380 | fn the_report_is_framed_for_a_client_that_asked_for_side_band() { |
| 381 | let old = "c71546fcd893ef8b0f57388b65e620d759705dda"; |
| 382 | let body = [ |
| 383 | pkt(&format!( |
| 384 | "{old} {ZERO_ID} refs/heads/main\0 report-status side-band-64k\n" |
| 385 | )), |
| 386 | b"0000".to_vec(), |
| 387 | ] |
| 388 | .concat(); |
| 389 | let report = refusal(&body, "main").unwrap(); |
| 390 | // A length, then channel 1, then the report itself. |
| 391 | assert_eq!(report[4], 1); |
| 392 | assert_eq!(&report[5..18], b"000eunpack ok"); |
| 393 | assert!(report.ends_with(b"00000000")); |
| 394 | } |
| 395 | |
| 396 | #[test] |
| 397 | fn other_branches_and_a_first_push_are_let_through() { |
| 398 | let old = "c71546fcd893ef8b0f57388b65e620d759705dda"; |
| 399 | let new = "4807077b296e6edbf410d55e72749d3e1170c291"; |
| 400 | let feature = [ |
| 401 | pkt(&format!("{old} {new} refs/heads/feature\0 report-status\n")), |
| 402 | b"0000".to_vec(), |
| 403 | ] |
| 404 | .concat(); |
| 405 | assert!(refusal(&feature, "main").is_none()); |
| 406 | // An empty repository has to be able to receive its first commits. |
| 407 | let first = [ |
| 408 | pkt(&format!( |
| 409 | "{ZERO_ID} {new} refs/heads/main\0 report-status\n" |
| 410 | )), |
| 411 | b"0000".to_vec(), |
| 412 | ] |
| 413 | .concat(); |
| 414 | assert!(refusal(&first, "main").is_none()); |
| 415 | } |
| 416 | |
| 417 | #[test] |
| 418 | fn a_fetch_request_names_no_branches() { |
| 419 | assert!( |
| 420 | pushed_branches(b"0032want c71546fcd893ef8b0f57388b65e620d759705dda\n0000").is_empty() |
| 421 | ); |
| 422 | } |
| 423 | } |