g1t/services/repos/src/git_http.rs
Pick any line to see why it is the way it is: the commit, the pull request and issue it came from, and what the agent was thinking.
| Rust repos service with shipping; pull requests kept in the model | 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 | ||
| Events service in Rust, with RFC 3339 times and accurate push events | 128 | const ZERO_ID: &str = "0000000000000000000000000000000000000000"; |
| 129 | const HEADS: &str = "refs/heads/"; | |
| 130 | ||
| Agents as a team: lifecycle, merge queue, billing and a new shell | 131 | /// One ref a push asks to change. |
| 132 | struct Command { | |
| 133 | old: String, | |
| 134 | new: String, | |
| 135 | name: String, | |
| 136 | } | |
| 137 | ||
| 138 | /// The commands at the start of a receive-pack request, and the | |
| 139 | /// capabilities the client sent with the first of them. | |
| 140 | fn commands(body: &[u8]) -> (Vec<Command>, String) { | |
| 141 | let mut commands = Vec::new(); | |
| 142 | let mut capabilities = String::new(); | |
| Events service in Rust, with RFC 3339 times and accurate push events | 143 | let mut position = 0; |
| 144 | // Commands are pkt-lines; a flush packet ends them and the pack follows. | |
| 145 | while let Some(length) = body | |
| 146 | .get(position..position + 4) | |
| 147 | .and_then(|hex| std::str::from_utf8(hex).ok()) | |
| 148 | .and_then(|hex| usize::from_str_radix(hex, 16).ok()) | |
| 149 | { | |
| 150 | if length < 4 || position + length > body.len() { | |
| 151 | break; | |
| 152 | } | |
| 153 | let line = &body[position + 4..position + length]; | |
| 154 | position += length; | |
| 155 | // `<old> <new> <ref>`, and on the first command a NUL then capabilities. | |
| Agents as a team: lifecycle, merge queue, billing and a new shell | 156 | let mut halves = line.splitn(2, |byte| *byte == 0); |
| 157 | let command = halves.next().unwrap_or_default(); | |
| 158 | if let Some(rest) = halves.next() { | |
| 159 | capabilities = String::from_utf8_lossy(rest).trim().to_owned(); | |
| 160 | } | |
| 161 | let Ok(command) = std::str::from_utf8(command) else { | |
| Events service in Rust, with RFC 3339 times and accurate push events | 162 | continue; |
| 163 | }; | |
| Agents as a team: lifecycle, merge queue, billing and a new shell | 164 | let mut parts = command.trim_end().splitn(3, ' '); |
| 165 | if let (Some(old), Some(new), Some(name)) = (parts.next(), parts.next(), parts.next()) { | |
| 166 | commands.push(Command { | |
| 167 | old: old.to_owned(), | |
| 168 | new: new.to_owned(), | |
| 169 | name: name.to_owned(), | |
| 170 | }); | |
| Events service in Rust, with RFC 3339 times and accurate push events | 171 | } |
| 172 | } | |
| Agents as a team: lifecycle, merge queue, billing and a new shell | 173 | (commands, capabilities) |
| Events service in Rust, with RFC 3339 times and accurate push events | 174 | } |
| 175 | ||
| Agents as a team: lifecycle, merge queue, billing and a new shell | 176 | fn pkt_line(payload: &[u8]) -> Vec<u8> { |
| 177 | let mut line = format!("{:04x}", payload.len() + 4).into_bytes(); | |
| 178 | line.extend_from_slice(payload); | |
| 179 | line | |
| 180 | } | |
| 181 | ||
| 182 | /// What git is told when a push would change a protected branch: every ref | |
| 183 | /// in it is declined, with the reason against the protected one, so that | |
| 184 | /// git prints it beside the branch. `None` if the push leaves the branch | |
| 185 | /// alone, or creates it in a repository that does not have it yet. | |
| 186 | fn refusal(body: &[u8], protected: &str) -> Option<Vec<u8>> { | |
| 187 | let (commands, capabilities) = commands(body); | |
| 188 | let reference = format!("{HEADS}{protected}"); | |
| 189 | if !commands | |
| 190 | .iter() | |
| 191 | .any(|command| command.name == reference && command.old != ZERO_ID) | |
| 192 | { | |
| 193 | return None; | |
| 194 | } | |
| 195 | let mut report = pkt_line(b"unpack ok\n"); | |
| 196 | for command in &commands { | |
| 197 | let reason = if command.name == reference { | |
| 198 | format!("{protected} is protected: push a branch and open a pull request") | |
| 199 | } else { | |
| 200 | format!("not pushed, because the same push would change {protected}") | |
| 201 | }; | |
| 202 | report.extend(pkt_line( | |
| 203 | format!("ng {} {reason}\n", command.name).as_bytes(), | |
| 204 | )); | |
| 205 | } | |
| 206 | report.extend_from_slice(b"0000"); | |
| 207 | // With side-band the report travels inside channel 1. | |
| 208 | let sideband = capabilities | |
| 209 | .split(' ') | |
| 210 | .any(|capability| capability.starts_with("side-band")); | |
| 211 | Some(if sideband { | |
| 212 | let mut framed = vec![1u8]; | |
| 213 | framed.extend(report); | |
| 214 | let mut body = pkt_line(&framed); | |
| 215 | body.extend_from_slice(b"0000"); | |
| 216 | body | |
| 217 | } else { | |
| 218 | report | |
| 219 | }) | |
| 220 | } | |
| 221 | ||
| 222 | /// The branches a push asks to move, as `(branch, new commit)`, read from | |
| 223 | /// the commands at the start of a receive-pack request. Deletions and refs | |
| 224 | /// that are not branches are left out. | |
| 225 | fn pushed_branches(body: &[u8]) -> Vec<(String, String)> { | |
| 226 | commands(body) | |
| 227 | .0 | |
| 228 | .into_iter() | |
| 229 | .filter(|command| command.new != ZERO_ID) | |
| 230 | .filter_map(|Command { new, name, .. }| { | |
| 231 | name.strip_prefix(HEADS) | |
| 232 | .map(|branch| (branch.to_owned(), new)) | |
| 233 | }) | |
| 234 | .collect() | |
| 235 | } | |
| 236 | ||
| Events service in Rust, with RFC 3339 times and accurate push events | 237 | /// The git store's answer, and what the request asked it to change. |
| 238 | pub struct Forwarded { | |
| 239 | pub response: Response, | |
| 240 | /// For a push: the branches it asks to move and the commits to move | |
| 241 | /// them to. Whether each moved is for the caller to confirm. | |
| 242 | pub pushed: Vec<(String, String)>, | |
| 243 | } | |
| 244 | ||
| Agents as a team: lifecycle, merge queue, billing and a new shell | 245 | /// What became of a git request. |
| 246 | pub enum Push { | |
| 247 | Forwarded(Forwarded), | |
| 248 | /// A push to a protected branch, answered here without reaching the store. | |
| 249 | Refused(Response), | |
| 250 | } | |
| 251 | ||
| 252 | /// Sends the request on to the git store and returns its response as is, | |
| 253 | /// unless it is a push that would change the `protected` branch. | |
| Rust repos service with shipping; pull requests kept in the model | 254 | pub async fn forward( |
| 255 | mut request: Request, | |
| 256 | git: &GitRequest, | |
| 257 | access: &GitAccess, | |
| Agents as a team: lifecycle, merge queue, billing and a new shell | 258 | protected: Option<&str>, |
| 259 | ) -> Result<Push> { | |
| Rust repos service with shipping; pull requests kept in the model | 260 | let headers = Headers::new(); |
| 261 | headers.set("authorization", &format!("Bearer {}", access.token))?; | |
| 262 | for name in FORWARDED_HEADERS { | |
| 263 | if let Some(value) = request.headers().get(name)? { | |
| 264 | headers.set(name, &value)?; | |
| 265 | } | |
| 266 | } | |
| 267 | let query = request | |
| 268 | .url()? | |
| 269 | .query() | |
| 270 | .map(|query| format!("?{query}")) | |
| 271 | .unwrap_or_default(); | |
| 272 | let mut init = RequestInit::new(); | |
| 273 | init.with_method(request.method()).with_headers(headers); | |
| Events service in Rust, with RFC 3339 times and accurate push events | 274 | let mut pushed = Vec::new(); |
| Rust repos service with shipping; pull requests kept in the model | 275 | if request.method() == Method::Post { |
| 276 | // Pushes are capped at 100 MB by the platform, so buffering is safe. | |
| 277 | let body = request.bytes().await?; | |
| Events service in Rust, with RFC 3339 times and accurate push events | 278 | if git.endpoint == "git-receive-pack" { |
| Agents as a team: lifecycle, merge queue, billing and a new shell | 279 | if let Some(report) = protected.and_then(|branch| refusal(&body, branch)) { |
| 280 | let headers = Headers::new(); | |
| 281 | headers.set("content-type", "application/x-git-receive-pack-result")?; | |
| 282 | headers.set("cache-control", "no-cache")?; | |
| 283 | return Ok(Push::Refused( | |
| 284 | Response::from_bytes(report)?.with_headers(headers), | |
| 285 | )); | |
| 286 | } | |
| Events service in Rust, with RFC 3339 times and accurate push events | 287 | pushed = pushed_branches(&body); |
| 288 | } | |
| Rust repos service with shipping; pull requests kept in the model | 289 | init.with_body(Some(Uint8Array::from(body.as_slice()).into())); |
| 290 | } | |
| 291 | let upstream = | |
| 292 | Request::new_with_init(&format!("{}/{}{query}", access.remote, git.endpoint), &init)?; | |
| Agents as a team: lifecycle, merge queue, billing and a new shell | 293 | Ok(Push::Forwarded(Forwarded { |
| Events service in Rust, with RFC 3339 times and accurate push events | 294 | response: Fetch::Request(upstream).send().await?, |
| 295 | pushed, | |
| Agents as a team: lifecycle, merge queue, billing and a new shell | 296 | })) |
| Events service in Rust, with RFC 3339 times and accurate push events | 297 | } |
| 298 | ||
| 299 | #[cfg(test)] | |
| 300 | mod tests { | |
| Agents as a team: lifecycle, merge queue, billing and a new shell | 301 | use super::{ZERO_ID, pushed_branches, refusal}; |
| Events service in Rust, with RFC 3339 times and accurate push events | 302 | |
| 303 | fn pkt(payload: &str) -> Vec<u8> { | |
| 304 | format!("{:04x}{payload}", payload.len() + 4).into_bytes() | |
| 305 | } | |
| 306 | ||
| 307 | #[test] | |
| 308 | fn pushed_branches_are_read_from_the_commands() { | |
| 309 | let old = "c71546fcd893ef8b0f57388b65e620d759705dda"; | |
| 310 | let new = "4807077b296e6edbf410d55e72749d3e1170c291"; | |
| 311 | let body = [ | |
| 312 | pkt(&format!( | |
| 313 | "{old} {new} refs/heads/main\0 report-status side-band-64k\n" | |
| 314 | )), | |
| 315 | pkt(&format!("{ZERO_ID} {new} refs/heads/feature/x\n")), | |
| 316 | pkt(&format!("{old} {ZERO_ID} refs/heads/gone\n")), | |
| 317 | pkt(&format!("{ZERO_ID} {new} refs/tags/v1\n")), | |
| 318 | b"0000".to_vec(), | |
| 319 | b"PACK\0\0\0\x02\0\0\0\0".to_vec(), | |
| 320 | ] | |
| 321 | .concat(); | |
| 322 | assert_eq!( | |
| 323 | pushed_branches(&body), | |
| 324 | [ | |
| 325 | ("main".to_owned(), new.to_owned()), | |
| 326 | ("feature/x".to_owned(), new.to_owned()), | |
| 327 | ] | |
| 328 | ); | |
| 329 | } | |
| 330 | ||
| 331 | #[test] | |
| Agents as a team: lifecycle, merge queue, billing and a new shell | 332 | fn a_push_to_a_protected_branch_is_declined_with_the_reason() { |
| 333 | let old = "c71546fcd893ef8b0f57388b65e620d759705dda"; | |
| 334 | let new = "4807077b296e6edbf410d55e72749d3e1170c291"; | |
| 335 | let body = [ | |
| 336 | pkt(&format!("{old} {new} refs/heads/main\0 report-status\n")), | |
| 337 | pkt(&format!("{ZERO_ID} {new} refs/heads/feature\n")), | |
| 338 | b"0000".to_vec(), | |
| 339 | ] | |
| 340 | .concat(); | |
| 341 | let report = String::from_utf8(refusal(&body, "main").unwrap()).unwrap(); | |
| 342 | assert!(report.starts_with("000eunpack ok\n")); | |
| 343 | assert!(report.contains("ng refs/heads/main main is protected")); | |
| 344 | assert!(report.contains("ng refs/heads/feature not pushed")); | |
| 345 | assert!(report.ends_with("0000")); | |
| 346 | } | |
| 347 | ||
| 348 | #[test] | |
| 349 | fn the_report_is_framed_for_a_client_that_asked_for_side_band() { | |
| 350 | let old = "c71546fcd893ef8b0f57388b65e620d759705dda"; | |
| 351 | let body = [ | |
| 352 | pkt(&format!( | |
| 353 | "{old} {ZERO_ID} refs/heads/main\0 report-status side-band-64k\n" | |
| 354 | )), | |
| 355 | b"0000".to_vec(), | |
| 356 | ] | |
| 357 | .concat(); | |
| 358 | let report = refusal(&body, "main").unwrap(); | |
| 359 | // A length, then channel 1, then the report itself. | |
| 360 | assert_eq!(report[4], 1); | |
| 361 | assert_eq!(&report[5..18], b"000eunpack ok"); | |
| 362 | assert!(report.ends_with(b"00000000")); | |
| 363 | } | |
| 364 | ||
| 365 | #[test] | |
| 366 | fn other_branches_and_a_first_push_are_let_through() { | |
| 367 | let old = "c71546fcd893ef8b0f57388b65e620d759705dda"; | |
| 368 | let new = "4807077b296e6edbf410d55e72749d3e1170c291"; | |
| 369 | let feature = [ | |
| 370 | pkt(&format!("{old} {new} refs/heads/feature\0 report-status\n")), | |
| 371 | b"0000".to_vec(), | |
| 372 | ] | |
| 373 | .concat(); | |
| 374 | assert!(refusal(&feature, "main").is_none()); | |
| 375 | // An empty repository has to be able to receive its first commits. | |
| 376 | let first = [ | |
| 377 | pkt(&format!( | |
| 378 | "{ZERO_ID} {new} refs/heads/main\0 report-status\n" | |
| 379 | )), | |
| 380 | b"0000".to_vec(), | |
| 381 | ] | |
| 382 | .concat(); | |
| 383 | assert!(refusal(&first, "main").is_none()); | |
| 384 | } | |
| 385 | ||
| 386 | #[test] | |
| Events service in Rust, with RFC 3339 times and accurate push events | 387 | fn a_fetch_request_names_no_branches() { |
| 388 | assert!( | |
| 389 | pushed_branches(b"0032want c71546fcd893ef8b0f57388b65e620d759705dda\n0000").is_empty() | |
| 390 | ); | |
| 391 | } | |
| Rust repos service with shipping; pull requests kept in the model | 392 | } |