diff options
Diffstat (limited to 'src/libfetchers/github.cc')
| -rw-r--r-- | src/libfetchers/github.cc | 304 |
1 files changed, 191 insertions, 113 deletions
diff --git a/src/libfetchers/github.cc b/src/libfetchers/github.cc index 0bee1d6b3..4c1a140ff 100644 --- a/src/libfetchers/github.cc +++ b/src/libfetchers/github.cc @@ -8,81 +8,142 @@ namespace nix::fetchers { -std::regex ownerRegex("[a-zA-Z][a-zA-Z0-9_-]*", std::regex::ECMAScript); -std::regex repoRegex("[a-zA-Z][a-zA-Z0-9_-]*", std::regex::ECMAScript); +// A github or gitlab url +const static std::string urlRegexS = "[a-zA-Z0-9.]*"; // FIXME: check +std::regex urlRegex(urlRegexS, std::regex::ECMAScript); -struct GitHubInput : Input +struct GitArchiveInputScheme : InputScheme { - std::string owner; - std::string repo; - std::optional<std::string> ref; - std::optional<Hash> rev; + virtual std::string type() = 0; - std::string type() const override { return "github"; } - - bool operator ==(const Input & other) const override + std::optional<Input> inputFromURL(const ParsedURL & url) override { - auto other2 = dynamic_cast<const GitHubInput *>(&other); - return - other2 - && owner == other2->owner - && repo == other2->repo - && rev == other2->rev - && ref == other2->ref; + if (url.scheme != type()) return {}; + + auto path = tokenizeString<std::vector<std::string>>(url.path, "/"); + + std::optional<Hash> rev; + std::optional<std::string> ref; + std::optional<std::string> host_url; + + if (path.size() == 2) { + } else if (path.size() == 3) { + if (std::regex_match(path[2], revRegex)) + rev = Hash(path[2], htSHA1); + else if (std::regex_match(path[2], refRegex)) + ref = path[2]; + else + throw BadURL("in URL '%s', '%s' is not a commit hash or branch/tag name", url.url, path[2]); + } else + throw BadURL("URL '%s' is invalid", url.url); + + for (auto &[name, value] : url.query) { + if (name == "rev") { + if (rev) + throw BadURL("URL '%s' contains multiple commit hashes", url.url); + rev = Hash(value, htSHA1); + } + else if (name == "ref") { + if (!std::regex_match(value, refRegex)) + throw BadURL("URL '%s' contains an invalid branch/tag name", url.url); + if (ref) + throw BadURL("URL '%s' contains multiple branch/tag names", url.url); + ref = value; + } + else if (name == "url") { + if (!std::regex_match(value, urlRegex)) + throw BadURL("URL '%s' contains an invalid instance url", url.url); + host_url = value; + } + // FIXME: barf on unsupported attributes + } + + if (ref && rev) + throw BadURL("URL '%s' contains both a commit hash and a branch/tag name %s %s", url.url, *ref, rev->gitRev()); + + Input input; + input.attrs.insert_or_assign("type", type()); + input.attrs.insert_or_assign("owner", path[0]); + input.attrs.insert_or_assign("repo", path[1]); + if (rev) input.attrs.insert_or_assign("rev", rev->gitRev()); + if (ref) input.attrs.insert_or_assign("ref", *ref); + if (host_url) input.attrs.insert_or_assign("url", *host_url); + + return input; } - bool isImmutable() const override + std::optional<Input> inputFromAttrs(const Attrs & attrs) override { - return (bool) rev || narHash; - } + if (maybeGetStrAttr(attrs, "type") != type()) return {}; + + for (auto & [name, value] : attrs) + if (name != "type" && name != "owner" && name != "repo" && name != "ref" && name != "rev" && name != "narHash" && name != "lastModified") + throw Error("unsupported input attribute '%s'", name); - std::optional<std::string> getRef() const override { return ref; } + getStrAttr(attrs, "owner"); + getStrAttr(attrs, "repo"); - std::optional<Hash> getRev() const override { return rev; } + Input input; + input.attrs = attrs; + return input; + } - ParsedURL toURL() const override + ParsedURL toURL(const Input & input) override { + auto owner = getStrAttr(input.attrs, "owner"); + auto repo = getStrAttr(input.attrs, "repo"); + auto ref = input.getRef(); + auto rev = input.getRev(); auto path = owner + "/" + repo; assert(!(ref && rev)); if (ref) path += "/" + *ref; if (rev) path += "/" + rev->to_string(Base16, false); return ParsedURL { - .scheme = "github", + .scheme = type(), .path = path, }; } - Attrs toAttrsInternal() const override + bool hasAllInfo(const Input & input) override { - Attrs attrs; - attrs.emplace("owner", owner); - attrs.emplace("repo", repo); - if (ref) - attrs.emplace("ref", *ref); - if (rev) - attrs.emplace("rev", rev->gitRev()); - return attrs; + return input.getRev() && maybeGetIntAttr(input.attrs, "lastModified"); } - std::pair<Tree, std::shared_ptr<const Input>> fetchTreeInternal(nix::ref<Store> store) const override + Input applyOverrides( + const Input & _input, + std::optional<std::string> ref, + std::optional<Hash> rev) override { - auto rev = this->rev; - auto ref = this->ref.value_or("master"); - - if (!rev) { - auto url = fmt("https://api.github.com/repos/%s/%s/commits/%s", - owner, repo, ref); - auto json = nlohmann::json::parse( - readFile( - store->toRealPath( - downloadFile(store, url, "source", false).storePath))); - rev = Hash(std::string { json["sha"] }, htSHA1); - debug("HEAD revision for '%s' is %s", url, rev->gitRev()); + auto input(_input); + if (rev && ref) + throw BadURL("cannot apply both a commit hash (%s) and a branch/tag name ('%s') to input '%s'", + rev->gitRev(), *ref, input.to_string()); + if (rev) { + input.attrs.insert_or_assign("rev", rev->gitRev()); + input.attrs.erase("ref"); } + if (ref) { + input.attrs.insert_or_assign("ref", *ref); + input.attrs.erase("rev"); + } + return input; + } + + virtual Hash getRevFromRef(nix::ref<Store> store, const Input & input) const = 0; + + virtual std::string getDownloadUrl(const Input & input) const = 0; + + std::pair<Tree, Input> fetch(ref<Store> store, const Input & _input) override + { + Input input(_input); - auto input = std::make_shared<GitHubInput>(*this); - input->ref = {}; - input->rev = *rev; + if (!maybeGetStrAttr(input.attrs, "ref")) input.attrs.insert_or_assign("ref", "HEAD"); + + auto rev = input.getRev(); + if (!rev) rev = getRevFromRef(store, input); + + input.attrs.erase("ref"); + input.attrs.insert_or_assign("rev", rev->gitRev()); Attrs immutableAttrs({ {"type", "git-tarball"}, @@ -90,36 +151,25 @@ struct GitHubInput : Input }); if (auto res = getCache()->lookup(store, immutableAttrs)) { + input.attrs.insert_or_assign("lastModified", getIntAttr(res->first, "lastModified")); return { - Tree{ - .actualPath = store->toRealPath(res->second), - .storePath = std::move(res->second), - .info = TreeInfo { - .lastModified = getIntAttr(res->first, "lastModified"), - }, - }, + Tree(store->toRealPath(res->second), std::move(res->second)), input }; } - // FIXME: use regular /archive URLs instead? api.github.com - // might have stricter rate limits. + auto url = getDownloadUrl(input); - auto url = fmt("https://api.github.com/repos/%s/%s/tarball/%s", - owner, repo, rev->to_string(Base16, false)); + auto [tree, lastModified] = downloadTarball(store, url, "source", true); - std::string accessToken = settings.githubAccessToken.get(); - if (accessToken != "") - url += "?access_token=" + accessToken; - - auto tree = downloadTarball(store, url, "source", true); + input.attrs.insert_or_assign("lastModified", lastModified); getCache()->add( store, immutableAttrs, { {"rev", rev->gitRev()}, - {"lastModified", *tree.info.lastModified} + {"lastModified", lastModified} }, tree.storePath, true); @@ -128,68 +178,96 @@ struct GitHubInput : Input } }; -struct GitHubInputScheme : InputScheme +struct GitHubInputScheme : GitArchiveInputScheme { - std::unique_ptr<Input> inputFromURL(const ParsedURL & url) override + std::string type() override { return "github"; } + + Hash getRevFromRef(nix::ref<Store> store, const Input & input) const override { - if (url.scheme != "github") return nullptr; + auto host_url = maybeGetStrAttr(input.attrs, "url").value_or("github.com"); + auto url = fmt("https://api.%s/repos/%s/%s/commits/%s", // FIXME: check + host_url, getStrAttr(input.attrs, "owner"), getStrAttr(input.attrs, "repo"), *input.getRef()); + auto json = nlohmann::json::parse( + readFile( + store->toRealPath( + downloadFile(store, url, "source", false).storePath))); + auto rev = Hash(std::string { json["sha"] }, htSHA1); + debug("HEAD revision for '%s' is %s", url, rev.gitRev()); + return rev; + } - auto path = tokenizeString<std::vector<std::string>>(url.path, "/"); - auto input = std::make_unique<GitHubInput>(); + std::string getDownloadUrl(const Input & input) const override + { + // FIXME: use regular /archive URLs instead? api.github.com + // might have stricter rate limits. + auto host_url = maybeGetStrAttr(input.attrs, "url").value_or("github.com"); + auto url = fmt("https://api.%s/repos/%s/%s/tarball/%s", // FIXME: check if this is correct for self hosted instances + host_url, getStrAttr(input.attrs, "owner"), getStrAttr(input.attrs, "repo"), + input.getRev()->to_string(Base16, false)); - if (path.size() == 2) { - } else if (path.size() == 3) { - if (std::regex_match(path[2], revRegex)) - input->rev = Hash(path[2], htSHA1); - else if (std::regex_match(path[2], refRegex)) - input->ref = path[2]; - else - throw BadURL("in GitHub URL '%s', '%s' is not a commit hash or branch/tag name", url.url, path[2]); - } else - throw BadURL("GitHub URL '%s' is invalid", url.url); + std::string accessToken = settings.githubAccessToken.get(); + if (accessToken != "") + url += "?access_token=" + accessToken; - for (auto &[name, value] : url.query) { - if (name == "rev") { - if (input->rev) - throw BadURL("GitHub URL '%s' contains multiple commit hashes", url.url); - input->rev = Hash(value, htSHA1); - } - else if (name == "ref") { - if (!std::regex_match(value, refRegex)) - throw BadURL("GitHub URL '%s' contains an invalid branch/tag name", url.url); - if (input->ref) - throw BadURL("GitHub URL '%s' contains multiple branch/tag names", url.url); - input->ref = value; - } - } + return url; + } - if (input->ref && input->rev) - throw BadURL("GitHub URL '%s' contains both a commit hash and a branch/tag name", url.url); + void clone(const Input & input, const Path & destDir) override + { + auto host_url = maybeGetStrAttr(input.attrs, "url").value_or("github.com"); + Input::fromURL(fmt("git+ssh://git@%s/%s/%s.git", + host_url, getStrAttr(input.attrs, "owner"), getStrAttr(input.attrs, "repo"))) + .applyOverrides(input.getRef().value_or("HEAD"), input.getRev()) + .clone(destDir); + } +}; - input->owner = path[0]; - input->repo = path[1]; +struct GitLabInputScheme : GitArchiveInputScheme +{ + std::string type() override { return "gitlab"; } - return input; + Hash getRevFromRef(nix::ref<Store> store, const Input & input) const override + { + auto host_url = maybeGetStrAttr(input.attrs, "url").value_or("gitlab.com"); + auto url = fmt("https://%s/api/v4/projects/%s%%2F%s/repository/branches/%s", + host_url, getStrAttr(input.attrs, "owner"), getStrAttr(input.attrs, "repo"), *input.getRef()); + auto json = nlohmann::json::parse( + readFile( + store->toRealPath( + downloadFile(store, url, "source", false).storePath))); + auto rev = Hash(std::string(json["commit"]["id"]), htSHA1); + debug("HEAD revision for '%s' is %s", url, rev.gitRev()); + return rev; } - std::unique_ptr<Input> inputFromAttrs(const Attrs & attrs) override + std::string getDownloadUrl(const Input & input) const override { - if (maybeGetStrAttr(attrs, "type") != "github") return {}; + // FIXME: This endpoint has a rate limit threshold of 5 requests per minute + auto host_url = maybeGetStrAttr(input.attrs, "url").value_or("gitlab.com"); + auto url = fmt("https://%s/api/v4/projects/%s%%2F%s/repository/archive.tar.gz?sha=%s", + host_url, getStrAttr(input.attrs, "owner"), getStrAttr(input.attrs, "repo"), + input.getRev()->to_string(Base16, false)); - for (auto & [name, value] : attrs) - if (name != "type" && name != "owner" && name != "repo" && name != "ref" && name != "rev") - throw Error("unsupported GitHub input attribute '%s'", name); - - auto input = std::make_unique<GitHubInput>(); - input->owner = getStrAttr(attrs, "owner"); - input->repo = getStrAttr(attrs, "repo"); - input->ref = maybeGetStrAttr(attrs, "ref"); - if (auto rev = maybeGetStrAttr(attrs, "rev")) - input->rev = Hash(*rev, htSHA1); - return input; + /* # FIXME: add privat token auth (`curl --header "PRIVATE-TOKEN: <your_access_token>"`) + std::string accessToken = settings.githubAccessToken.get(); + if (accessToken != "") + url += "?access_token=" + accessToken;*/ + + return url; + } + + void clone(const Input & input, const Path & destDir) override + { + auto host_url = maybeGetStrAttr(input.attrs, "url").value_or("gitlab.com"); + // FIXME: get username somewhere + Input::fromURL(fmt("git+ssh://git@%s/%s/%s.git", + host_url, getStrAttr(input.attrs, "owner"), getStrAttr(input.attrs, "repo"))) + .applyOverrides(input.getRef().value_or("HEAD"), input.getRev()) + .clone(destDir); } }; static auto r1 = OnStartup([] { registerInputScheme(std::make_unique<GitHubInputScheme>()); }); +static auto r2 = OnStartup([] { registerInputScheme(std::make_unique<GitLabInputScheme>()); }); } |