Merge remote-tracking branch 'origin/master' into enum-FileIngestionMethod

14 files changed, 1892 insertions, 0 deletions

diff --git a/src/libfetchers/attrs.cc b/src/libfetchers/attrs.cc
new file mode 100644
index 000000000..feb0a6085
--- /dev/null
+++ b/src/libfetchers/attrs.cc
@@ -0,0 +1,107 @@
+#include "attrs.hh"
+#include "fetchers.hh"
+
+#include <nlohmann/json.hpp>
+
+namespace nix::fetchers {
+
+Attrs jsonToAttrs(const nlohmann::json & json)
+{
+    Attrs attrs;
+
+    for (auto & i : json.items()) {
+        if (i.value().is_number())
+            attrs.emplace(i.key(), i.value().get<int64_t>());
+        else if (i.value().is_string())
+            attrs.emplace(i.key(), i.value().get<std::string>());
+        else if (i.value().is_boolean())
+            attrs.emplace(i.key(), i.value().get<bool>());
+        else
+            throw Error("unsupported input attribute type in lock file");
+    }
+
+    return attrs;
+}
+
+nlohmann::json attrsToJson(const Attrs & attrs)
+{
+    nlohmann::json json;
+    for (auto & attr : attrs) {
+        if (auto v = std::get_if<int64_t>(&attr.second)) {
+            json[attr.first] = *v;
+        } else if (auto v = std::get_if<std::string>(&attr.second)) {
+            json[attr.first] = *v;
+        } else if (auto v = std::get_if<Explicit<bool>>(&attr.second)) {
+            json[attr.first] = v->t;
+        } else abort();
+    }
+    return json;
+}
+
+std::optional<std::string> maybeGetStrAttr(const Attrs & attrs, const std::string & name)
+{
+    auto i = attrs.find(name);
+    if (i == attrs.end()) return {};
+    if (auto v = std::get_if<std::string>(&i->second))
+        return *v;
+    throw Error("input attribute '%s' is not a string %s", name, attrsToJson(attrs).dump());
+}
+
+std::string getStrAttr(const Attrs & attrs, const std::string & name)
+{
+    auto s = maybeGetStrAttr(attrs, name);
+    if (!s)
+        throw Error("input attribute '%s' is missing", name);
+    return *s;
+}
+
+std::optional<int64_t> maybeGetIntAttr(const Attrs & attrs, const std::string & name)
+{
+    auto i = attrs.find(name);
+    if (i == attrs.end()) return {};
+    if (auto v = std::get_if<int64_t>(&i->second))
+        return *v;
+    throw Error("input attribute '%s' is not an integer", name);
+}
+
+int64_t getIntAttr(const Attrs & attrs, const std::string & name)
+{
+    auto s = maybeGetIntAttr(attrs, name);
+    if (!s)
+        throw Error("input attribute '%s' is missing", name);
+    return *s;
+}
+
+std::optional<bool> maybeGetBoolAttr(const Attrs & attrs, const std::string & name)
+{
+    auto i = attrs.find(name);
+    if (i == attrs.end()) return {};
+    if (auto v = std::get_if<int64_t>(&i->second))
+        return *v;
+    throw Error("input attribute '%s' is not a Boolean", name);
+}
+
+bool getBoolAttr(const Attrs & attrs, const std::string & name)
+{
+    auto s = maybeGetBoolAttr(attrs, name);
+    if (!s)
+        throw Error("input attribute '%s' is missing", name);
+    return *s;
+}
+
+std::map<std::string, std::string> attrsToQuery(const Attrs & attrs)
+{
+    std::map<std::string, std::string> query;
+    for (auto & attr : attrs) {
+        if (auto v = std::get_if<int64_t>(&attr.second)) {
+            query.insert_or_assign(attr.first, fmt("%d", *v));
+        } else if (auto v = std::get_if<std::string>(&attr.second)) {
+            query.insert_or_assign(attr.first, *v);
+        } else if (auto v = std::get_if<Explicit<bool>>(&attr.second)) {
+            query.insert_or_assign(attr.first, v->t ? "1" : "0");
+        } else abort();
+    }
+    return query;
+}
+
+}
diff --git a/src/libfetchers/attrs.hh b/src/libfetchers/attrs.hh
new file mode 100644
index 000000000..d6e0ae000
--- /dev/null
+++ b/src/libfetchers/attrs.hh
@@ -0,0 +1,39 @@
+#pragma once
+
+#include "types.hh"
+
+#include <variant>
+
+#include <nlohmann/json_fwd.hpp>
+
+namespace nix::fetchers {
+
+/* Wrap bools to prevent string literals (i.e. 'char *') from being
+   cast to a bool in Attr. */
+template<typename T>
+struct Explicit {
+    T t;
+};
+
+typedef std::variant<std::string, int64_t, Explicit<bool>> Attr;
+typedef std::map<std::string, Attr> Attrs;
+
+Attrs jsonToAttrs(const nlohmann::json & json);
+
+nlohmann::json attrsToJson(const Attrs & attrs);
+
+std::optional<std::string> maybeGetStrAttr(const Attrs & attrs, const std::string & name);
+
+std::string getStrAttr(const Attrs & attrs, const std::string & name);
+
+std::optional<int64_t> maybeGetIntAttr(const Attrs & attrs, const std::string & name);
+
+int64_t getIntAttr(const Attrs & attrs, const std::string & name);
+
+std::optional<bool> maybeGetBoolAttr(const Attrs & attrs, const std::string & name);
+
+bool getBoolAttr(const Attrs & attrs, const std::string & name);
+
+std::map<std::string, std::string> attrsToQuery(const Attrs & attrs);
+
+}
diff --git a/src/libfetchers/cache.cc b/src/libfetchers/cache.cc
new file mode 100644
index 000000000..e1c7f3dee
--- /dev/null
+++ b/src/libfetchers/cache.cc
@@ -0,0 +1,121 @@
+#include "cache.hh"
+#include "sqlite.hh"
+#include "sync.hh"
+#include "store-api.hh"
+
+#include <nlohmann/json.hpp>
+
+namespace nix::fetchers {
+
+static const char * schema = R"sql(
+
+create table if not exists Cache (
+    input     text not null,
+    info      text not null,
+    path      text not null,
+    immutable integer not null,
+    timestamp integer not null,
+    primary key (input)
+);
+)sql";
+
+struct CacheImpl : Cache
+{
+    struct State
+    {
+        SQLite db;
+        SQLiteStmt add, lookup;
+    };
+
+    Sync<State> _state;
+
+    CacheImpl()
+    {
+        auto state(_state.lock());
+
+        auto dbPath = getCacheDir() + "/nix/fetcher-cache-v1.sqlite";
+        createDirs(dirOf(dbPath));
+
+        state->db = SQLite(dbPath);
+        state->db.isCache();
+        state->db.exec(schema);
+
+        state->add.create(state->db,
+            "insert or replace into Cache(input, info, path, immutable, timestamp) values (?, ?, ?, ?, ?)");
+
+        state->lookup.create(state->db,
+            "select info, path, immutable, timestamp from Cache where input = ?");
+    }
+
+    void add(
+        ref<Store> store,
+        const Attrs & inAttrs,
+        const Attrs & infoAttrs,
+        const StorePath & storePath,
+        bool immutable) override
+    {
+        _state.lock()->add.use()
+            (attrsToJson(inAttrs).dump())
+            (attrsToJson(infoAttrs).dump())
+            (store->printStorePath(storePath))
+            (immutable)
+            (time(0)).exec();
+    }
+
+    std::optional<std::pair<Attrs, StorePath>> lookup(
+        ref<Store> store,
+        const Attrs & inAttrs) override
+    {
+        if (auto res = lookupExpired(store, inAttrs)) {
+            if (!res->expired)
+                return std::make_pair(std::move(res->infoAttrs), std::move(res->storePath));
+            debug("ignoring expired cache entry '%s'",
+                attrsToJson(inAttrs).dump());
+        }
+        return {};
+    }
+
+    std::optional<Result> lookupExpired(
+        ref<Store> store,
+        const Attrs & inAttrs) override
+    {
+        auto state(_state.lock());
+
+        auto inAttrsJson = attrsToJson(inAttrs).dump();
+
+        auto stmt(state->lookup.use()(inAttrsJson));
+        if (!stmt.next()) {
+            debug("did not find cache entry for '%s'", inAttrsJson);
+            return {};
+        }
+
+        auto infoJson = stmt.getStr(0);
+        auto storePath = store->parseStorePath(stmt.getStr(1));
+        auto immutable = stmt.getInt(2) != 0;
+        auto timestamp = stmt.getInt(3);
+
+        store->addTempRoot(storePath);
+        if (!store->isValidPath(storePath)) {
+            // FIXME: we could try to substitute 'storePath'.
+            debug("ignoring disappeared cache entry '%s'", inAttrsJson);
+            return {};
+        }
+
+        debug("using cache entry '%s' -> '%s', '%s'",
+            inAttrsJson, infoJson, store->printStorePath(storePath));
+
+        return Result {
+            .expired = !immutable && (settings.tarballTtl.get() == 0 || timestamp + settings.tarballTtl < time(0)),
+            .infoAttrs = jsonToAttrs(nlohmann::json::parse(infoJson)),
+            .storePath = std::move(storePath)
+        };
+    }
+};
+
+ref<Cache> getCache()
+{
+    static auto cache = std::make_shared<CacheImpl>();
+    return ref<Cache>(cache);
+}
+
+}
diff --git a/src/libfetchers/cache.hh b/src/libfetchers/cache.hh
new file mode 100644
index 000000000..d76ab1233
--- /dev/null
+++ b/src/libfetchers/cache.hh
@@ -0,0 +1,34 @@
+#pragma once
+
+#include "fetchers.hh"
+
+namespace nix::fetchers {
+
+struct Cache
+{
+    virtual void add(
+        ref<Store> store,
+        const Attrs & inAttrs,
+        const Attrs & infoAttrs,
+        const StorePath & storePath,
+        bool immutable) = 0;
+
+    virtual std::optional<std::pair<Attrs, StorePath>> lookup(
+        ref<Store> store,
+        const Attrs & inAttrs) = 0;
+
+    struct Result
+    {
+        bool expired = false;
+        Attrs infoAttrs;
+        StorePath storePath;
+    };
+
+    virtual std::optional<Result> lookupExpired(
+        ref<Store> store,
+        const Attrs & inAttrs) = 0;
+};
+
+ref<Cache> getCache();
+
+}
diff --git a/src/libfetchers/fetchers.cc b/src/libfetchers/fetchers.cc
new file mode 100644
index 000000000..94ac30e38
--- /dev/null
+++ b/src/libfetchers/fetchers.cc
@@ -0,0 +1,75 @@
+#include "fetchers.hh"
+#include "store-api.hh"
+
+#include <nlohmann/json.hpp>
+
+namespace nix::fetchers {
+
+std::unique_ptr<std::vector<std::unique_ptr<InputScheme>>> inputSchemes = nullptr;
+
+void registerInputScheme(std::unique_ptr<InputScheme> && inputScheme)
+{
+    if (!inputSchemes) inputSchemes = std::make_unique<std::vector<std::unique_ptr<InputScheme>>>();
+    inputSchemes->push_back(std::move(inputScheme));
+}
+
+std::unique_ptr<Input> inputFromURL(const ParsedURL & url)
+{
+    for (auto & inputScheme : *inputSchemes) {
+        auto res = inputScheme->inputFromURL(url);
+        if (res) return res;
+    }
+    throw Error("input '%s' is unsupported", url.url);
+}
+
+std::unique_ptr<Input> inputFromURL(const std::string & url)
+{
+    return inputFromURL(parseURL(url));
+}
+
+std::unique_ptr<Input> inputFromAttrs(const Attrs & attrs)
+{
+    auto attrs2(attrs);
+    attrs2.erase("narHash");
+    for (auto & inputScheme : *inputSchemes) {
+        auto res = inputScheme->inputFromAttrs(attrs2);
+        if (res) {
+            if (auto narHash = maybeGetStrAttr(attrs, "narHash"))
+                // FIXME: require SRI hash.
+                res->narHash = Hash(*narHash);
+            return res;
+        }
+    }
+    throw Error("input '%s' is unsupported", attrsToJson(attrs));
+}
+
+Attrs Input::toAttrs() const
+{
+    auto attrs = toAttrsInternal();
+    if (narHash)
+        attrs.emplace("narHash", narHash->to_string(SRI));
+    attrs.emplace("type", type());
+    return attrs;
+}
+
+std::pair<Tree, std::shared_ptr<const Input>> Input::fetchTree(ref<Store> store) const
+{
+    auto [tree, input] = fetchTreeInternal(store);
+
+    if (tree.actualPath == "")
+        tree.actualPath = store->toRealPath(tree.storePath);
+
+    if (!tree.info.narHash)
+        tree.info.narHash = store->queryPathInfo(tree.storePath)->narHash;
+
+    if (input->narHash)
+        assert(input->narHash == tree.info.narHash);
+
+    if (narHash && narHash != input->narHash)
+        throw Error("NAR hash mismatch in input '%s' (%s), expected '%s', got '%s'",
+            to_string(), tree.actualPath, narHash->to_string(SRI), input->narHash->to_string(SRI));
+
+    return {std::move(tree), input};
+}
+
+}
diff --git a/src/libfetchers/fetchers.hh b/src/libfetchers/fetchers.hh
new file mode 100644
index 000000000..59a58ae67
--- /dev/null
+++ b/src/libfetchers/fetchers.hh
@@ -0,0 +1,103 @@
+#pragma once
+
+#include "types.hh"
+#include "hash.hh"
+#include "path.hh"
+#include "tree-info.hh"
+#include "attrs.hh"
+#include "url.hh"
+
+#include <memory>
+
+namespace nix { class Store; }
+
+namespace nix::fetchers {
+
+struct Input;
+
+struct Tree
+{
+    Path actualPath;
+    StorePath storePath;
+    TreeInfo info;
+};
+
+struct Input : std::enable_shared_from_this<Input>
+{
+    std::optional<Hash> narHash; // FIXME: implement
+
+    virtual std::string type() const = 0;
+
+    virtual ~Input() { }
+
+    virtual bool operator ==(const Input & other) const { return false; }
+
+    /* Check whether this is a "direct" input, that is, not
+       one that goes through a registry. */
+    virtual bool isDirect() const { return true; }
+
+    /* Check whether this is an "immutable" input, that is,
+       one that contains a commit hash or content hash. */
+    virtual bool isImmutable() const { return (bool) narHash; }
+
+    virtual bool contains(const Input & other) const { return false; }
+
+    virtual std::optional<std::string> getRef() const { return {}; }
+
+    virtual std::optional<Hash> getRev() const { return {}; }
+
+    virtual ParsedURL toURL() const = 0;
+
+    std::string to_string() const
+    {
+        return toURL().to_string();
+    }
+
+    Attrs toAttrs() const;
+
+    std::pair<Tree, std::shared_ptr<const Input>> fetchTree(ref<Store> store) const;
+
+private:
+
+    virtual std::pair<Tree, std::shared_ptr<const Input>> fetchTreeInternal(ref<Store> store) const = 0;
+
+    virtual Attrs toAttrsInternal() const = 0;
+};
+
+struct InputScheme
+{
+    virtual ~InputScheme() { }
+
+    virtual std::unique_ptr<Input> inputFromURL(const ParsedURL & url) = 0;
+
+    virtual std::unique_ptr<Input> inputFromAttrs(const Attrs & attrs) = 0;
+};
+
+std::unique_ptr<Input> inputFromURL(const ParsedURL & url);
+
+std::unique_ptr<Input> inputFromURL(const std::string & url);
+
+std::unique_ptr<Input> inputFromAttrs(const Attrs & attrs);
+
+void registerInputScheme(std::unique_ptr<InputScheme> && fetcher);
+
+struct DownloadFileResult
+{
+    StorePath storePath;
+    std::string etag;
+    std::string effectiveUrl;
+};
+
+DownloadFileResult downloadFile(
+    ref<Store> store,
+    const std::string & url,
+    const std::string & name,
+    bool immutable);
+
+Tree downloadTarball(
+    ref<Store> store,
+    const std::string & url,
+    const std::string & name,
+    bool immutable);
+
+}
diff --git a/src/libfetchers/git.cc b/src/libfetchers/git.cc
new file mode 100644
index 000000000..7c18cf67f
--- /dev/null
+++ b/src/libfetchers/git.cc
@@ -0,0 +1,438 @@
+#include "fetchers.hh"
+#include "cache.hh"
+#include "globals.hh"
+#include "tarfile.hh"
+#include "store-api.hh"
+
+#include <sys/time.h>
+
+using namespace std::string_literals;
+
+namespace nix::fetchers {
+
+static std::string readHead(const Path & path)
+{
+    return chomp(runProgram("git", true, { "-C", path, "rev-parse", "--abbrev-ref", "HEAD" }));
+}
+
+static bool isNotDotGitDirectory(const Path & path)
+{
+    static const std::regex gitDirRegex("^(?:.*/)?\\.git$");
+
+    return not std::regex_match(path, gitDirRegex);
+}
+
+struct GitInput : Input
+{
+    ParsedURL url;
+    std::optional<std::string> ref;
+    std::optional<Hash> rev;
+    bool shallow = false;
+    bool submodules = false;
+
+    GitInput(const ParsedURL & url) : url(url)
+    { }
+
+    std::string type() const override { return "git"; }
+
+    bool operator ==(const Input & other) const override
+    {
+        auto other2 = dynamic_cast<const GitInput *>(&other);
+        return
+            other2
+            && url == other2->url
+            && rev == other2->rev
+            && ref == other2->ref;
+    }
+
+    bool isImmutable() const override
+    {
+        return (bool) rev || narHash;
+    }
+
+    std::optional<std::string> getRef() const override { return ref; }
+
+    std::optional<Hash> getRev() const override { return rev; }
+
+    ParsedURL toURL() const override
+    {
+        ParsedURL url2(url);
+        if (url2.scheme != "git") url2.scheme = "git+" + url2.scheme;
+        if (rev) url2.query.insert_or_assign("rev", rev->gitRev());
+        if (ref) url2.query.insert_or_assign("ref", *ref);
+        if (shallow) url2.query.insert_or_assign("shallow", "1");
+        return url2;
+    }
+
+    Attrs toAttrsInternal() const override
+    {
+        Attrs attrs;
+        attrs.emplace("url", url.to_string());
+        if (ref)
+            attrs.emplace("ref", *ref);
+        if (rev)
+            attrs.emplace("rev", rev->gitRev());
+        if (shallow)
+            attrs.emplace("shallow", true);
+        if (submodules)
+            attrs.emplace("submodules", true);
+        return attrs;
+    }
+
+    std::pair<bool, std::string> getActualUrl() const
+    {
+        // Don't clone file:// URIs (but otherwise treat them the
+        // same as remote URIs, i.e. don't use the working tree or
+        // HEAD).
+        static bool forceHttp = getEnv("_NIX_FORCE_HTTP") == "1"; // for testing
+        bool isLocal = url.scheme == "file" && !forceHttp;
+        return {isLocal, isLocal ? url.path : url.base};
+    }
+
+    std::pair<Tree, std::shared_ptr<const Input>> fetchTreeInternal(nix::ref<Store> store) const override
+    {
+        auto name = "source";
+
+        auto input = std::make_shared<GitInput>(*this);
+
+        assert(!rev || rev->type == htSHA1);
+
+        std::string cacheType = "git";
+        if (shallow) cacheType += "-shallow";
+        if (submodules) cacheType += "-submodules";
+
+        auto getImmutableAttrs = [&]()
+        {
+            return Attrs({
+                {"type", cacheType},
+                {"name", name},
+                {"rev", input->rev->gitRev()},
+            });
+        };
+
+        auto makeResult = [&](const Attrs & infoAttrs, StorePath && storePath)
+            -> std::pair<Tree, std::shared_ptr<const Input>>
+        {
+            assert(input->rev);
+            assert(!rev || rev == input->rev);
+            return {
+                Tree {
+                    .actualPath = store->toRealPath(storePath),
+                    .storePath = std::move(storePath),
+                    .info = TreeInfo {
+                        .revCount = shallow ? std::nullopt : std::optional(getIntAttr(infoAttrs, "revCount")),
+                        .lastModified = getIntAttr(infoAttrs, "lastModified"),
+                    },
+                },
+                input
+            };
+        };
+
+        if (rev) {
+            if (auto res = getCache()->lookup(store, getImmutableAttrs()))
+                return makeResult(res->first, std::move(res->second));
+        }
+
+        auto [isLocal, actualUrl_] = getActualUrl();
+        auto actualUrl = actualUrl_; // work around clang bug
+
+        // If this is a local directory and no ref or revision is
+        // given, then allow the use of an unclean working tree.
+        if (!input->ref && !input->rev && isLocal) {
+            bool clean = false;
+
+            /* Check whether this repo has any commits. There are
+               probably better ways to do this. */
+            auto gitDir = actualUrl + "/.git";
+            auto commonGitDir = chomp(runProgram(
+                "git",
+                true,
+                { "-C", actualUrl, "rev-parse", "--git-common-dir" }
+            ));
+            if (commonGitDir != ".git")
+                    gitDir = commonGitDir;
+
+            bool haveCommits = !readDirectory(gitDir + "/refs/heads").empty();
+
+            try {
+                if (haveCommits) {
+                    runProgram("git", true, { "-C", actualUrl, "diff-index", "--quiet", "HEAD", "--" });
+                    clean = true;
+                }
+            } catch (ExecError & e) {
+                if (!WIFEXITED(e.status) || WEXITSTATUS(e.status) != 1) throw;
+            }
+
+            if (!clean) {
+
+                /* This is an unclean working tree. So copy all tracked files. */
+
+                if (!settings.allowDirty)
+                    throw Error("Git tree '%s' is dirty", actualUrl);
+
+                if (settings.warnDirty)
+                    warn("Git tree '%s' is dirty", actualUrl);
+
+                auto gitOpts = Strings({ "-C", actualUrl, "ls-files", "-z" });
+                if (submodules)
+                    gitOpts.emplace_back("--recurse-submodules");
+
+                auto files = tokenizeString<std::set<std::string>>(
+                    runProgram("git", true, gitOpts), "\0"s);
+
+                PathFilter filter = [&](const Path & p) -> bool {
+                    assert(hasPrefix(p, actualUrl));
+                    std::string file(p, actualUrl.size() + 1);
+
+                    auto st = lstat(p);
+
+                    if (S_ISDIR(st.st_mode)) {
+                        auto prefix = file + "/";
+                        auto i = files.lower_bound(prefix);
+                        return i != files.end() && hasPrefix(*i, prefix);
+                    }
+
+                    return files.count(file);
+                };
+
+                auto storePath = store->addToStore("source", actualUrl, true, htSHA256, filter);
+
+                auto tree = Tree {
+                    .actualPath = store->printStorePath(storePath),
+                    .storePath = std::move(storePath),
+                    .info = TreeInfo {
+                        // FIXME: maybe we should use the timestamp of the last
+                        // modified dirty file?
+                        .lastModified = haveCommits ? std::stoull(runProgram("git", true, { "-C", actualUrl, "log", "-1", "--format=%ct", "HEAD" })) : 0,
+                    }
+                };
+
+                return {std::move(tree), input};
+            }
+        }
+
+        if (!input->ref) input->ref = isLocal ? readHead(actualUrl) : "master";
+
+        Attrs mutableAttrs({
+            {"type", cacheType},
+            {"name", name},
+            {"url", actualUrl},
+            {"ref", *input->ref},
+        });
+
+        Path repoDir;
+
+        if (isLocal) {
+
+            if (!input->rev)
+                input->rev = Hash(chomp(runProgram("git", true, { "-C", actualUrl, "rev-parse", *input->ref })), htSHA1);
+
+            repoDir = actualUrl;
+
+        } else {
+
+            if (auto res = getCache()->lookup(store, mutableAttrs)) {
+                auto rev2 = Hash(getStrAttr(res->first, "rev"), htSHA1);
+                if (!rev || rev == rev2) {
+                    input->rev = rev2;
+                    return makeResult(res->first, std::move(res->second));
+                }
+            }
+
+            Path cacheDir = getCacheDir() + "/nix/gitv3/" + hashString(htSHA256, actualUrl).to_string(Base32, false);
+            repoDir = cacheDir;
+
+            if (!pathExists(cacheDir)) {
+                createDirs(dirOf(cacheDir));
+                runProgram("git", true, { "init", "--bare", repoDir });
+            }
+
+            Path localRefFile =
+                input->ref->compare(0, 5, "refs/") == 0
+                ? cacheDir + "/" + *input->ref
+                : cacheDir + "/refs/heads/" + *input->ref;
+
+            bool doFetch;
+            time_t now = time(0);
+
+            /* If a rev was specified, we need to fetch if it's not in the
+               repo. */
+            if (input->rev) {
+                try {
+                    runProgram("git", true, { "-C", repoDir, "cat-file", "-e", input->rev->gitRev() });
+                    doFetch = false;
+                } catch (ExecError & e) {
+                    if (WIFEXITED(e.status)) {
+                        doFetch = true;
+                    } else {
+                        throw;
+                    }
+                }
+            } else {
+                /* If the local ref is older than ‘tarball-ttl’ seconds, do a
+                   git fetch to update the local ref to the remote ref. */
+                struct stat st;
+                doFetch = stat(localRefFile.c_str(), &st) != 0 ||
+                    (uint64_t) st.st_mtime + settings.tarballTtl <= (uint64_t) now;
+            }
+
+            if (doFetch) {
+                Activity act(*logger, lvlTalkative, actUnknown, fmt("fetching Git repository '%s'", actualUrl));
+
+                // FIXME: git stderr messes up our progress indicator, so
+                // we're using --quiet for now. Should process its stderr.
+                try {
+                    runProgram("git", true, { "-C", repoDir, "fetch", "--quiet", "--force", "--", actualUrl, fmt("%s:%s", *input->ref, *input->ref) });
+                } catch (Error & e) {
+                    if (!pathExists(localRefFile)) throw;
+                    warn("could not update local clone of Git repository '%s'; continuing with the most recent version", actualUrl);
+                }
+
+                struct timeval times[2];
+                times[0].tv_sec = now;
+                times[0].tv_usec = 0;
+                times[1].tv_sec = now;
+                times[1].tv_usec = 0;
+
+                utimes(localRefFile.c_str(), times);
+            }
+
+            if (!input->rev)
+                input->rev = Hash(chomp(readFile(localRefFile)), htSHA1);
+        }
+
+        bool isShallow = chomp(runProgram("git", true, { "-C", repoDir, "rev-parse", "--is-shallow-repository" })) == "true";
+
+        if (isShallow && !shallow)
+            throw Error("'%s' is a shallow Git repository, but a non-shallow repository is needed", actualUrl);
+
+        // FIXME: check whether rev is an ancestor of ref.
+
+        printTalkative("using revision %s of repo '%s'", input->rev->gitRev(), actualUrl);
+
+        /* Now that we know the ref, check again whether we have it in
+           the store. */
+        if (auto res = getCache()->lookup(store, getImmutableAttrs()))
+            return makeResult(res->first, std::move(res->second));
+
+        Path tmpDir = createTempDir();
+        AutoDelete delTmpDir(tmpDir, true);
+        PathFilter filter = defaultPathFilter;
+
+        if (submodules) {
+            Path tmpGitDir = createTempDir();
+            AutoDelete delTmpGitDir(tmpGitDir, true);
+
+            runProgram("git", true, { "init", tmpDir, "--separate-git-dir", tmpGitDir });
+            // TODO: repoDir might lack the ref (it only checks if rev
+            // exists, see FIXME above) so use a big hammer and fetch
+            // everything to ensure we get the rev.
+            runProgram("git", true, { "-C", tmpDir, "fetch", "--quiet", "--force",
+                                      "--update-head-ok", "--", repoDir, "refs/*:refs/*" });
+
+            runProgram("git", true, { "-C", tmpDir, "checkout", "--quiet", input->rev->gitRev() });
+            runProgram("git", true, { "-C", tmpDir, "remote", "add", "origin", actualUrl });
+            runProgram("git", true, { "-C", tmpDir, "submodule", "--quiet", "update", "--init", "--recursive" });
+
+            filter = isNotDotGitDirectory;
+        } else {
+            // FIXME: should pipe this, or find some better way to extract a
+            // revision.
+            auto source = sinkToSource([&](Sink & sink) {
+                RunOptions gitOptions("git", { "-C", repoDir, "archive", input->rev->gitRev() });
+                gitOptions.standardOut = &sink;
+                runProgram2(gitOptions);
+            });
+
+            unpackTarfile(*source, tmpDir);
+        }
+
+        auto storePath = store->addToStore(name, tmpDir, true, htSHA256, filter);
+
+        auto lastModified = std::stoull(runProgram("git", true, { "-C", repoDir, "log", "-1", "--format=%ct", input->rev->gitRev() }));
+
+        Attrs infoAttrs({
+            {"rev", input->rev->gitRev()},
+            {"lastModified", lastModified},
+        });
+
+        if (!shallow)
+            infoAttrs.insert_or_assign("revCount",
+                std::stoull(runProgram("git", true, { "-C", repoDir, "rev-list", "--count", input->rev->gitRev() })));
+
+        if (!this->rev)
+            getCache()->add(
+                store,
+                mutableAttrs,
+                infoAttrs,
+                storePath,
+                false);
+
+        getCache()->add(
+            store,
+            getImmutableAttrs(),
+            infoAttrs,
+            storePath,
+            true);
+
+        return makeResult(infoAttrs, std::move(storePath));
+    }
+};
+
+struct GitInputScheme : InputScheme
+{
+    std::unique_ptr<Input> inputFromURL(const ParsedURL & url) override
+    {
+        if (url.scheme != "git" &&
+            url.scheme != "git+http" &&
+            url.scheme != "git+https" &&
+            url.scheme != "git+ssh" &&
+            url.scheme != "git+file") return nullptr;
+
+        auto url2(url);
+        if (hasPrefix(url2.scheme, "git+")) url2.scheme = std::string(url2.scheme, 4);
+        url2.query.clear();
+
+        Attrs attrs;
+        attrs.emplace("type", "git");
+
+        for (auto &[name, value] : url.query) {
+            if (name == "rev" || name == "ref")
+                attrs.emplace(name, value);
+            else
+                url2.query.emplace(name, value);
+        }
+
+        attrs.emplace("url", url2.to_string());
+
+        return inputFromAttrs(attrs);
+    }
+
+    std::unique_ptr<Input> inputFromAttrs(const Attrs & attrs) override
+    {
+        if (maybeGetStrAttr(attrs, "type") != "git") return {};
+
+        for (auto & [name, value] : attrs)
+            if (name != "type" && name != "url" && name != "ref" && name != "rev" && name != "shallow" && name != "submodules")
+                throw Error("unsupported Git input attribute '%s'", name);
+
+        auto input = std::make_unique<GitInput>(parseURL(getStrAttr(attrs, "url")));
+        if (auto ref = maybeGetStrAttr(attrs, "ref")) {
+            if (!std::regex_match(*ref, refRegex))
+                throw BadURL("invalid Git branch/tag name '%s'", *ref);
+            input->ref = *ref;
+        }
+        if (auto rev = maybeGetStrAttr(attrs, "rev"))
+            input->rev = Hash(*rev, htSHA1);
+
+        input->shallow = maybeGetBoolAttr(attrs, "shallow").value_or(false);
+
+        input->submodules = maybeGetBoolAttr(attrs, "submodules").value_or(false);
+
+        return input;
+    }
+};
+
+static auto r1 = OnStartup([] { registerInputScheme(std::make_unique<GitInputScheme>()); });
+
+}
diff --git a/src/libfetchers/github.cc b/src/libfetchers/github.cc
new file mode 100644
index 000000000..8675a5a66
--- /dev/null
+++ b/src/libfetchers/github.cc
@@ -0,0 +1,195 @@
+#include "filetransfer.hh"
+#include "cache.hh"
+#include "fetchers.hh"
+#include "globals.hh"
+#include "store-api.hh"
+
+#include <nlohmann/json.hpp>
+
+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);
+
+struct GitHubInput : Input
+{
+    std::string owner;
+    std::string repo;
+    std::optional<std::string> ref;
+    std::optional<Hash> rev;
+
+    std::string type() const override { return "github"; }
+
+    bool operator ==(const Input & other) const override
+    {
+        auto other2 = dynamic_cast<const GitHubInput *>(&other);
+        return
+            other2
+            && owner == other2->owner
+            && repo == other2->repo
+            && rev == other2->rev
+            && ref == other2->ref;
+    }
+
+    bool isImmutable() const override
+    {
+        return (bool) rev || narHash;
+    }
+
+    std::optional<std::string> getRef() const override { return ref; }
+
+    std::optional<Hash> getRev() const override { return rev; }
+
+    ParsedURL toURL() const override
+    {
+        auto path = owner + "/" + repo;
+        assert(!(ref && rev));
+        if (ref) path += "/" + *ref;
+        if (rev) path += "/" + rev->to_string(Base16, false);
+        return ParsedURL {
+            .scheme = "github",
+            .path = path,
+        };
+    }
+
+    Attrs toAttrsInternal() const 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;
+    }
+
+    std::pair<Tree, std::shared_ptr<const Input>> fetchTreeInternal(nix::ref<Store> store) const 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(json["sha"], htSHA1);
+            debug("HEAD revision for '%s' is %s", url, rev->gitRev());
+        }
+
+        auto input = std::make_shared<GitHubInput>(*this);
+        input->ref = {};
+        input->rev = *rev;
+
+        Attrs immutableAttrs({
+            {"type", "git-tarball"},
+            {"rev", rev->gitRev()},
+        });
+
+        if (auto res = getCache()->lookup(store, immutableAttrs)) {
+            return {
+                Tree{
+                    .actualPath = store->toRealPath(res->second),
+                    .storePath = std::move(res->second),
+                    .info = TreeInfo {
+                        .lastModified = getIntAttr(res->first, "lastModified"),
+                    },
+                },
+                input
+            };
+        }
+
+        // FIXME: use regular /archive URLs instead? api.github.com
+        // might have stricter rate limits.
+
+        auto url = fmt("https://api.github.com/repos/%s/%s/tarball/%s",
+            owner, repo, rev->to_string(Base16, false));
+
+        std::string accessToken = settings.githubAccessToken.get();
+        if (accessToken != "")
+            url += "?access_token=" + accessToken;
+
+        auto tree = downloadTarball(store, url, "source", true);
+
+        getCache()->add(
+            store,
+            immutableAttrs,
+            {
+                {"rev", rev->gitRev()},
+                {"lastModified", *tree.info.lastModified}
+            },
+            tree.storePath,
+            true);
+
+        return {std::move(tree), input};
+    }
+};
+
+struct GitHubInputScheme : InputScheme
+{
+    std::unique_ptr<Input> inputFromURL(const ParsedURL & url) override
+    {
+        if (url.scheme != "github") return nullptr;
+
+        auto path = tokenizeString<std::vector<std::string>>(url.path, "/");
+        auto input = std::make_unique<GitHubInput>();
+
+        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);
+
+        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;
+            }
+        }
+
+        if (input->ref && input->rev)
+            throw BadURL("GitHub URL '%s' contains both a commit hash and a branch/tag name", url.url);
+
+        input->owner = path[0];
+        input->repo = path[1];
+
+        return input;
+    }
+
+    std::unique_ptr<Input> inputFromAttrs(const Attrs & attrs) override
+    {
+        if (maybeGetStrAttr(attrs, "type") != "github") return {};
+
+        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;
+    }
+};
+
+static auto r1 = OnStartup([] { registerInputScheme(std::make_unique<GitHubInputScheme>()); });
+
+}
diff --git a/src/libfetchers/local.mk b/src/libfetchers/local.mk
new file mode 100644
index 000000000..d7143d8a6
--- /dev/null
+++ b/src/libfetchers/local.mk
@@ -0,0 +1,11 @@
+libraries += libfetchers
+
+libfetchers_NAME = libnixfetchers
+
+libfetchers_DIR := $(d)
+
+libfetchers_SOURCES := $(wildcard $(d)/*.cc)
+
+libfetchers_CXXFLAGS += -I src/libutil -I src/libstore
+
+libfetchers_LIBS = libutil libstore libnixrust
diff --git a/src/libfetchers/mercurial.cc b/src/libfetchers/mercurial.cc
new file mode 100644
index 000000000..1d6571571
--- /dev/null
+++ b/src/libfetchers/mercurial.cc
@@ -0,0 +1,303 @@
+#include "fetchers.hh"
+#include "cache.hh"
+#include "globals.hh"
+#include "tarfile.hh"
+#include "store-api.hh"
+
+#include <sys/time.h>
+
+using namespace std::string_literals;
+
+namespace nix::fetchers {
+
+struct MercurialInput : Input
+{
+    ParsedURL url;
+    std::optional<std::string> ref;
+    std::optional<Hash> rev;
+
+    MercurialInput(const ParsedURL & url) : url(url)
+    { }
+
+    std::string type() const override { return "hg"; }
+
+    bool operator ==(const Input & other) const override
+    {
+        auto other2 = dynamic_cast<const MercurialInput *>(&other);
+        return
+            other2
+            && url == other2->url
+            && rev == other2->rev
+            && ref == other2->ref;
+    }
+
+    bool isImmutable() const override
+    {
+        return (bool) rev || narHash;
+    }
+
+    std::optional<std::string> getRef() const override { return ref; }
+
+    std::optional<Hash> getRev() const override { return rev; }
+
+    ParsedURL toURL() const override
+    {
+        ParsedURL url2(url);
+        url2.scheme = "hg+" + url2.scheme;
+        if (rev) url2.query.insert_or_assign("rev", rev->gitRev());
+        if (ref) url2.query.insert_or_assign("ref", *ref);
+        return url;
+    }
+
+    Attrs toAttrsInternal() const override
+    {
+        Attrs attrs;
+        attrs.emplace("url", url.to_string());
+        if (ref)
+            attrs.emplace("ref", *ref);
+        if (rev)
+            attrs.emplace("rev", rev->gitRev());
+        return attrs;
+    }
+
+    std::pair<bool, std::string> getActualUrl() const
+    {
+        bool isLocal = url.scheme == "file";
+        return {isLocal, isLocal ? url.path : url.base};
+    }
+
+    std::pair<Tree, std::shared_ptr<const Input>> fetchTreeInternal(nix::ref<Store> store) const override
+    {
+        auto name = "source";
+
+        auto input = std::make_shared<MercurialInput>(*this);
+
+        auto [isLocal, actualUrl_] = getActualUrl();
+        auto actualUrl = actualUrl_; // work around clang bug
+
+        // FIXME: return lastModified.
+
+        // FIXME: don't clone local repositories.
+
+        if (!input->ref && !input->rev && isLocal && pathExists(actualUrl + "/.hg")) {
+
+            bool clean = runProgram("hg", true, { "status", "-R", actualUrl, "--modified", "--added", "--removed" }) == "";
+
+            if (!clean) {
+
+                /* This is an unclean working tree. So copy all tracked
+                   files. */
+
+                if (!settings.allowDirty)
+                    throw Error("Mercurial tree '%s' is unclean", actualUrl);
+
+                if (settings.warnDirty)
+                    warn("Mercurial tree '%s' is unclean", actualUrl);
+
+                input->ref = chomp(runProgram("hg", true, { "branch", "-R", actualUrl }));
+
+                auto files = tokenizeString<std::set<std::string>>(
+                    runProgram("hg", true, { "status", "-R", actualUrl, "--clean", "--modified", "--added", "--no-status", "--print0" }), "\0"s);
+
+                PathFilter filter = [&](const Path & p) -> bool {
+                    assert(hasPrefix(p, actualUrl));
+                    std::string file(p, actualUrl.size() + 1);
+
+                    auto st = lstat(p);
+
+                    if (S_ISDIR(st.st_mode)) {
+                        auto prefix = file + "/";
+                        auto i = files.lower_bound(prefix);
+                        return i != files.end() && hasPrefix(*i, prefix);
+                    }
+
+                    return files.count(file);
+                };
+
+                auto storePath = store->addToStore("source", actualUrl, true, htSHA256, filter);
+
+                return {Tree {
+                    .actualPath = store->printStorePath(storePath),
+                    .storePath = std::move(storePath),
+                }, input};
+            }
+        }
+
+        if (!input->ref) input->ref = "default";
+
+        auto getImmutableAttrs = [&]()
+        {
+            return Attrs({
+                {"type", "hg"},
+                {"name", name},
+                {"rev", input->rev->gitRev()},
+            });
+        };
+
+        auto makeResult = [&](const Attrs & infoAttrs, StorePath && storePath)
+            -> std::pair<Tree, std::shared_ptr<const Input>>
+        {
+            assert(input->rev);
+            assert(!rev || rev == input->rev);
+            return {
+                Tree{
+                    .actualPath = store->toRealPath(storePath),
+                    .storePath = std::move(storePath),
+                    .info = TreeInfo {
+                        .revCount = getIntAttr(infoAttrs, "revCount"),
+                    },
+                },
+                input
+            };
+        };
+
+        if (input->rev) {
+            if (auto res = getCache()->lookup(store, getImmutableAttrs()))
+                return makeResult(res->first, std::move(res->second));
+        }
+
+        assert(input->rev || input->ref);
+        auto revOrRef = input->rev ? input->rev->gitRev() : *input->ref;
+
+        Attrs mutableAttrs({
+            {"type", "hg"},
+            {"name", name},
+            {"url", actualUrl},
+            {"ref", *input->ref},
+        });
+
+        if (auto res = getCache()->lookup(store, mutableAttrs)) {
+            auto rev2 = Hash(getStrAttr(res->first, "rev"), htSHA1);
+            if (!rev || rev == rev2) {
+                input->rev = rev2;
+                return makeResult(res->first, std::move(res->second));
+            }
+        }
+
+        Path cacheDir = fmt("%s/nix/hg/%s", getCacheDir(), hashString(htSHA256, actualUrl).to_string(Base32, false));
+
+        /* If this is a commit hash that we already have, we don't
+           have to pull again. */
+        if (!(input->rev
+                && pathExists(cacheDir)
+                && runProgram(
+                    RunOptions("hg", { "log", "-R", cacheDir, "-r", input->rev->gitRev(), "--template", "1" })
+                    .killStderr(true)).second == "1"))
+        {
+            Activity act(*logger, lvlTalkative, actUnknown, fmt("fetching Mercurial repository '%s'", actualUrl));
+
+            if (pathExists(cacheDir)) {
+                try {
+                    runProgram("hg", true, { "pull", "-R", cacheDir, "--", actualUrl });
+                }
+                catch (ExecError & e) {
+                    string transJournal = cacheDir + "/.hg/store/journal";
+                    /* hg throws "abandoned transaction" error only if this file exists */
+                    if (pathExists(transJournal)) {
+                        runProgram("hg", true, { "recover", "-R", cacheDir });
+                        runProgram("hg", true, { "pull", "-R", cacheDir, "--", actualUrl });
+                    } else {
+                        throw ExecError(e.status, fmt("'hg pull' %s", statusToString(e.status)));
+                    }
+                }
+            } else {
+                createDirs(dirOf(cacheDir));
+                runProgram("hg", true, { "clone", "--noupdate", "--", actualUrl, cacheDir });
+            }
+        }
+
+        auto tokens = tokenizeString<std::vector<std::string>>(
+            runProgram("hg", true, { "log", "-R", cacheDir, "-r", revOrRef, "--template", "{node} {rev} {branch}" }));
+        assert(tokens.size() == 3);
+
+        input->rev = Hash(tokens[0], htSHA1);
+        auto revCount = std::stoull(tokens[1]);
+        input->ref = tokens[2];
+
+        if (auto res = getCache()->lookup(store, getImmutableAttrs()))
+            return makeResult(res->first, std::move(res->second));
+
+        Path tmpDir = createTempDir();
+        AutoDelete delTmpDir(tmpDir, true);
+
+        runProgram("hg", true, { "archive", "-R", cacheDir, "-r", input->rev->gitRev(), tmpDir });
+
+        deletePath(tmpDir + "/.hg_archival.txt");
+
+        auto storePath = store->addToStore(name, tmpDir);
+
+        Attrs infoAttrs({
+            {"rev", input->rev->gitRev()},
+            {"revCount", (int64_t) revCount},
+        });
+
+        if (!this->rev)
+            getCache()->add(
+                store,
+                mutableAttrs,
+                infoAttrs,
+                storePath,
+                false);
+
+        getCache()->add(
+            store,
+            getImmutableAttrs(),
+            infoAttrs,
+            storePath,
+            true);
+
+        return makeResult(infoAttrs, std::move(storePath));
+    }
+};
+
+struct MercurialInputScheme : InputScheme
+{
+    std::unique_ptr<Input> inputFromURL(const ParsedURL & url) override
+    {
+        if (url.scheme != "hg+http" &&
+            url.scheme != "hg+https" &&
+            url.scheme != "hg+ssh" &&
+            url.scheme != "hg+file") return nullptr;
+
+        auto url2(url);
+        url2.scheme = std::string(url2.scheme, 3);
+        url2.query.clear();
+
+        Attrs attrs;
+        attrs.emplace("type", "hg");
+
+        for (auto &[name, value] : url.query) {
+            if (name == "rev" || name == "ref")
+                attrs.emplace(name, value);
+            else
+                url2.query.emplace(name, value);
+        }
+
+        attrs.emplace("url", url2.to_string());
+
+        return inputFromAttrs(attrs);
+    }
+
+    std::unique_ptr<Input> inputFromAttrs(const Attrs & attrs) override
+    {
+        if (maybeGetStrAttr(attrs, "type") != "hg") return {};
+
+        for (auto & [name, value] : attrs)
+            if (name != "type" && name != "url" && name != "ref" && name != "rev")
+                throw Error("unsupported Mercurial input attribute '%s'", name);
+
+        auto input = std::make_unique<MercurialInput>(parseURL(getStrAttr(attrs, "url")));
+        if (auto ref = maybeGetStrAttr(attrs, "ref")) {
+            if (!std::regex_match(*ref, refRegex))
+                throw BadURL("invalid Mercurial branch/tag name '%s'", *ref);
+            input->ref = *ref;
+        }
+        if (auto rev = maybeGetStrAttr(attrs, "rev"))
+            input->rev = Hash(*rev, htSHA1);
+        return input;
+    }
+};
+
+static auto r1 = OnStartup([] { registerInputScheme(std::make_unique<MercurialInputScheme>()); });
+
+}
diff --git a/src/libfetchers/path.cc b/src/libfetchers/path.cc
new file mode 100644
index 000000000..ba2cc192e
--- /dev/null
+++ b/src/libfetchers/path.cc
@@ -0,0 +1,148 @@
+#include "fetchers.hh"
+#include "store-api.hh"
+
+namespace nix::fetchers {
+
+struct PathInput : Input
+{
+    Path path;
+
+    /* Allow the user to pass in "fake" tree info attributes. This is
+       useful for making a pinned tree work the same as the repository
+       from which is exported
+       (e.g. path:/nix/store/...-source?lastModified=1585388205&rev=b0c285...). */
+    std::optional<Hash> rev;
+    std::optional<uint64_t> revCount;
+    std::optional<time_t> lastModified;
+
+    std::string type() const override { return "path"; }
+
+    std::optional<Hash> getRev() const override { return rev; }
+
+    bool operator ==(const Input & other) const override
+    {
+        auto other2 = dynamic_cast<const PathInput *>(&other);
+        return
+            other2
+            && path == other2->path
+            && rev == other2->rev
+            && revCount == other2->revCount
+            && lastModified == other2->lastModified;
+    }
+
+    bool isImmutable() const override
+    {
+        return (bool) narHash;
+    }
+
+    ParsedURL toURL() const override
+    {
+        auto query = attrsToQuery(toAttrsInternal());
+        query.erase("path");
+        return ParsedURL {
+            .scheme = "path",
+            .path = path,
+            .query = query,
+        };
+    }
+
+    Attrs toAttrsInternal() const override
+    {
+        Attrs attrs;
+        attrs.emplace("path", path);
+        if (rev)
+            attrs.emplace("rev", rev->gitRev());
+        if (revCount)
+            attrs.emplace("revCount", *revCount);
+        if (lastModified)
+            attrs.emplace("lastModified", *lastModified);
+        return attrs;
+    }
+
+    std::pair<Tree, std::shared_ptr<const Input>> fetchTreeInternal(nix::ref<Store> store) const override
+    {
+        auto input = std::make_shared<PathInput>(*this);
+
+        // FIXME: check whether access to 'path' is allowed.
+
+        auto storePath = store->maybeParseStorePath(path);
+
+        if (storePath)
+            store->addTempRoot(*storePath);
+
+        if (!storePath || storePath->name() != "source" || !store->isValidPath(*storePath))
+            // FIXME: try to substitute storePath.
+            storePath = store->addToStore("source", path);
+
+        return
+            {
+                Tree {
+                    .actualPath = store->toRealPath(*storePath),
+                    .storePath = std::move(*storePath),
+                    .info = TreeInfo {
+                        .revCount = revCount,
+                        .lastModified = lastModified
+                    }
+                },
+                input
+            };
+    }
+
+};
+
+struct PathInputScheme : InputScheme
+{
+    std::unique_ptr<Input> inputFromURL(const ParsedURL & url) override
+    {
+        if (url.scheme != "path") return nullptr;
+
+        auto input = std::make_unique<PathInput>();
+        input->path = url.path;
+
+        for (auto & [name, value] : url.query)
+            if (name == "rev")
+                input->rev = Hash(value, htSHA1);
+            else if (name == "revCount") {
+                uint64_t revCount;
+                if (!string2Int(value, revCount))
+                    throw Error("path URL '%s' has invalid parameter '%s'", url.to_string(), name);
+                input->revCount = revCount;
+            }
+            else if (name == "lastModified") {
+                time_t lastModified;
+                if (!string2Int(value, lastModified))
+                    throw Error("path URL '%s' has invalid parameter '%s'", url.to_string(), name);
+                input->lastModified = lastModified;
+            }
+            else
+                throw Error("path URL '%s' has unsupported parameter '%s'", url.to_string(), name);
+
+        return input;
+    }
+
+    std::unique_ptr<Input> inputFromAttrs(const Attrs & attrs) override
+    {
+        if (maybeGetStrAttr(attrs, "type") != "path") return {};
+
+        auto input = std::make_unique<PathInput>();
+        input->path = getStrAttr(attrs, "path");
+
+        for (auto & [name, value] : attrs)
+            if (name == "rev")
+                input->rev = Hash(getStrAttr(attrs, "rev"), htSHA1);
+            else if (name == "revCount")
+                input->revCount = getIntAttr(attrs, "revCount");
+            else if (name == "lastModified")
+                input->lastModified = getIntAttr(attrs, "lastModified");
+            else if (name == "type" || name == "path")
+                ;
+            else
+                throw Error("unsupported path input attribute '%s'", name);
+
+        return input;
+    }
+};
+
+static auto r1 = OnStartup([] { registerInputScheme(std::make_unique<PathInputScheme>()); });
+
+}
diff --git a/src/libfetchers/tarball.cc b/src/libfetchers/tarball.cc
new file mode 100644
index 000000000..695525b31
--- /dev/null
+++ b/src/libfetchers/tarball.cc
@@ -0,0 +1,275 @@
+#include "fetchers.hh"
+#include "cache.hh"
+#include "filetransfer.hh"
+#include "globals.hh"
+#include "store-api.hh"
+#include "archive.hh"
+#include "tarfile.hh"
+
+namespace nix::fetchers {
+
+DownloadFileResult downloadFile(
+    ref<Store> store,
+    const std::string & url,
+    const std::string & name,
+    bool immutable)
+{
+    // FIXME: check store
+
+    Attrs inAttrs({
+        {"type", "file"},
+        {"url", url},
+        {"name", name},
+    });
+
+    auto cached = getCache()->lookupExpired(store, inAttrs);
+
+    auto useCached = [&]() -> DownloadFileResult
+    {
+        return {
+            .storePath = std::move(cached->storePath),
+            .etag = getStrAttr(cached->infoAttrs, "etag"),
+            .effectiveUrl = getStrAttr(cached->infoAttrs, "url")
+        };
+    };
+
+    if (cached && !cached->expired)
+        return useCached();
+
+    FileTransferRequest request(url);
+    if (cached)
+        request.expectedETag = getStrAttr(cached->infoAttrs, "etag");
+    FileTransferResult res;
+    try {
+        res = getFileTransfer()->download(request);
+    } catch (FileTransferError & e) {
+        if (cached) {
+            warn("%s; using cached version", e.msg());
+            return useCached();
+        } else
+            throw;
+    }
+
+    // FIXME: write to temporary file.
+
+    Attrs infoAttrs({
+        {"etag", res.etag},
+        {"url", res.effectiveUri},
+    });
+
+    std::optional<StorePath> storePath;
+
+    if (res.cached) {
+        assert(cached);
+        assert(request.expectedETag == res.etag);
+        storePath = std::move(cached->storePath);
+    } else {
+        StringSink sink;
+        dumpString(*res.data, sink);
+        auto hash = hashString(htSHA256, *res.data);
+        ValidPathInfo info(store->makeFixedOutputPath(false, hash, name));
+        info.narHash = hashString(htSHA256, *sink.s);
+        info.narSize = sink.s->size();
+        info.ca = makeFixedOutputCA(false, hash);
+        store->addToStore(info, sink.s, NoRepair, NoCheckSigs);
+        storePath = std::move(info.path);
+    }
+
+    getCache()->add(
+        store,
+        inAttrs,
+        infoAttrs,
+        *storePath,
+        immutable);
+
+    if (url != res.effectiveUri)
+        getCache()->add(
+            store,
+            {
+                {"type", "file"},
+                {"url", res.effectiveUri},
+                {"name", name},
+            },
+            infoAttrs,
+            *storePath,
+            immutable);
+
+    return {
+        .storePath = std::move(*storePath),
+        .etag = res.etag,
+        .effectiveUrl = res.effectiveUri,
+    };
+}
+
+Tree downloadTarball(
+    ref<Store> store,
+    const std::string & url,
+    const std::string & name,
+    bool immutable)
+{
+    Attrs inAttrs({
+        {"type", "tarball"},
+        {"url", url},
+        {"name", name},
+    });
+
+    auto cached = getCache()->lookupExpired(store, inAttrs);
+
+    if (cached && !cached->expired)
+        return Tree {
+            .actualPath = store->toRealPath(cached->storePath),
+            .storePath = std::move(cached->storePath),
+            .info = TreeInfo {
+                .lastModified = getIntAttr(cached->infoAttrs, "lastModified"),
+            },
+        };
+
+    auto res = downloadFile(store, url, name, immutable);
+
+    std::optional<StorePath> unpackedStorePath;
+    time_t lastModified;
+
+    if (cached && res.etag != "" && getStrAttr(cached->infoAttrs, "etag") == res.etag) {
+        unpackedStorePath = std::move(cached->storePath);
+        lastModified = getIntAttr(cached->infoAttrs, "lastModified");
+    } else {
+        Path tmpDir = createTempDir();
+        AutoDelete autoDelete(tmpDir, true);
+        unpackTarfile(store->toRealPath(res.storePath), tmpDir);
+        auto members = readDirectory(tmpDir);
+        if (members.size() != 1)
+            throw nix::Error("tarball '%s' contains an unexpected number of top-level files", url);
+        auto topDir = tmpDir + "/" + members.begin()->name;
+        lastModified = lstat(topDir).st_mtime;
+        unpackedStorePath = store->addToStore(name, topDir, true, htSHA256, defaultPathFilter, NoRepair);
+    }
+
+    Attrs infoAttrs({
+        {"lastModified", lastModified},
+        {"etag", res.etag},
+    });
+
+    getCache()->add(
+        store,
+        inAttrs,
+        infoAttrs,
+        *unpackedStorePath,
+        immutable);
+
+    return Tree {
+        .actualPath = store->toRealPath(*unpackedStorePath),
+        .storePath = std::move(*unpackedStorePath),
+        .info = TreeInfo {
+            .lastModified = lastModified,
+        },
+    };
+}
+
+struct TarballInput : Input
+{
+    ParsedURL url;
+    std::optional<Hash> hash;
+
+    TarballInput(const ParsedURL & url) : url(url)
+    { }
+
+    std::string type() const override { return "tarball"; }
+
+    bool operator ==(const Input & other) const override
+    {
+        auto other2 = dynamic_cast<const TarballInput *>(&other);
+        return
+            other2
+            && to_string() == other2->to_string()
+            && hash == other2->hash;
+    }
+
+    bool isImmutable() const override
+    {
+        return hash || narHash;
+    }
+
+    ParsedURL toURL() const override
+    {
+        auto url2(url);
+        // NAR hashes are preferred over file hashes since tar/zip files
+        // don't have a canonical representation.
+        if (narHash)
+            url2.query.insert_or_assign("narHash", narHash->to_string(SRI));
+        else if (hash)
+            url2.query.insert_or_assign("hash", hash->to_string(SRI));
+        return url2;
+    }
+
+    Attrs toAttrsInternal() const override
+    {
+        Attrs attrs;
+        attrs.emplace("url", url.to_string());
+        if (hash)
+            attrs.emplace("hash", hash->to_string(SRI));
+        return attrs;
+    }
+
+    std::pair<Tree, std::shared_ptr<const Input>> fetchTreeInternal(nix::ref<Store> store) const override
+    {
+        auto tree = downloadTarball(store, url.to_string(), "source", false);
+
+        auto input = std::make_shared<TarballInput>(*this);
+        input->narHash = store->queryPathInfo(tree.storePath)->narHash;
+
+        return {std::move(tree), input};
+    }
+};
+
+struct TarballInputScheme : InputScheme
+{
+    std::unique_ptr<Input> inputFromURL(const ParsedURL & url) override
+    {
+        if (url.scheme != "file" && url.scheme != "http" && url.scheme != "https") return nullptr;
+
+        if (!hasSuffix(url.path, ".zip")
+            && !hasSuffix(url.path, ".tar")
+            && !hasSuffix(url.path, ".tar.gz")
+            && !hasSuffix(url.path, ".tar.xz")
+            && !hasSuffix(url.path, ".tar.bz2"))
+            return nullptr;
+
+        auto input = std::make_unique<TarballInput>(url);
+
+        auto hash = input->url.query.find("hash");
+        if (hash != input->url.query.end()) {
+            // FIXME: require SRI hash.
+            input->hash = Hash(hash->second);
+            input->url.query.erase(hash);
+        }
+
+        auto narHash = input->url.query.find("narHash");
+        if (narHash != input->url.query.end()) {
+            // FIXME: require SRI hash.
+            input->narHash = Hash(narHash->second);
+            input->url.query.erase(narHash);
+        }
+
+        return input;
+    }
+
+    std::unique_ptr<Input> inputFromAttrs(const Attrs & attrs) override
+    {
+        if (maybeGetStrAttr(attrs, "type") != "tarball") return {};
+
+        for (auto & [name, value] : attrs)
+            if (name != "type" && name != "url" && name != "hash")
+                throw Error("unsupported tarball input attribute '%s'", name);
+
+        auto input = std::make_unique<TarballInput>(parseURL(getStrAttr(attrs, "url")));
+        if (auto hash = maybeGetStrAttr(attrs, "hash"))
+            // FIXME: require SRI hash.
+            input->hash = Hash(*hash);
+
+        return input;
+    }
+};
+
+static auto r1 = OnStartup([] { registerInputScheme(std::make_unique<TarballInputScheme>()); });
+
+}
diff --git a/src/libfetchers/tree-info.cc b/src/libfetchers/tree-info.cc
new file mode 100644
index 000000000..5788e94a1
--- /dev/null
+++ b/src/libfetchers/tree-info.cc
@@ -0,0 +1,14 @@
+#include "tree-info.hh"
+#include "store-api.hh"
+
+#include <nlohmann/json.hpp>
+
+namespace nix::fetchers {
+
+StorePath TreeInfo::computeStorePath(Store & store) const
+{
+    assert(narHash);
+    return store.makeFixedOutputPath(true, narHash, "source");
+}
+
+}
diff --git a/src/libfetchers/tree-info.hh b/src/libfetchers/tree-info.hh
new file mode 100644
index 000000000..2c7347281
--- /dev/null
+++ b/src/libfetchers/tree-info.hh
@@ -0,0 +1,29 @@
+#pragma once
+
+#include "path.hh"
+#include "hash.hh"
+
+#include <nlohmann/json_fwd.hpp>
+
+namespace nix { class Store; }
+
+namespace nix::fetchers {
+
+struct TreeInfo
+{
+    Hash narHash;
+    std::optional<uint64_t> revCount;
+    std::optional<time_t> lastModified;
+
+    bool operator ==(const TreeInfo & other) const
+    {
+        return
+            narHash == other.narHash
+            && revCount == other.revCount
+            && lastModified == other.lastModified;
+    }
+
+    StorePath computeStorePath(Store & store) const;
+};
+
+}