Merge remote-tracking branch 'upstream/master' into hash-always-has-type

14 files changed, 1249 insertions, 688 deletions

diff --git a/src/libfetchers/attrs.cc b/src/libfetchers/attrs.cc
index feb0a6085..1e59faa73 100644
--- a/src/libfetchers/attrs.cc
+++ b/src/libfetchers/attrs.cc
@@ -27,7 +27,7 @@ nlohmann::json attrsToJson(const Attrs & attrs)
 {
     nlohmann::json json;
     for (auto & attr : attrs) {
-        if (auto v = std::get_if<int64_t>(&attr.second)) {
+        if (auto v = std::get_if<uint64_t>(&attr.second)) {
             json[attr.first] = *v;
         } else if (auto v = std::get_if<std::string>(&attr.second)) {
             json[attr.first] = *v;
@@ -55,16 +55,16 @@ std::string getStrAttr(const Attrs & attrs, const std::string & name)
     return *s;
 }
 
-std::optional<int64_t> maybeGetIntAttr(const Attrs & attrs, const std::string & name)
+std::optional<uint64_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))
+    if (auto v = std::get_if<uint64_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)
+uint64_t getIntAttr(const Attrs & attrs, const std::string & name)
 {
     auto s = maybeGetIntAttr(attrs, name);
     if (!s)
@@ -76,8 +76,8 @@ std::optional<bool> maybeGetBoolAttr(const Attrs & attrs, const std::string & na
 {
     auto i = attrs.find(name);
     if (i == attrs.end()) return {};
-    if (auto v = std::get_if<int64_t>(&i->second))
-        return *v;
+    if (auto v = std::get_if<Explicit<bool>>(&i->second))
+        return v->t;
     throw Error("input attribute '%s' is not a Boolean", name);
 }
 
@@ -93,7 +93,7 @@ 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)) {
+        if (auto v = std::get_if<uint64_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);
diff --git a/src/libfetchers/attrs.hh b/src/libfetchers/attrs.hh
index d6e0ae000..4b4630c80 100644
--- a/src/libfetchers/attrs.hh
+++ b/src/libfetchers/attrs.hh
@@ -13,9 +13,14 @@ namespace nix::fetchers {
 template<typename T>
 struct Explicit {
     T t;
+
+    bool operator ==(const Explicit<T> & other) const
+    {
+        return t == other.t;
+    }
 };
 
-typedef std::variant<std::string, int64_t, Explicit<bool>> Attr;
+typedef std::variant<std::string, uint64_t, Explicit<bool>> Attr;
 typedef std::map<std::string, Attr> Attrs;
 
 Attrs jsonToAttrs(const nlohmann::json & json);
@@ -26,9 +31,9 @@ std::optional<std::string> maybeGetStrAttr(const Attrs & attrs, const std::strin
 
 std::string getStrAttr(const Attrs & attrs, const std::string & name);
 
-std::optional<int64_t> maybeGetIntAttr(const Attrs & attrs, const std::string & name);
+std::optional<uint64_t> maybeGetIntAttr(const Attrs & attrs, const std::string & name);
 
-int64_t getIntAttr(const Attrs & attrs, const std::string & name);
+uint64_t getIntAttr(const Attrs & attrs, const std::string & name);
 
 std::optional<bool> maybeGetBoolAttr(const Attrs & attrs, const std::string & name);
 
diff --git a/src/libfetchers/fetchers.cc b/src/libfetchers/fetchers.cc
index 9174c3de4..c3ee9bf43 100644
--- a/src/libfetchers/fetchers.cc
+++ b/src/libfetchers/fetchers.cc
@@ -5,71 +5,265 @@
 
 namespace nix::fetchers {
 
-std::unique_ptr<std::vector<std::unique_ptr<InputScheme>>> inputSchemes = nullptr;
+std::unique_ptr<std::vector<std::shared_ptr<InputScheme>>> inputSchemes = nullptr;
 
-void registerInputScheme(std::unique_ptr<InputScheme> && inputScheme)
+void registerInputScheme(std::shared_ptr<InputScheme> && inputScheme)
 {
-    if (!inputSchemes) inputSchemes = std::make_unique<std::vector<std::unique_ptr<InputScheme>>>();
+    if (!inputSchemes) inputSchemes = std::make_unique<std::vector<std::shared_ptr<InputScheme>>>();
     inputSchemes->push_back(std::move(inputScheme));
 }
 
-std::unique_ptr<Input> inputFromURL(const ParsedURL & url)
+Input Input::fromURL(const std::string & url)
+{
+    return fromURL(parseURL(url));
+}
+
+static void fixupInput(Input & input)
+{
+    // Check common attributes.
+    input.getType();
+    input.getRef();
+    if (input.getRev())
+        input.immutable = true;
+    input.getRevCount();
+    input.getLastModified();
+    if (input.getNarHash())
+        input.immutable = true;
+}
+
+Input Input::fromURL(const ParsedURL & url)
 {
     for (auto & inputScheme : *inputSchemes) {
         auto res = inputScheme->inputFromURL(url);
-        if (res) return res;
+        if (res) {
+            res->scheme = inputScheme;
+            fixupInput(*res);
+            return std::move(*res);
+        }
     }
-    throw Error("input '%s' is unsupported", url.url);
-}
 
-std::unique_ptr<Input> inputFromURL(const std::string & url)
-{
-    return inputFromURL(parseURL(url));
+    throw Error("input '%s' is unsupported", url.url);
 }
 
-std::unique_ptr<Input> inputFromAttrs(const Attrs & attrs)
+Input Input::fromAttrs(Attrs && attrs)
 {
-    auto attrs2(attrs);
-    attrs2.erase("narHash");
     for (auto & inputScheme : *inputSchemes) {
-        auto res = inputScheme->inputFromAttrs(attrs2);
+        auto res = inputScheme->inputFromAttrs(attrs);
         if (res) {
-            if (auto narHash = maybeGetStrAttr(attrs, "narHash"))
-                // FIXME: require SRI hash.
-                res->narHash = newHashAllowEmpty(*narHash, {});
-            return res;
+            res->scheme = inputScheme;
+            fixupInput(*res);
+            return std::move(*res);
         }
     }
-    throw Error("input '%s' is unsupported", attrsToJson(attrs));
+
+    Input input;
+    input.attrs = attrs;
+    fixupInput(input);
+    return input;
+}
+
+ParsedURL Input::toURL() const
+{
+    if (!scheme)
+        throw Error("cannot show unsupported input '%s'", attrsToJson(attrs));
+    return scheme->toURL(*this);
+}
+
+std::string Input::to_string() const
+{
+    return toURL().to_string();
 }
 
 Attrs Input::toAttrs() const
 {
-    auto attrs = toAttrsInternal();
-    if (narHash)
-        attrs.emplace("narHash", narHash->to_string(SRI, true));
-    attrs.emplace("type", type());
     return attrs;
 }
 
-std::pair<Tree, std::shared_ptr<const Input>> Input::fetchTree(ref<Store> store) const
+bool Input::hasAllInfo() const
 {
-    auto [tree, input] = fetchTreeInternal(store);
+    return getNarHash() && scheme && scheme->hasAllInfo(*this);
+}
+
+bool Input::operator ==(const Input & other) const
+{
+    return attrs == other.attrs;
+}
+
+bool Input::contains(const Input & other) const
+{
+    if (*this == other) return true;
+    auto other2(other);
+    other2.attrs.erase("ref");
+    other2.attrs.erase("rev");
+    if (*this == other2) return true;
+    return false;
+}
+
+std::pair<Tree, Input> Input::fetch(ref<Store> store) const
+{
+    if (!scheme)
+        throw Error("cannot fetch unsupported input '%s'", attrsToJson(toAttrs()));
+
+    /* The tree may already be in the Nix store, or it could be
+       substituted (which is often faster than fetching from the
+       original source). So check that. */
+    if (hasAllInfo()) {
+        try {
+            auto storePath = computeStorePath(*store);
+
+            store->ensurePath(storePath);
+
+            debug("using substituted/cached input '%s' in '%s'",
+                to_string(), store->printStorePath(storePath));
+
+            auto actualPath = store->toRealPath(storePath);
+
+            return {fetchers::Tree(std::move(actualPath), std::move(storePath)), *this};
+        } catch (Error & e) {
+            debug("substitution of input '%s' failed: %s", to_string(), e.what());
+        }
+    }
+
+    auto [tree, input] = scheme->fetch(store, *this);
 
     if (tree.actualPath == "")
         tree.actualPath = store->toRealPath(tree.storePath);
 
-    if (!tree.info.narHash)
-        tree.info.narHash = store->queryPathInfo(tree.storePath)->narHash;
+    auto narHash = store->queryPathInfo(tree.storePath)->narHash;
+    input.attrs.insert_or_assign("narHash", narHash->to_string(SRI, true));
+
+    if (auto prevNarHash = getNarHash()) {
+        if (narHash != *prevNarHash)
+            throw Error("NAR hash mismatch in input '%s' (%s), expected '%s', got '%s'",
+                to_string(), tree.actualPath, prevNarHash->to_string(SRI, true), narHash->to_string(SRI, true));
+    }
 
-    if (input->narHash)
-        assert(input->narHash == tree.info.narHash);
+    if (auto prevLastModified = getLastModified()) {
+        if (input.getLastModified() != prevLastModified)
+            throw Error("'lastModified' attribute mismatch in input '%s', expected %d",
+                input.to_string(), *prevLastModified);
+    }
 
-    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, true), input->narHash->to_string(SRI, true));
+    if (auto prevRevCount = getRevCount()) {
+        if (input.getRevCount() != prevRevCount)
+            throw Error("'revCount' attribute mismatch in input '%s', expected %d",
+                input.to_string(), *prevRevCount);
+    }
+
+    input.immutable = true;
+
+    assert(input.hasAllInfo());
 
     return {std::move(tree), input};
 }
 
+Input Input::applyOverrides(
+    std::optional<std::string> ref,
+    std::optional<Hash> rev) const
+{
+    if (!scheme) return *this;
+    return scheme->applyOverrides(*this, ref, rev);
+}
+
+void Input::clone(const Path & destDir) const
+{
+    assert(scheme);
+    scheme->clone(*this, destDir);
+}
+
+std::optional<Path> Input::getSourcePath() const
+{
+    assert(scheme);
+    return scheme->getSourcePath(*this);
+}
+
+void Input::markChangedFile(
+    std::string_view file,
+    std::optional<std::string> commitMsg) const
+{
+    assert(scheme);
+    return scheme->markChangedFile(*this, file, commitMsg);
+}
+
+StorePath Input::computeStorePath(Store & store) const
+{
+    auto narHash = getNarHash();
+    if (!narHash)
+        throw Error("cannot compute store path for mutable input '%s'", to_string());
+    return store.makeFixedOutputPath(FileIngestionMethod::Recursive, *narHash, "source");
+}
+
+std::string Input::getType() const
+{
+    return getStrAttr(attrs, "type");
+}
+
+std::optional<Hash> Input::getNarHash() const
+{
+    if (auto s = maybeGetStrAttr(attrs, "narHash"))
+        // FIXME: require SRI hash.
+        return newHashAllowEmpty(*s, htSHA256);
+    return {};
+}
+
+std::optional<std::string> Input::getRef() const
+{
+    if (auto s = maybeGetStrAttr(attrs, "ref"))
+        return *s;
+    return {};
+}
+
+std::optional<Hash> Input::getRev() const
+{
+    if (auto s = maybeGetStrAttr(attrs, "rev"))
+        return Hash(*s, htSHA1);
+    return {};
+}
+
+std::optional<uint64_t> Input::getRevCount() const
+{
+    if (auto n = maybeGetIntAttr(attrs, "revCount"))
+        return *n;
+    return {};
+}
+
+std::optional<time_t> Input::getLastModified() const
+{
+    if (auto n = maybeGetIntAttr(attrs, "lastModified"))
+        return *n;
+    return {};
+}
+
+ParsedURL InputScheme::toURL(const Input & input)
+{
+    throw Error("don't know how to convert input '%s' to a URL", attrsToJson(input.attrs));
+}
+
+Input InputScheme::applyOverrides(
+    const Input & input,
+    std::optional<std::string> ref,
+    std::optional<Hash> rev)
+{
+    if (ref)
+        throw Error("don't know how to set branch/tag name of input '%s' to '%s'", input.to_string(), *ref);
+    if (rev)
+        throw Error("don't know how to set revision of input '%s' to '%s'", input.to_string(), rev->gitRev());
+    return input;
+}
+
+std::optional<Path> InputScheme::getSourcePath(const Input & input)
+{
+    return {};
+}
+
+void InputScheme::markChangedFile(const Input & input, std::string_view file, std::optional<std::string> commitMsg)
+{
+    assert(false);
+}
+
+void InputScheme::clone(const Input & input, const Path & destDir)
+{
+    throw Error("do not know how to clone input '%s'", input.to_string());
+}
+
 }
diff --git a/src/libfetchers/fetchers.hh b/src/libfetchers/fetchers.hh
index 59a58ae67..be71b786b 100644
--- a/src/libfetchers/fetchers.hh
+++ b/src/libfetchers/fetchers.hh
@@ -3,7 +3,6 @@
 #include "types.hh"
 #include "hash.hh"
 #include "path.hh"
-#include "tree-info.hh"
 #include "attrs.hh"
 #include "url.hh"
 
@@ -13,73 +12,101 @@ namespace nix { class Store; }
 
 namespace nix::fetchers {
 
-struct Input;
-
 struct Tree
 {
     Path actualPath;
     StorePath storePath;
-    TreeInfo info;
+    Tree(Path && actualPath, StorePath && storePath) : actualPath(actualPath), storePath(std::move(storePath)) {}
 };
 
-struct Input : std::enable_shared_from_this<Input>
+struct InputScheme;
+
+struct Input
 {
-    std::optional<Hash> narHash; // FIXME: implement
+    friend class InputScheme;
+
+    std::shared_ptr<InputScheme> scheme; // note: can be null
+    Attrs attrs;
+    bool immutable = false;
+    bool direct = true;
+
+public:
+    static Input fromURL(const std::string & url);
 
-    virtual std::string type() const = 0;
+    static Input fromURL(const ParsedURL & url);
 
-    virtual ~Input() { }
+    static Input fromAttrs(Attrs && attrs);
 
-    virtual bool operator ==(const Input & other) const { return false; }
+    ParsedURL toURL() const;
+
+    std::string to_string() const;
+
+    Attrs toAttrs() const;
 
     /* Check whether this is a "direct" input, that is, not
        one that goes through a registry. */
-    virtual bool isDirect() const { return true; }
+    bool isDirect() const { return direct; }
 
     /* 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; }
+    bool isImmutable() const { return immutable; }
 
-    virtual bool contains(const Input & other) const { return false; }
+    bool hasAllInfo() const;
 
-    virtual std::optional<std::string> getRef() const { return {}; }
+    bool operator ==(const Input & other) const;
 
-    virtual std::optional<Hash> getRev() const { return {}; }
+    bool contains(const Input & other) const;
 
-    virtual ParsedURL toURL() const = 0;
+    std::pair<Tree, Input> fetch(ref<Store> store) const;
 
-    std::string to_string() const
-    {
-        return toURL().to_string();
-    }
+    Input applyOverrides(
+        std::optional<std::string> ref,
+        std::optional<Hash> rev) const;
 
-    Attrs toAttrs() const;
+    void clone(const Path & destDir) const;
 
-    std::pair<Tree, std::shared_ptr<const Input>> fetchTree(ref<Store> store) const;
+    std::optional<Path> getSourcePath() const;
 
-private:
+    void markChangedFile(
+        std::string_view file,
+        std::optional<std::string> commitMsg) const;
 
-    virtual std::pair<Tree, std::shared_ptr<const Input>> fetchTreeInternal(ref<Store> store) const = 0;
+    StorePath computeStorePath(Store & store) const;
 
-    virtual Attrs toAttrsInternal() const = 0;
+    // Convience functions for common attributes.
+    std::string getType() const;
+    std::optional<Hash> getNarHash() const;
+    std::optional<std::string> getRef() const;
+    std::optional<Hash> getRev() const;
+    std::optional<uint64_t> getRevCount() const;
+    std::optional<time_t> getLastModified() const;
 };
 
 struct InputScheme
 {
-    virtual ~InputScheme() { }
+    virtual std::optional<Input> inputFromURL(const ParsedURL & url) = 0;
 
-    virtual std::unique_ptr<Input> inputFromURL(const ParsedURL & url) = 0;
+    virtual std::optional<Input> inputFromAttrs(const Attrs & attrs) = 0;
 
-    virtual std::unique_ptr<Input> inputFromAttrs(const Attrs & attrs) = 0;
-};
+    virtual ParsedURL toURL(const Input & input);
+
+    virtual bool hasAllInfo(const Input & input) = 0;
 
-std::unique_ptr<Input> inputFromURL(const ParsedURL & url);
+    virtual Input applyOverrides(
+        const Input & input,
+        std::optional<std::string> ref,
+        std::optional<Hash> rev);
 
-std::unique_ptr<Input> inputFromURL(const std::string & url);
+    virtual void clone(const Input & input, const Path & destDir);
 
-std::unique_ptr<Input> inputFromAttrs(const Attrs & attrs);
+    virtual std::optional<Path> getSourcePath(const Input & input);
+
+    virtual void markChangedFile(const Input & input, std::string_view file, std::optional<std::string> commitMsg);
+
+    virtual std::pair<Tree, Input> fetch(ref<Store> store, const Input & input) = 0;
+};
 
-void registerInputScheme(std::unique_ptr<InputScheme> && fetcher);
+void registerInputScheme(std::shared_ptr<InputScheme> && fetcher);
 
 struct DownloadFileResult
 {
@@ -94,7 +121,7 @@ DownloadFileResult downloadFile(
     const std::string & name,
     bool immutable);
 
-Tree downloadTarball(
+std::pair<Tree, time_t> downloadTarball(
     ref<Store> store,
     const std::string & url,
     const std::string & name,
diff --git a/src/libfetchers/git.cc b/src/libfetchers/git.cc
index 75ce5ee8b..5d38e0c2b 100644
--- a/src/libfetchers/git.cc
+++ b/src/libfetchers/git.cc
@@ -22,80 +22,152 @@ static bool isNotDotGitDirectory(const Path & path)
     return not std::regex_match(path, gitDirRegex);
 }
 
-struct GitInput : Input
+struct GitInputScheme : InputScheme
 {
-    ParsedURL url;
-    std::optional<std::string> ref;
-    std::optional<Hash> rev;
-    bool shallow = false;
-    bool submodules = false;
+    std::optional<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 {};
+
+        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 if (name == "shallow")
+                attrs.emplace(name, Explicit<bool> { value == "1" });
+            else
+                url2.query.emplace(name, value);
+        }
+
+        attrs.emplace("url", url2.to_string());
+
+        return inputFromAttrs(attrs);
+    }
 
-    GitInput(const ParsedURL & url) : url(url)
-    { }
+    std::optional<Input> inputFromAttrs(const Attrs & attrs) override
+    {
+        if (maybeGetStrAttr(attrs, "type") != "git") return {};
 
-    std::string type() const override { return "git"; }
+        for (auto & [name, value] : attrs)
+            if (name != "type" && name != "url" && name != "ref" && name != "rev" && name != "shallow" && name != "submodules" && name != "lastModified" && name != "revCount" && name != "narHash")
+                throw Error("unsupported Git input attribute '%s'", name);
+
+        parseURL(getStrAttr(attrs, "url"));
+        maybeGetBoolAttr(attrs, "shallow");
+        maybeGetBoolAttr(attrs, "submodules");
+
+        if (auto ref = maybeGetStrAttr(attrs, "ref")) {
+            if (std::regex_search(*ref, badGitRefRegex))
+                throw BadURL("invalid Git branch/tag name '%s'", *ref);
+        }
+
+        Input input;
+        input.attrs = attrs;
+        return input;
+    }
 
-    bool operator ==(const Input & other) const override
+    ParsedURL toURL(const Input & input) override
     {
-        auto other2 = dynamic_cast<const GitInput *>(&other);
+        auto url = parseURL(getStrAttr(input.attrs, "url"));
+        if (url.scheme != "git") url.scheme = "git+" + url.scheme;
+        if (auto rev = input.getRev()) url.query.insert_or_assign("rev", rev->gitRev());
+        if (auto ref = input.getRef()) url.query.insert_or_assign("ref", *ref);
+        if (maybeGetBoolAttr(input.attrs, "shallow").value_or(false))
+            url.query.insert_or_assign("shallow", "1");
+        return url;
+    }
+
+    bool hasAllInfo(const Input & input) override
+    {
+        bool maybeDirty = !input.getRef();
+        bool shallow = maybeGetBoolAttr(input.attrs, "shallow").value_or(false);
         return
-            other2
-            && url == other2->url
-            && rev == other2->rev
-            && ref == other2->ref;
+            maybeGetIntAttr(input.attrs, "lastModified")
+            && (shallow || maybeDirty || maybeGetIntAttr(input.attrs, "revCount"));
     }
 
-    bool isImmutable() const override
+    Input applyOverrides(
+        const Input & input,
+        std::optional<std::string> ref,
+        std::optional<Hash> rev) override
     {
-        return (bool) rev || narHash;
+        auto res(input);
+        if (rev) res.attrs.insert_or_assign("rev", rev->gitRev());
+        if (ref) res.attrs.insert_or_assign("ref", *ref);
+        if (!res.getRef() && res.getRev())
+            throw Error("Git input '%s' has a commit hash but no branch/tag name", res.to_string());
+        return res;
     }
 
-    std::optional<std::string> getRef() const override { return ref; }
+    void clone(const Input & input, const Path & destDir) override
+    {
+        auto [isLocal, actualUrl] = getActualUrl(input);
+
+        Strings args = {"clone"};
+
+        args.push_back(actualUrl);
 
-    std::optional<Hash> getRev() const override { return rev; }
+        if (auto ref = input.getRef()) {
+            args.push_back("--branch");
+            args.push_back(*ref);
+        }
+
+        if (input.getRev()) throw Error("cloning a specific revision is not implemented");
+
+        args.push_back(destDir);
 
-    ParsedURL toURL() const override
+        runProgram("git", true, args);
+    }
+
+    std::optional<Path> getSourcePath(const Input & input) 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;
+        auto url = parseURL(getStrAttr(input.attrs, "url"));
+        if (url.scheme == "file" && !input.getRef() && !input.getRev())
+            return url.path;
+        return {};
     }
 
-    Attrs toAttrsInternal() const override
+    void markChangedFile(const Input & input, std::string_view file, std::optional<std::string> commitMsg) 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;
+        auto sourcePath = getSourcePath(input);
+        assert(sourcePath);
+
+        runProgram("git", true,
+            { "-C", *sourcePath, "add", "--force", "--intent-to-add", "--", std::string(file) });
+
+        if (commitMsg)
+            runProgram("git", true,
+                { "-C", *sourcePath, "commit", std::string(file), "-m", *commitMsg });
     }
 
-    std::pair<bool, std::string> getActualUrl() const
+    std::pair<bool, std::string> getActualUrl(const Input & input) 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
+        auto url = parseURL(getStrAttr(input.attrs, "url"));
         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
+    std::pair<Tree, Input> fetch(ref<Store> store, const Input & _input) override
     {
         auto name = "source";
 
-        auto input = std::make_shared<GitInput>(*this);
+        Input input(_input);
 
-        assert(!rev || rev->type == htSHA1);
+        bool shallow = maybeGetBoolAttr(input.attrs, "shallow").value_or(false);
+        bool submodules = maybeGetBoolAttr(input.attrs, "submodules").value_or(false);
 
         std::string cacheType = "git";
         if (shallow) cacheType += "-shallow";
@@ -106,39 +178,35 @@ struct GitInput : Input
             return Attrs({
                 {"type", cacheType},
                 {"name", name},
-                {"rev", input->rev->gitRev()},
+                {"rev", input.getRev()->gitRev()},
             });
         };
 
         auto makeResult = [&](const Attrs & infoAttrs, StorePath && storePath)
-            -> std::pair<Tree, std::shared_ptr<const Input>>
+            -> std::pair<Tree, Input>
         {
-            assert(input->rev);
-            assert(!rev || rev == input->rev);
+            assert(input.getRev());
+            assert(!_input.getRev() || _input.getRev() == input.getRev());
+            if (!shallow)
+                input.attrs.insert_or_assign("revCount", getIntAttr(infoAttrs, "revCount"));
+            input.attrs.insert_or_assign("lastModified", getIntAttr(infoAttrs, "lastModified"));
             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"),
-                    },
-                },
+                Tree(store->toRealPath(storePath), std::move(storePath)),
                 input
             };
         };
 
-        if (rev) {
+        if (input.getRev()) {
             if (auto res = getCache()->lookup(store, getImmutableAttrs()))
                 return makeResult(res->first, std::move(res->second));
         }
 
-        auto [isLocal, actualUrl_] = getActualUrl();
+        auto [isLocal, actualUrl_] = getActualUrl(input);
         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) {
+        if (!input.getRef() && !input.getRev() && isLocal) {
             bool clean = false;
 
             /* Check whether this repo has any commits. There are
@@ -197,35 +265,35 @@ struct GitInput : Input
 
                 auto storePath = store->addToStore("source", actualUrl, FileIngestionMethod::Recursive, 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,
-                    }
-                };
+                // FIXME: maybe we should use the timestamp of the last
+                // modified dirty file?
+                input.attrs.insert_or_assign(
+                    "lastModified",
+                    haveCommits ? std::stoull(runProgram("git", true, { "-C", actualUrl, "log", "-1", "--format=%ct", "HEAD" })) : 0);
 
-                return {std::move(tree), input};
+                return {
+                    Tree(store->printStorePath(storePath), std::move(storePath)),
+                    input
+                };
             }
         }
 
-        if (!input->ref) input->ref = isLocal ? readHead(actualUrl) : "master";
+        if (!input.getRef()) input.attrs.insert_or_assign("ref", isLocal ? readHead(actualUrl) : "master");
 
         Attrs mutableAttrs({
             {"type", cacheType},
             {"name", name},
             {"url", actualUrl},
-            {"ref", *input->ref},
+            {"ref", *input.getRef()},
         });
 
         Path repoDir;
 
         if (isLocal) {
 
-            if (!input->rev)
-                input->rev = Hash(chomp(runProgram("git", true, { "-C", actualUrl, "rev-parse", *input->ref })), htSHA1);
+            if (!input.getRev())
+                input.attrs.insert_or_assign("rev",
+                    Hash(chomp(runProgram("git", true, { "-C", actualUrl, "rev-parse", *input.getRef() })), htSHA1).gitRev());
 
             repoDir = actualUrl;
 
@@ -233,8 +301,8 @@ struct GitInput : Input
 
             if (auto res = getCache()->lookup(store, mutableAttrs)) {
                 auto rev2 = Hash(getStrAttr(res->first, "rev"), htSHA1);
-                if (!rev || rev == rev2) {
-                    input->rev = rev2;
+                if (!input.getRev() || input.getRev() == rev2) {
+                    input.attrs.insert_or_assign("rev", rev2.gitRev());
                     return makeResult(res->first, std::move(res->second));
                 }
             }
@@ -248,18 +316,18 @@ struct GitInput : Input
             }
 
             Path localRefFile =
-                input->ref->compare(0, 5, "refs/") == 0
-                ? cacheDir + "/" + *input->ref
-                : cacheDir + "/refs/heads/" + *input->ref;
+                input.getRef()->compare(0, 5, "refs/") == 0
+                ? cacheDir + "/" + *input.getRef()
+                : cacheDir + "/refs/heads/" + *input.getRef();
 
             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) {
+            if (input.getRev()) {
                 try {
-                    runProgram("git", true, { "-C", repoDir, "cat-file", "-e", input->rev->gitRev() });
+                    runProgram("git", true, { "-C", repoDir, "cat-file", "-e", input.getRev()->gitRev() });
                     doFetch = false;
                 } catch (ExecError & e) {
                     if (WIFEXITED(e.status)) {
@@ -282,9 +350,10 @@ struct GitInput : Input
                 // FIXME: git stderr messes up our progress indicator, so
                 // we're using --quiet for now. Should process its stderr.
                 try {
-                    auto fetchRef = input->ref->compare(0, 5, "refs/") == 0
-                        ? *input->ref
-                        : "refs/heads/" + *input->ref;
+                    auto ref = input.getRef();
+                    auto fetchRef = ref->compare(0, 5, "refs/") == 0
+                        ? *ref
+                        : "refs/heads/" + *ref;
                     runProgram("git", true, { "-C", repoDir, "fetch", "--quiet", "--force", "--", actualUrl, fmt("%s:%s", fetchRef, fetchRef) });
                 } catch (Error & e) {
                     if (!pathExists(localRefFile)) throw;
@@ -300,8 +369,8 @@ struct GitInput : Input
                 utimes(localRefFile.c_str(), times);
             }
 
-            if (!input->rev)
-                input->rev = Hash(chomp(readFile(localRefFile)), htSHA1);
+            if (!input.getRev())
+                input.attrs.insert_or_assign("rev", Hash(chomp(readFile(localRefFile)), htSHA1).gitRev());
         }
 
         bool isShallow = chomp(runProgram("git", true, { "-C", repoDir, "rev-parse", "--is-shallow-repository" })) == "true";
@@ -311,7 +380,7 @@ struct GitInput : Input
 
         // FIXME: check whether rev is an ancestor of ref.
 
-        printTalkative("using revision %s of repo '%s'", input->rev->gitRev(), actualUrl);
+        printTalkative("using revision %s of repo '%s'", input.getRev()->gitRev(), actualUrl);
 
         /* Now that we know the ref, check again whether we have it in
            the store. */
@@ -333,7 +402,7 @@ struct GitInput : Input
             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, "checkout", "--quiet", input.getRev()->gitRev() });
             runProgram("git", true, { "-C", tmpDir, "remote", "add", "origin", actualUrl });
             runProgram("git", true, { "-C", tmpDir, "submodule", "--quiet", "update", "--init", "--recursive" });
 
@@ -342,7 +411,7 @@ struct GitInput : Input
             // 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() });
+                RunOptions gitOptions("git", { "-C", repoDir, "archive", input.getRev()->gitRev() });
                 gitOptions.standardOut = &sink;
                 runProgram2(gitOptions);
             });
@@ -352,18 +421,18 @@ struct GitInput : Input
 
         auto storePath = store->addToStore(name, tmpDir, FileIngestionMethod::Recursive, htSHA256, filter);
 
-        auto lastModified = std::stoull(runProgram("git", true, { "-C", repoDir, "log", "-1", "--format=%ct", input->rev->gitRev() }));
+        auto lastModified = std::stoull(runProgram("git", true, { "-C", repoDir, "log", "-1", "--format=%ct", input.getRev()->gitRev() }));
 
         Attrs infoAttrs({
-            {"rev", input->rev->gitRev()},
+            {"rev", input.getRev()->gitRev()},
             {"lastModified", lastModified},
         });
 
         if (!shallow)
             infoAttrs.insert_or_assign("revCount",
-                std::stoull(runProgram("git", true, { "-C", repoDir, "rev-list", "--count", input->rev->gitRev() })));
+                std::stoull(runProgram("git", true, { "-C", repoDir, "rev-list", "--count", input.getRev()->gitRev() })));
 
-        if (!this->rev)
+        if (!_input.getRev())
             getCache()->add(
                 store,
                 mutableAttrs,
@@ -382,60 +451,6 @@ struct GitInput : Input
     }
 };
 
-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_search(*ref, badGitRefRegex))
-                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
index 0bee1d6b3..8bb7c2c1d 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/commits?ref_name=%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[0]["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>()); });
 
 }
diff --git a/src/libfetchers/indirect.cc b/src/libfetchers/indirect.cc
new file mode 100644
index 000000000..91dc83740
--- /dev/null
+++ b/src/libfetchers/indirect.cc
@@ -0,0 +1,104 @@
+#include "fetchers.hh"
+
+namespace nix::fetchers {
+
+std::regex flakeRegex("[a-zA-Z][a-zA-Z0-9_-]*", std::regex::ECMAScript);
+
+struct IndirectInputScheme : InputScheme
+{
+    std::optional<Input> inputFromURL(const ParsedURL & url) override
+    {
+        if (url.scheme != "flake") return {};
+
+        auto path = tokenizeString<std::vector<std::string>>(url.path, "/");
+
+        std::optional<Hash> rev;
+        std::optional<std::string> ref;
+
+        if (path.size() == 1) {
+        } else if (path.size() == 2) {
+            if (std::regex_match(path[1], revRegex))
+                rev = Hash(path[1], htSHA1);
+            else if (std::regex_match(path[1], refRegex))
+                ref = path[1];
+            else
+                throw BadURL("in flake URL '%s', '%s' is not a commit hash or branch/tag name", url.url, path[1]);
+        } else if (path.size() == 3) {
+            if (!std::regex_match(path[1], refRegex))
+                throw BadURL("in flake URL '%s', '%s' is not a branch/tag name", url.url, path[1]);
+            ref = path[1];
+            if (!std::regex_match(path[2], revRegex))
+                throw BadURL("in flake URL '%s', '%s' is not a commit hash", url.url, path[2]);
+            rev = Hash(path[2], htSHA1);
+        } else
+            throw BadURL("GitHub URL '%s' is invalid", url.url);
+
+        std::string id = path[0];
+        if (!std::regex_match(id, flakeRegex))
+            throw BadURL("'%s' is not a valid flake ID", id);
+
+        // FIXME: forbid query params?
+
+        Input input;
+        input.direct = false;
+        input.attrs.insert_or_assign("type", "indirect");
+        input.attrs.insert_or_assign("id", id);
+        if (rev) input.attrs.insert_or_assign("rev", rev->gitRev());
+        if (ref) input.attrs.insert_or_assign("ref", *ref);
+
+        return input;
+    }
+
+    std::optional<Input> inputFromAttrs(const Attrs & attrs) override
+    {
+        if (maybeGetStrAttr(attrs, "type") != "indirect") return {};
+
+        for (auto & [name, value] : attrs)
+            if (name != "type" && name != "id" && name != "ref" && name != "rev" && name != "narHash")
+                throw Error("unsupported indirect input attribute '%s'", name);
+
+        auto id = getStrAttr(attrs, "id");
+        if (!std::regex_match(id, flakeRegex))
+            throw BadURL("'%s' is not a valid flake ID", id);
+
+        Input input;
+        input.direct = false;
+        input.attrs = attrs;
+        return input;
+    }
+
+    ParsedURL toURL(const Input & input) override
+    {
+        ParsedURL url;
+        url.scheme = "flake";
+        url.path = getStrAttr(input.attrs, "id");
+        if (auto ref = input.getRef()) { url.path += '/'; url.path += *ref; };
+        if (auto rev = input.getRev()) { url.path += '/'; url.path += rev->gitRev(); };
+        return url;
+    }
+
+    bool hasAllInfo(const Input & input) override
+    {
+        return false;
+    }
+
+    Input applyOverrides(
+        const Input & _input,
+        std::optional<std::string> ref,
+        std::optional<Hash> rev) override
+    {
+        auto input(_input);
+        if (rev) input.attrs.insert_or_assign("rev", rev->gitRev());
+        if (ref) input.attrs.insert_or_assign("ref", *ref);
+        return input;
+    }
+
+    std::pair<Tree, Input> fetch(ref<Store> store, const Input & input) override
+    {
+        throw Error("indirect input '%s' cannot be fetched directly", input.to_string());
+    }
+};
+
+static auto r1 = OnStartup([] { registerInputScheme(std::make_unique<IndirectInputScheme>()); });
+
+}
diff --git a/src/libfetchers/mercurial.cc b/src/libfetchers/mercurial.cc
index 2e0d4bf4d..c48cb6fd1 100644
--- a/src/libfetchers/mercurial.cc
+++ b/src/libfetchers/mercurial.cc
@@ -10,76 +10,124 @@ using namespace std::string_literals;
 
 namespace nix::fetchers {
 
-struct MercurialInput : Input
+struct MercurialInputScheme : InputScheme
 {
-    ParsedURL url;
-    std::optional<std::string> ref;
-    std::optional<Hash> rev;
+    std::optional<Input> inputFromURL(const ParsedURL & url) override
+    {
+        if (url.scheme != "hg+http" &&
+            url.scheme != "hg+https" &&
+            url.scheme != "hg+ssh" &&
+            url.scheme != "hg+file") return {};
 
-    MercurialInput(const ParsedURL & url) : url(url)
-    { }
+        auto url2(url);
+        url2.scheme = std::string(url2.scheme, 3);
+        url2.query.clear();
+
+        Attrs attrs;
+        attrs.emplace("type", "hg");
 
-    std::string type() const override { return "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);
+    }
 
-    bool operator ==(const Input & other) const override
+    std::optional<Input> inputFromAttrs(const Attrs & attrs) override
     {
-        auto other2 = dynamic_cast<const MercurialInput *>(&other);
-        return
-            other2
-            && url == other2->url
-            && rev == other2->rev
-            && ref == other2->ref;
+        if (maybeGetStrAttr(attrs, "type") != "hg") return {};
+
+        for (auto & [name, value] : attrs)
+            if (name != "type" && name != "url" && name != "ref" && name != "rev" && name != "revCount" && name != "narHash")
+                throw Error("unsupported Mercurial input attribute '%s'", name);
+
+        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 input;
+        input.attrs = attrs;
+        return input;
     }
 
-    bool isImmutable() const override
+    ParsedURL toURL(const Input & input) override
     {
-        return (bool) rev || narHash;
+        auto url = parseURL(getStrAttr(input.attrs, "url"));
+        url.scheme = "hg+" + url.scheme;
+        if (auto rev = input.getRev()) url.query.insert_or_assign("rev", rev->gitRev());
+        if (auto ref = input.getRef()) url.query.insert_or_assign("ref", *ref);
+        return url;
     }
 
-    std::optional<std::string> getRef() const override { return ref; }
+    bool hasAllInfo(const Input & input) override
+    {
+        // FIXME: ugly, need to distinguish between dirty and clean
+        // default trees.
+        return input.getRef() == "default" || maybeGetIntAttr(input.attrs, "revCount");
+    }
 
-    std::optional<Hash> getRev() const override { return rev; }
+    Input applyOverrides(
+        const Input & input,
+        std::optional<std::string> ref,
+        std::optional<Hash> rev) override
+    {
+        auto res(input);
+        if (rev) res.attrs.insert_or_assign("rev", rev->gitRev());
+        if (ref) res.attrs.insert_or_assign("ref", *ref);
+        return res;
+    }
 
-    ParsedURL toURL() const override
+    std::optional<Path> getSourcePath(const Input & input) 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;
+        auto url = parseURL(getStrAttr(input.attrs, "url"));
+        if (url.scheme == "file" && !input.getRef() && !input.getRev())
+            return url.path;
+        return {};
     }
 
-    Attrs toAttrsInternal() const override
+    void markChangedFile(const Input & input, std::string_view file, std::optional<std::string> commitMsg) override
     {
-        Attrs attrs;
-        attrs.emplace("url", url.to_string());
-        if (ref)
-            attrs.emplace("ref", *ref);
-        if (rev)
-            attrs.emplace("rev", rev->gitRev());
-        return attrs;
+        auto sourcePath = getSourcePath(input);
+        assert(sourcePath);
+
+        // FIXME: shut up if file is already tracked.
+        runProgram("hg", true,
+            { "add", *sourcePath + "/" + std::string(file) });
+
+        if (commitMsg)
+            runProgram("hg", true,
+                { "commit", *sourcePath + "/" + std::string(file), "-m", *commitMsg });
     }
 
-    std::pair<bool, std::string> getActualUrl() const
+    std::pair<bool, std::string> getActualUrl(const Input & input) const
     {
+        auto url = parseURL(getStrAttr(input.attrs, "url"));
         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
+    std::pair<Tree, Input> fetch(ref<Store> store, const Input & _input) override
     {
         auto name = "source";
 
-        auto input = std::make_shared<MercurialInput>(*this);
+        Input input(_input);
 
-        auto [isLocal, actualUrl_] = getActualUrl();
+        auto [isLocal, actualUrl_] = getActualUrl(input);
         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")) {
+        if (!input.getRef() && !input.getRev() && isLocal && pathExists(actualUrl + "/.hg")) {
 
             bool clean = runProgram("hg", true, { "status", "-R", actualUrl, "--modified", "--added", "--removed" }) == "";
 
@@ -94,7 +142,7 @@ struct MercurialInput : Input
                 if (settings.warnDirty)
                     warn("Mercurial tree '%s' is unclean", actualUrl);
 
-                input->ref = chomp(runProgram("hg", true, { "branch", "-R", actualUrl }));
+                input.attrs.insert_or_assign("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);
@@ -116,60 +164,54 @@ struct MercurialInput : Input
 
                 auto storePath = store->addToStore("source", actualUrl, FileIngestionMethod::Recursive, htSHA256, filter);
 
-                return {Tree {
-                    .actualPath = store->printStorePath(storePath),
-                    .storePath = std::move(storePath),
-                }, input};
+                return {
+                    Tree(store->printStorePath(storePath), std::move(storePath)),
+                    input
+                };
             }
         }
 
-        if (!input->ref) input->ref = "default";
+        if (!input.getRef()) input.attrs.insert_or_assign("ref", "default");
 
         auto getImmutableAttrs = [&]()
         {
             return Attrs({
                 {"type", "hg"},
                 {"name", name},
-                {"rev", input->rev->gitRev()},
+                {"rev", input.getRev()->gitRev()},
             });
         };
 
         auto makeResult = [&](const Attrs & infoAttrs, StorePath && storePath)
-            -> std::pair<Tree, std::shared_ptr<const Input>>
+            -> std::pair<Tree, Input>
         {
-            assert(input->rev);
-            assert(!rev || rev == input->rev);
+            assert(input.getRev());
+            assert(!_input.getRev() || _input.getRev() == input.getRev());
+            input.attrs.insert_or_assign("revCount", getIntAttr(infoAttrs, "revCount"));
             return {
-                Tree{
-                    .actualPath = store->toRealPath(storePath),
-                    .storePath = std::move(storePath),
-                    .info = TreeInfo {
-                        .revCount = getIntAttr(infoAttrs, "revCount"),
-                    },
-                },
+                Tree(store->toRealPath(storePath), std::move(storePath)),
                 input
             };
         };
 
-        if (input->rev) {
+        if (input.getRev()) {
             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;
+        auto revOrRef = input.getRev() ? input.getRev()->gitRev() : *input.getRef();
 
         Attrs mutableAttrs({
             {"type", "hg"},
             {"name", name},
             {"url", actualUrl},
-            {"ref", *input->ref},
+            {"ref", *input.getRef()},
         });
 
         if (auto res = getCache()->lookup(store, mutableAttrs)) {
             auto rev2 = Hash(getStrAttr(res->first, "rev"), htSHA1);
-            if (!rev || rev == rev2) {
-                input->rev = rev2;
+            if (!input.getRev() || input.getRev() == rev2) {
+                input.attrs.insert_or_assign("rev", rev2.gitRev());
                 return makeResult(res->first, std::move(res->second));
             }
         }
@@ -178,10 +220,10 @@ struct MercurialInput : Input
 
         /* If this is a commit hash that we already have, we don't
            have to pull again. */
-        if (!(input->rev
+        if (!(input.getRev()
                 && pathExists(cacheDir)
                 && runProgram(
-                    RunOptions("hg", { "log", "-R", cacheDir, "-r", input->rev->gitRev(), "--template", "1" })
+                    RunOptions("hg", { "log", "-R", cacheDir, "-r", input.getRev()->gitRev(), "--template", "1" })
                     .killStderr(true)).second == "1"))
         {
             Activity act(*logger, lvlTalkative, actUnknown, fmt("fetching Mercurial repository '%s'", actualUrl));
@@ -210,9 +252,9 @@ struct MercurialInput : Input
             runProgram("hg", true, { "log", "-R", cacheDir, "-r", revOrRef, "--template", "{node} {rev} {branch}" }));
         assert(tokens.size() == 3);
 
-        input->rev = Hash(tokens[0], htSHA1);
+        input.attrs.insert_or_assign("rev", Hash(tokens[0], htSHA1).gitRev());
         auto revCount = std::stoull(tokens[1]);
-        input->ref = tokens[2];
+        input.attrs.insert_or_assign("ref", tokens[2]);
 
         if (auto res = getCache()->lookup(store, getImmutableAttrs()))
             return makeResult(res->first, std::move(res->second));
@@ -220,18 +262,18 @@ struct MercurialInput : Input
         Path tmpDir = createTempDir();
         AutoDelete delTmpDir(tmpDir, true);
 
-        runProgram("hg", true, { "archive", "-R", cacheDir, "-r", input->rev->gitRev(), tmpDir });
+        runProgram("hg", true, { "archive", "-R", cacheDir, "-r", input.getRev()->gitRev(), tmpDir });
 
         deletePath(tmpDir + "/.hg_archival.txt");
 
         auto storePath = store->addToStore(name, tmpDir);
 
         Attrs infoAttrs({
-            {"rev", input->rev->gitRev()},
+            {"rev", input.getRev()->gitRev()},
             {"revCount", (int64_t) revCount},
         });
 
-        if (!this->rev)
+        if (!_input.getRev())
             getCache()->add(
                 store,
                 mutableAttrs,
@@ -250,54 +292,6 @@ struct MercurialInput : Input
     }
 };
 
-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
index ba2cc192e..99d4b4e8f 100644
--- a/src/libfetchers/path.cc
+++ b/src/libfetchers/path.cc
@@ -3,65 +3,86 @@
 
 namespace nix::fetchers {
 
-struct PathInput : Input
+struct PathInputScheme : InputScheme
 {
-    Path path;
+    std::optional<Input> inputFromURL(const ParsedURL & url) override
+    {
+        if (url.scheme != "path") return {};
 
-    /* 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;
+        if (url.authority && *url.authority != "")
+            throw Error("path URL '%s' should not have an authority ('%s')", url.url, *url.authority);
 
-    std::string type() const override { return "path"; }
+        Input input;
+        input.attrs.insert_or_assign("type", "path");
+        input.attrs.insert_or_assign("path", url.path);
 
-    std::optional<Hash> getRev() const override { return rev; }
+        for (auto & [name, value] : url.query)
+            if (name == "rev" || name == "narHash")
+                input.attrs.insert_or_assign(name, value);
+            else if (name == "revCount" || name == "lastModified") {
+                uint64_t n;
+                if (!string2Int(value, n))
+                    throw Error("path URL '%s' has invalid parameter '%s'", url.to_string(), name);
+                input.attrs.insert_or_assign(name, n);
+            }
+            else
+                throw Error("path URL '%s' has unsupported parameter '%s'", url.to_string(), name);
 
-    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;
+        return input;
     }
 
-    bool isImmutable() const override
+    std::optional<Input> inputFromAttrs(const Attrs & attrs) override
     {
-        return (bool) narHash;
+        if (maybeGetStrAttr(attrs, "type") != "path") return {};
+
+        getStrAttr(attrs, "path");
+
+        for (auto & [name, value] : attrs)
+            /* 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...). */
+            if (name == "type" || name == "rev" || name == "revCount" || name == "lastModified" || name == "narHash" || name == "path")
+                // checked in Input::fromAttrs
+                ;
+            else
+                throw Error("unsupported path input attribute '%s'", name);
+
+        Input input;
+        input.attrs = attrs;
+        return input;
     }
 
-    ParsedURL toURL() const override
+    ParsedURL toURL(const Input & input) override
     {
-        auto query = attrsToQuery(toAttrsInternal());
+        auto query = attrsToQuery(input.attrs);
         query.erase("path");
+        query.erase("type");
         return ParsedURL {
             .scheme = "path",
-            .path = path,
+            .path = getStrAttr(input.attrs, "path"),
             .query = query,
         };
     }
 
-    Attrs toAttrsInternal() const override
+    bool hasAllInfo(const Input & input) override
+    {
+        return true;
+    }
+
+    std::optional<Path> getSourcePath(const Input & input) 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;
+        return getStrAttr(input.attrs, "path");
     }
 
-    std::pair<Tree, std::shared_ptr<const Input>> fetchTreeInternal(nix::ref<Store> store) const override
+    void markChangedFile(const Input & input, std::string_view file, std::optional<std::string> commitMsg) override
     {
-        auto input = std::make_shared<PathInput>(*this);
+        // nothing to do
+    }
+
+    std::pair<Tree, Input> fetch(ref<Store> store, const Input & input) override
+    {
+        auto path = getStrAttr(input.attrs, "path");
 
         // FIXME: check whether access to 'path' is allowed.
 
@@ -74,72 +95,10 @@ struct PathInput : Input
             // 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;
+        return {
+            Tree(store->toRealPath(*storePath), std::move(*storePath)),
+            input
+        };
     }
 };
 
diff --git a/src/libfetchers/registry.cc b/src/libfetchers/registry.cc
new file mode 100644
index 000000000..d4134ce29
--- /dev/null
+++ b/src/libfetchers/registry.cc
@@ -0,0 +1,212 @@
+#include "registry.hh"
+#include "fetchers.hh"
+#include "util.hh"
+#include "globals.hh"
+#include "store-api.hh"
+
+#include <nlohmann/json.hpp>
+
+namespace nix::fetchers {
+
+std::shared_ptr<Registry> Registry::read(
+    const Path & path, RegistryType type)
+{
+    auto registry = std::make_shared<Registry>(type);
+
+    if (!pathExists(path))
+        return std::make_shared<Registry>(type);
+
+    try {
+
+        auto json = nlohmann::json::parse(readFile(path));
+
+        auto version = json.value("version", 0);
+
+        if (version == 2) {
+            for (auto & i : json["flakes"]) {
+                auto toAttrs = jsonToAttrs(i["to"]);
+                Attrs extraAttrs;
+                auto j = toAttrs.find("dir");
+                if (j != toAttrs.end()) {
+                    extraAttrs.insert(*j);
+                    toAttrs.erase(j);
+                }
+                auto exact = i.find("exact");
+                registry->entries.push_back(
+                    Entry {
+                        .from = Input::fromAttrs(jsonToAttrs(i["from"])),
+                        .to = Input::fromAttrs(std::move(toAttrs)),
+                        .extraAttrs = extraAttrs,
+                        .exact = exact != i.end() && exact.value()
+                    });
+            }
+        }
+
+        else
+            throw Error("flake registry '%s' has unsupported version %d", path, version);
+
+    } catch (nlohmann::json::exception & e) {
+        warn("cannot parse flake registry '%s': %s", path, e.what());
+    } catch (Error & e) {
+        warn("cannot read flake registry '%s': %s", path, e.what());
+    }
+
+    return registry;
+}
+
+void Registry::write(const Path & path)
+{
+    nlohmann::json arr;
+    for (auto & entry : entries) {
+        nlohmann::json obj;
+        obj["from"] = attrsToJson(entry.from.toAttrs());
+        obj["to"] = attrsToJson(entry.to.toAttrs());
+        if (!entry.extraAttrs.empty())
+            obj["to"].update(attrsToJson(entry.extraAttrs));
+        if (entry.exact)
+            obj["exact"] = true;
+        arr.emplace_back(std::move(obj));
+    }
+
+    nlohmann::json json;
+    json["version"] = 2;
+    json["flakes"] = std::move(arr);
+
+    createDirs(dirOf(path));
+    writeFile(path, json.dump(2));
+}
+
+void Registry::add(
+    const Input & from,
+    const Input & to,
+    const Attrs & extraAttrs)
+{
+    entries.emplace_back(
+        Entry {
+            .from = from,
+            .to = to,
+            .extraAttrs = extraAttrs
+        });
+}
+
+void Registry::remove(const Input & input)
+{
+    // FIXME: use C++20 std::erase.
+    for (auto i = entries.begin(); i != entries.end(); )
+        if (i->from == input)
+            i = entries.erase(i);
+        else
+            ++i;
+}
+
+static Path getSystemRegistryPath()
+{
+    return settings.nixConfDir + "/registry.json";
+}
+
+static std::shared_ptr<Registry> getSystemRegistry()
+{
+    static auto systemRegistry =
+        Registry::read(getSystemRegistryPath(), Registry::System);
+    return systemRegistry;
+}
+
+Path getUserRegistryPath()
+{
+    return getHome() + "/.config/nix/registry.json";
+}
+
+std::shared_ptr<Registry> getUserRegistry()
+{
+    static auto userRegistry =
+        Registry::read(getUserRegistryPath(), Registry::User);
+    return userRegistry;
+}
+
+static std::shared_ptr<Registry> flagRegistry =
+    std::make_shared<Registry>(Registry::Flag);
+
+std::shared_ptr<Registry> getFlagRegistry()
+{
+    return flagRegistry;
+}
+
+void overrideRegistry(
+    const Input & from,
+    const Input & to,
+    const Attrs & extraAttrs)
+{
+    flagRegistry->add(from, to, extraAttrs);
+}
+
+static std::shared_ptr<Registry> getGlobalRegistry(ref<Store> store)
+{
+    static auto reg = [&]() {
+        auto path = settings.flakeRegistry.get();
+
+        if (!hasPrefix(path, "/")) {
+            auto storePath = downloadFile(store, path, "flake-registry.json", false).storePath;
+            if (auto store2 = store.dynamic_pointer_cast<LocalFSStore>())
+                store2->addPermRoot(storePath, getCacheDir() + "/nix/flake-registry.json", true);
+            path = store->toRealPath(storePath);
+        }
+
+        return Registry::read(path, Registry::Global);
+    }();
+
+    return reg;
+}
+
+Registries getRegistries(ref<Store> store)
+{
+    Registries registries;
+    registries.push_back(getFlagRegistry());
+    registries.push_back(getUserRegistry());
+    registries.push_back(getSystemRegistry());
+    registries.push_back(getGlobalRegistry(store));
+    return registries;
+}
+
+std::pair<Input, Attrs> lookupInRegistries(
+    ref<Store> store,
+    const Input & _input)
+{
+    Attrs extraAttrs;
+    int n = 0;
+    Input input(_input);
+
+ restart:
+
+    n++;
+    if (n > 100) throw Error("cycle detected in flake registry for '%s'", input.to_string());
+
+    for (auto & registry : getRegistries(store)) {
+        // FIXME: O(n)
+        for (auto & entry : registry->entries) {
+            if (entry.exact) {
+                if (entry.from == input) {
+                    input = entry.to;
+                    extraAttrs = entry.extraAttrs;
+                    goto restart;
+                }
+            } else {
+                if (entry.from.contains(input)) {
+                    input = entry.to.applyOverrides(
+                        !entry.from.getRef() && input.getRef() ? input.getRef() : std::optional<std::string>(),
+                        !entry.from.getRev() && input.getRev() ? input.getRev() : std::optional<Hash>());
+                    extraAttrs = entry.extraAttrs;
+                    goto restart;
+                }
+            }
+        }
+    }
+
+    if (!input.isDirect())
+        throw Error("cannot find flake '%s' in the flake registries", input.to_string());
+
+    debug("looked up '%s' -> '%s'", _input.to_string(), input.to_string());
+
+    return {input, extraAttrs};
+}
+
+}
diff --git a/src/libfetchers/registry.hh b/src/libfetchers/registry.hh
new file mode 100644
index 000000000..1077af020
--- /dev/null
+++ b/src/libfetchers/registry.hh
@@ -0,0 +1,64 @@
+#pragma once
+
+#include "types.hh"
+#include "fetchers.hh"
+
+namespace nix { class Store; }
+
+namespace nix::fetchers {
+
+struct Registry
+{
+    enum RegistryType {
+        Flag = 0,
+        User = 1,
+        System = 2,
+        Global = 3,
+    };
+
+    RegistryType type;
+
+    struct Entry
+    {
+        Input from, to;
+        Attrs extraAttrs;
+        bool exact = false;
+    };
+
+    std::vector<Entry> entries;
+
+    Registry(RegistryType type)
+        : type(type)
+    { }
+
+    static std::shared_ptr<Registry> read(
+        const Path & path, RegistryType type);
+
+    void write(const Path & path);
+
+    void add(
+        const Input & from,
+        const Input & to,
+        const Attrs & extraAttrs);
+
+    void remove(const Input & input);
+};
+
+typedef std::vector<std::shared_ptr<Registry>> Registries;
+
+std::shared_ptr<Registry> getUserRegistry();
+
+Path getUserRegistryPath();
+
+Registries getRegistries(ref<Store> store);
+
+void overrideRegistry(
+    const Input & from,
+    const Input & to,
+    const Attrs & extraAttrs);
+
+std::pair<Input, Attrs> lookupInRegistries(
+    ref<Store> store,
+    const Input & input);
+
+}
diff --git a/src/libfetchers/tarball.cc b/src/libfetchers/tarball.cc
index f5356f0af..55158cece 100644
--- a/src/libfetchers/tarball.cc
+++ b/src/libfetchers/tarball.cc
@@ -105,7 +105,7 @@ DownloadFileResult downloadFile(
     };
 }
 
-Tree downloadTarball(
+std::pair<Tree, time_t> downloadTarball(
     ref<Store> store,
     const std::string & url,
     const std::string & name,
@@ -120,12 +120,9 @@ Tree downloadTarball(
     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"),
-            },
+        return {
+            Tree(store->toRealPath(cached->storePath), std::move(cached->storePath)),
+            getIntAttr(cached->infoAttrs, "lastModified")
         };
 
     auto res = downloadFile(store, url, name, immutable);
@@ -160,117 +157,72 @@ Tree downloadTarball(
         *unpackedStorePath,
         immutable);
 
-    return Tree {
-        .actualPath = store->toRealPath(*unpackedStorePath),
-        .storePath = std::move(*unpackedStorePath),
-        .info = TreeInfo {
-            .lastModified = lastModified,
-        },
+    return {
+        Tree(store->toRealPath(*unpackedStorePath), std::move(*unpackedStorePath)),
+        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, true));
-        else if (hash)
-            url2.query.insert_or_assign("hash", hash->to_string(SRI, true));
-        return url2;
-    }
-
-    Attrs toAttrsInternal() const override
-    {
-        Attrs attrs;
-        attrs.emplace("url", url.to_string());
-        if (hash)
-            attrs.emplace("hash", hash->to_string(SRI, true));
-        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
+    std::optional<Input> inputFromURL(const ParsedURL & url) override
     {
-        if (url.scheme != "file" && url.scheme != "http" && url.scheme != "https") return nullptr;
+        if (url.scheme != "file" && url.scheme != "http" && url.scheme != "https") return {};
 
         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 input;
+        input.attrs.insert_or_assign("type", "tarball");
+        input.attrs.insert_or_assign("url", url.to_string());
+        auto narHash = url.query.find("narHash");
+        if (narHash != url.query.end())
+            input.attrs.insert_or_assign("narHash", narHash->second);
         return input;
     }
 
-    std::unique_ptr<Input> inputFromAttrs(const Attrs & attrs) override
+    std::optional<Input> inputFromAttrs(const Attrs & attrs) override
     {
         if (maybeGetStrAttr(attrs, "type") != "tarball") return {};
 
         for (auto & [name, value] : attrs)
-            if (name != "type" && name != "url" && name != "hash")
+            if (name != "type" && name != "url" && /* name != "hash" && */ name != "narHash")
                 throw Error("unsupported tarball input attribute '%s'", name);
 
-        auto input = std::make_unique<TarballInput>(parseURL(getStrAttr(attrs, "url")));
-        if (auto hash = maybeGetStrAttr(attrs, "hash"))
-            input->hash = newHashAllowEmpty(*hash, {});
-
+        Input input;
+        input.attrs = attrs;
+        //input.immutable = (bool) maybeGetStrAttr(input.attrs, "hash");
         return input;
     }
+
+    ParsedURL toURL(const Input & input) override
+    {
+        auto url = parseURL(getStrAttr(input.attrs, "url"));
+        // NAR hashes are preferred over file hashes since tar/zip files
+        // don't have a canonical representation.
+        if (auto narHash = input.getNarHash())
+            url.query.insert_or_assign("narHash", narHash->to_string(SRI, true));
+        /*
+        else if (auto hash = maybeGetStrAttr(input.attrs, "hash"))
+            url.query.insert_or_assign("hash", Hash(*hash).to_string(SRI, true));
+        */
+        return url;
+    }
+
+    bool hasAllInfo(const Input & input) override
+    {
+        return true;
+    }
+
+    std::pair<Tree, Input> fetch(ref<Store> store, const Input & input) override
+    {
+        auto tree = downloadTarball(store, getStrAttr(input.attrs, "url"), "source", false).first;
+        return {std::move(tree), input};
+    }
 };
 
 static auto r1 = OnStartup([] { registerInputScheme(std::make_unique<TarballInputScheme>()); });
diff --git a/src/libfetchers/tree-info.cc b/src/libfetchers/tree-info.cc
deleted file mode 100644
index 432aa6182..000000000
--- a/src/libfetchers/tree-info.cc
+++ /dev/null
@@ -1,14 +0,0 @@
-#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(FileIngestionMethod::Recursive, *narHash, "source");
-}
-
-}
diff --git a/src/libfetchers/tree-info.hh b/src/libfetchers/tree-info.hh
deleted file mode 100644
index 9d1872097..000000000
--- a/src/libfetchers/tree-info.hh
+++ /dev/null
@@ -1,29 +0,0 @@
-#pragma once
-
-#include "path.hh"
-#include "hash.hh"
-
-#include <nlohmann/json_fwd.hpp>
-
-namespace nix { class Store; }
-
-namespace nix::fetchers {
-
-struct TreeInfo
-{
-    std::optional<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;
-};
-
-}