Move StringData to its own header

Now that it isn't just used for `Value`, it doesn't really belong in
there.

Rename `static-string-data.hh` to share the same prefix to keep them
close together when sorting lines, also.

src/libcmd/repl.cc

@@ -656,7 +656,7 @@ ProcessLineResult NixRepl::processLine(std::string line)
                             + "\n\n";
             }
 
-            markdown += stripIndentation(doc->doc);
+            markdown += stripIndentation(doc->doc.view());
 
             logger->cout(trim(renderMarkdownToTerminal(markdown)));
         } else if (fallbackPos) {

src/libexpr-c/nix_api_value.cc

@@ -134,7 +134,7 @@ PrimOp * nix_alloc_primop(
                     .name = name,
                     .args = {},
                     .arity = (size_t) arity,
-                    .doc = doc,
+                    .doc = &nix::StringData::make(doc),
                     .fun = std::bind(nix_c_primop_wrapper, fun, user_data, _1, _2, _3, _4)};
         if (args)
             for (size_t i = 0; args[i]; i++)

src/libexpr-tests/json.cc

@@ -1,6 +1,6 @@
 #include "nix/expr/tests/libexpr.hh"
 #include "nix/expr/value-to-json.hh"
-#include "nix/expr/static-string-data.hh"
+#include "nix/expr/string-data-static.hh"
 
 namespace nix {
 // Testing the conversion to JSON

src/libexpr-tests/value/print.cc

@@ -1,5 +1,5 @@
 #include "nix/expr/tests/libexpr.hh"
-#include "nix/expr/static-string-data.hh"
+#include "nix/expr/string-data-static.hh"
 
 #include "nix/expr/value.hh"
 #include "nix/expr/print.hh"

src/libexpr-tests/value/value.cc

@@ -1,5 +1,5 @@
 #include "nix/expr/value.hh"
-#include "nix/expr/static-string-data.hh"
+#include "nix/expr/string-data-static.hh"
 
 #include "nix/store/tests/libstore.hh"
 #include <gtest/gtest.h>

src/libexpr/eval.cc

@@ -51,36 +51,6 @@ using json = nlohmann::json;
 
 namespace nix {
 
-/**
- * Just for doc strings. Not for regular string values.
- */
-static char * allocString(size_t size)
-{
-    char * t;
-    t = (char *) GC_MALLOC_ATOMIC(size);
-    if (!t)
-        throw std::bad_alloc();
-    return t;
-}
-
-// When there's no need to write to the string, we can optimize away empty
-// string allocations.
-// This function handles makeImmutableString(std::string_view()) by returning
-// the empty string.
-/**
- * Just for doc strings. Not for regular string values.
- */
-static const char * makeImmutableString(std::string_view s)
-{
-    const size_t size = s.size();
-    if (size == 0)
-        return "";
-    auto t = allocString(size + 1);
-    memcpy(t, s.data(), size);
-    t[size] = '\0';
-    return t;
-}
-
 StringData & StringData::alloc(size_t size)
 {
     void * t = GC_MALLOC_ATOMIC(sizeof(StringData) + size + 1);
@@ -517,16 +487,15 @@ Value * EvalState::addPrimOp(PrimOp && primOp)
     if (primOp.arity == 0) {
         primOp.arity = 1;
         auto vPrimOp = allocValue();
-        vPrimOp->mkPrimOp(new PrimOp(std::move(primOp)));
+        vPrimOp->mkPrimOp(new PrimOp(primOp));
         Value v;
         v.mkApp(vPrimOp, vPrimOp);
-        auto & primOp1 = *vPrimOp->primOp();
         return addConstant(
-            primOp1.name,
+            primOp.name,
             v,
             {
                 .type = nThunk, // FIXME
-                .doc = primOp1.doc ? primOp1.doc->c_str() : nullptr,
+                .doc = primOp.doc,
             });
     }
 
@@ -566,14 +535,13 @@ std::optional<EvalState::Doc> EvalState::getDoc(Value & v)
 {
     if (v.isPrimOp()) {
         auto v2 = &v;
-        auto & primOp = *v2->primOp();
-        if (primOp.doc)
+        if (auto * doc = v2->primOp()->doc)
             return Doc{
                 .pos = {},
-                .name = primOp.name,
-                .arity = primOp.arity,
-                .args = primOp.args,
-                .doc = primOp.doc->c_str(),
+                .name = v2->primOp()->name,
+                .arity = v2->primOp()->arity,
+                .args = v2->primOp()->args,
+                .doc = *doc,
             };
     }
     if (v.isLambda()) {
@@ -615,9 +583,8 @@ std::optional<EvalState::Doc> EvalState::getDoc(Value & v)
             .name = name,
             .arity = 0, // FIXME: figure out how deep by syntax only? It's not semantically useful though...
             .args = {},
-            /* N.B. Can't use StringData here, because that would lead to an interior pointer.
-               NOTE: memory leak when compiled without GC. */
-            .doc = makeImmutableString(s.view()),
+            /* NOTE: memory leak when compiled without GC. */
+            .doc = StringData::make(s.view()),
         };
     }
     if (isFunctor(v)) {

src/libexpr/include/nix/expr/eval.hh

@@ -108,7 +108,7 @@ struct PrimOp
     /**
      * Optional free-form documentation about the primop.
      */
-    std::optional<std::string> doc;
+    const StringData * doc = nullptr;
 
     /**
      * Add a trace item, while calling the `<name>` builtin.
@@ -156,7 +156,7 @@ struct Constant
     /**
      * Optional free-form documentation about the constant.
      */
-    const char * doc = nullptr;
+    const StringData * doc = nullptr;
 
     /**
      * Whether the constant is impure, and not available in pure mode.
@@ -837,11 +837,7 @@ public:
         std::optional<std::string> name;
         size_t arity;
         std::vector<std::string> args;
-        /**
-         * Unlike the other `doc` fields in this file, this one should never be
-         * `null`.
-         */
-        const char * doc;
+        const StringData & doc;
     };
 
     /**

src/libexpr/include/nix/expr/meson.build

@@ -31,7 +31,8 @@ headers = [ config_pub_h ] + files(
   'print.hh',
   'repl-exit-status.hh',
   'search-path.hh',
-  'static-string-data.hh',
+  'string-data.hh',
+  'string-data-static.hh',
   'symbol-table.hh',
   'value-to-json.hh',
   'value-to-xml.hh',

src/libexpr/include/nix/expr/nixexpr.hh

@@ -12,7 +12,7 @@
 #include "nix/expr/value.hh"
 #include "nix/expr/symbol-table.hh"
 #include "nix/expr/eval-error.hh"
-#include "nix/expr/static-string-data.hh"
+#include "nix/expr/string-data-static.hh"
 #include "nix/util/pos-idx.hh"
 #include "nix/expr/counter.hh"
 #include "nix/util/pos-table.hh"

src/libexpr/include/nix/expr/parser-state.hh

@@ -5,7 +5,7 @@
 
 #include "nix/expr/eval.hh"
 #include "nix/expr/value.hh"
-#include "nix/expr/static-string-data.hh"
+#include "nix/expr/string-data-static.hh"
 
 namespace nix {
 

src/libexpr/include/nix/expr/string-data-static.hh

@@ -1,7 +1,7 @@
 #pragma once
 ///@file
 
-#include "nix/expr/value.hh"
+#include "nix/expr/string-data.hh"
 
 namespace nix {
 

src/libexpr/include/nix/expr/string-data.hh

@@ -0,0 +1,96 @@
+#pragma once
+///@file
+
+#include <cstddef>
+#include <cstring>
+#include <memory_resource>
+#include <string_view>
+
+namespace nix {
+
+class StringData
+{
+public:
+    using size_type = std::size_t;
+
+    size_type size_;
+    char data_[];
+
+    /*
+     * This in particular ensures that we cannot have a `StringData`
+     * that we use by value, which is just what we want!
+     *
+     * Dynamically sized types aren't a thing in C++ and even flexible array
+     * members are a language extension and beyond the realm of standard C++.
+     * Technically, sizeof data_ member is 0 and the intended way to use flexible
+     * array members is to allocate sizeof(StrindData) + count * sizeof(char) bytes
+     * and the compiler will consider alignment restrictions for the FAM.
+     *
+     */
+
+    StringData(StringData &&) = delete;
+    StringData & operator=(StringData &&) = delete;
+    StringData(const StringData &) = delete;
+    StringData & operator=(const StringData &) = delete;
+    ~StringData() = default;
+
+private:
+    StringData() = delete;
+
+    explicit StringData(size_type size)
+        : size_(size)
+    {
+    }
+
+public:
+    /**
+     * Allocate StringData on the (possibly) GC-managed heap and copy
+     * the contents of s to it.
+     */
+    static const StringData & make(std::string_view s);
+
+    /**
+     * Allocate StringData on the (possibly) GC-managed heap.
+     * @param size Length of the string (without the NUL terminator).
+     */
+    static StringData & alloc(size_t size);
+
+    size_t size() const
+    {
+        return size_;
+    }
+
+    char * data() noexcept
+    {
+        return data_;
+    }
+
+    const char * data() const noexcept
+    {
+        return data_;
+    }
+
+    const char * c_str() const noexcept
+    {
+        return data_;
+    }
+
+    constexpr std::string_view view() const noexcept
+    {
+        return std::string_view(data_, size_);
+    }
+
+    template<size_t N>
+    struct Static;
+
+    static StringData & make(std::pmr::memory_resource & resource, std::string_view s)
+    {
+        auto & res =
+            *new (resource.allocate(sizeof(StringData) + s.size() + 1, alignof(StringData))) StringData(s.size());
+        std::memcpy(res.data_, s.data(), s.size());
+        res.data_[s.size()] = '\0';
+        return res;
+    }
+};
+
+} // namespace nix

src/libexpr/include/nix/expr/symbol-table.hh

@@ -3,7 +3,7 @@
 
 #include <memory_resource>
 #include "nix/expr/value.hh"
-#include "nix/expr/static-string-data.hh"
+#include "nix/expr/string-data-static.hh"
 #include "nix/util/chunked-vector.hh"
 #include "nix/util/error.hh"
 

src/libexpr/include/nix/expr/value.hh

@@ -3,16 +3,14 @@
 
 #include <bit>
 #include <cassert>
-#include <cstddef>
-#include <cstring>
 #include <memory>
-#include <memory_resource>
 #include <span>
 #include <string_view>
 #include <type_traits>
 #include <concepts>
 
 #include "nix/expr/eval-gc.hh"
+#include "nix/expr/string-data.hh"
 #include "nix/expr/value/context.hh"
 #include "nix/util/source-path.hh"
 #include "nix/expr/print-options.hh"
@@ -193,91 +191,6 @@ public:
     friend struct Value;
 };
 
-class StringData
-{
-public:
-    using size_type = std::size_t;
-
-    size_type size_;
-    char data_[];
-
-    /*
-     * This in particular ensures that we cannot have a `StringData`
-     * that we use by value, which is just what we want!
-     *
-     * Dynamically sized types aren't a thing in C++ and even flexible array
-     * members are a language extension and beyond the realm of standard C++.
-     * Technically, sizeof data_ member is 0 and the intended way to use flexible
-     * array members is to allocate sizeof(StrindData) + count * sizeof(char) bytes
-     * and the compiler will consider alignment restrictions for the FAM.
-     *
-     */
-
-    StringData(StringData &&) = delete;
-    StringData & operator=(StringData &&) = delete;
-    StringData(const StringData &) = delete;
-    StringData & operator=(const StringData &) = delete;
-    ~StringData() = default;
-
-private:
-    StringData() = delete;
-
-    explicit StringData(size_type size)
-        : size_(size)
-    {
-    }
-
-public:
-    /**
-     * Allocate StringData on the (possibly) GC-managed heap and copy
-     * the contents of s to it.
-     */
-    static const StringData & make(std::string_view s);
-
-    /**
-     * Allocate StringData on the (possibly) GC-managed heap.
-     * @param size Length of the string (without the NUL terminator).
-     */
-    static StringData & alloc(size_t size);
-
-    size_t size() const
-    {
-        return size_;
-    }
-
-    char * data() noexcept
-    {
-        return data_;
-    }
-
-    const char * data() const noexcept
-    {
-        return data_;
-    }
-
-    const char * c_str() const noexcept
-    {
-        return data_;
-    }
-
-    constexpr std::string_view view() const noexcept
-    {
-        return std::string_view(data_, size_);
-    }
-
-    template<size_t N>
-    struct Static;
-
-    static StringData & make(std::pmr::memory_resource & resource, std::string_view s)
-    {
-        auto & res =
-            *new (resource.allocate(sizeof(StringData) + s.size() + 1, alignof(StringData))) StringData(s.size());
-        std::memcpy(res.data_, s.data(), s.size());
-        res.data_[s.size()] = '\0';
-        return res;
-    }
-};
-
 namespace detail {
 
 /**

src/libexpr/primops/context.cc

@@ -1,5 +1,6 @@
 #include "nix/expr/primops.hh"
 #include "nix/expr/eval-inline.hh"
+#include "nix/expr/string-data-static.hh"
 #include "nix/store/derivations.hh"
 #include "nix/store/store-api.hh"
 #include "nix/store/globals.hh"
@@ -17,9 +18,9 @@ static void prim_unsafeDiscardStringContext(EvalState & state, const PosIdx pos,
 static RegisterPrimOp primop_unsafeDiscardStringContext({
     .name = "__unsafeDiscardStringContext",
     .args = {"s"},
-    .doc = R"(
+    .doc = &R"(
         Discard the [string context](@docroot@/language/string-context.md) from a value that can be coerced to a string.
-    )",
+    )"_sds,
     .fun = prim_unsafeDiscardStringContext,
 });
 
@@ -33,7 +34,7 @@ static void prim_hasContext(EvalState & state, const PosIdx pos, Value ** args,
 static RegisterPrimOp primop_hasContext(
     {.name = "__hasContext",
      .args = {"s"},
-     .doc = R"(
+     .doc = &R"(
       Return `true` if string *s* has a non-empty context.
       The context can be obtained with
       [`getContext`](#builtins-getContext).
@@ -50,7 +51,7 @@ static RegisterPrimOp primop_hasContext(
       > then throw "package name cannot contain string context"
       > else { ${name} = meta; }
       > ```
-    )",
+    )"_sds,
      .fun = prim_hasContext});
 
 static void prim_unsafeDiscardOutputDependency(EvalState & state, const PosIdx pos, Value ** args, Value & v)
@@ -75,7 +76,7 @@ static void prim_unsafeDiscardOutputDependency(EvalState & state, const PosIdx p
 static RegisterPrimOp primop_unsafeDiscardOutputDependency(
     {.name = "__unsafeDiscardOutputDependency",
      .args = {"s"},
-     .doc = R"(
+     .doc = &R"(
       Create a copy of the given string where every
       [derivation deep](@docroot@/language/string-context.md#string-context-element-derivation-deep)
       string context element is turned into a
@@ -91,7 +92,7 @@ static RegisterPrimOp primop_unsafeDiscardOutputDependency(
       Replacing a constant string context element with a "derivation deep" element is a safe operation that just enlargens the string context without forgetting anything.
 
       [`builtins.addDrvOutputDependencies`]: #builtins-addDrvOutputDependencies
-    )",
+    )"_sds,
      .fun = prim_unsafeDiscardOutputDependency});
 
 static void prim_addDrvOutputDependencies(EvalState & state, const PosIdx pos, Value ** args, Value & v)
@@ -143,7 +144,7 @@ static void prim_addDrvOutputDependencies(EvalState & state, const PosIdx pos, V
 static RegisterPrimOp primop_addDrvOutputDependencies(
     {.name = "__addDrvOutputDependencies",
      .args = {"s"},
-     .doc = R"(
+     .doc = &R"(
       Create a copy of the given string where a single
       [constant](@docroot@/language/string-context.md#string-context-constant)
       string context element is turned into a
@@ -156,7 +157,7 @@ static RegisterPrimOp primop_addDrvOutputDependencies(
       The latter is supported so this function is idempotent.
 
       This is the opposite of [`builtins.unsafeDiscardOutputDependency`](#builtins-unsafeDiscardOutputDependency).
-    )",
+    )"_sds,
      .fun = prim_addDrvOutputDependencies});
 
 /* Extract the context of a string as a structured Nix value.
@@ -230,7 +231,7 @@ static void prim_getContext(EvalState & state, const PosIdx pos, Value ** args,
 static RegisterPrimOp primop_getContext(
     {.name = "__getContext",
      .args = {"s"},
-     .doc = R"(
+     .doc = &R"(
       Return the string context of *s*.
 
       The string context tracks references to derivations within a string.
@@ -248,7 +249,7 @@ static RegisterPrimOp primop_getContext(
       ```
       { "/nix/store/arhvjaf6zmlyn8vh8fgn55rpwnxq0n7l-a.drv" = { outputs = [ "out" ]; }; }
       ```
-    )",
+    )"_sds,
      .fun = prim_getContext});
 
 /* Append the given context to a given string.

src/libexpr/primops/fetchClosure.cc

@@ -1,4 +1,5 @@
 #include "nix/expr/primops.hh"
+#include "nix/expr/string-data-static.hh"
 #include "nix/store/store-open.hh"
 #include "nix/store/realisation.hh"
 #include "nix/store/make-content-addressed.hh"
@@ -216,7 +217,7 @@ static void prim_fetchClosure(EvalState & state, const PosIdx pos, Value ** args
 static RegisterPrimOp primop_fetchClosure({
     .name = "__fetchClosure",
     .args = {"args"},
-    .doc = R"(
+    .doc = &R"(
       Fetch a store path [closure](@docroot@/glossary.md#gloss-closure) from a binary cache, and return the store path as a string with context.
 
       This function can be invoked in three ways that we will discuss in order of preference.
@@ -284,7 +285,7 @@ static RegisterPrimOp primop_fetchClosure({
       `fetchClosure` is similar to [`builtins.storePath`](#builtins-storePath) in that it allows you to use a previously built store path in a Nix expression.
       However, `fetchClosure` is more reproducible because it specifies a binary cache from which the path can be fetched.
       Also, using content-addressed store paths does not require users to configure [`trusted-public-keys`](@docroot@/command-ref/conf-file.md#conf-trusted-public-keys) to ensure their authenticity.
-    )",
+    )"_sds,
     .fun = prim_fetchClosure,
     .experimentalFeature = Xp::FetchClosure,
 });

src/libexpr/primops/fetchTree.cc

@@ -2,6 +2,7 @@
 #include "nix/expr/primops.hh"
 #include "nix/expr/eval-inline.hh"
 #include "nix/expr/eval-settings.hh"
+#include "nix/expr/string-data-static.hh"
 #include "nix/store/store-api.hh"
 #include "nix/fetchers/fetchers.hh"
 #include "nix/store/filetransfer.hh"
@@ -235,127 +236,229 @@ static void prim_fetchTree(EvalState & state, const PosIdx pos, Value ** args, V
 static RegisterPrimOp primop_fetchTree({
     .name = "fetchTree",
     .args = {"input"},
-    .doc = []() -> std::string {
-        std::string doc = stripIndentation(R"(
-          Fetch a file system tree or a plain file using one of the supported backends and return an attribute set with:
+    .doc = &R"(
+      Fetch a file system tree or a plain file using one of the supported backends and return an attribute set with:
 
-          - the resulting fixed-output [store path](@docroot@/store/store-path.md)
-          - the corresponding [NAR](@docroot@/store/file-system-object/content-address.md#serial-nix-archive) hash
-          - backend-specific metadata (currently not documented). <!-- TODO: document output attributes -->
+      - the resulting fixed-output [store path](@docroot@/store/store-path.md)
+      - the corresponding [NAR](@docroot@/store/file-system-object/content-address.md#serial-nix-archive) hash
+      - backend-specific metadata (currently not documented). <!-- TODO: document output attributes -->
 
-          *input* must be an attribute set with the following attributes:
+      *input* must be an attribute set with the following attributes:
 
-          - `type` (String, required)
+      - `type` (String, required)
 
-            One of the [supported source types](#source-types).
-            This determines other required and allowed input attributes.
+        One of the [supported source types](#source-types).
+        This determines other required and allowed input attributes.
 
-          - `narHash` (String, optional)
+      - `narHash` (String, optional)
 
-            The `narHash` parameter can be used to substitute the source of the tree.
-            It also allows for verification of tree contents that may not be provided by the underlying transfer mechanism.
-            If `narHash` is set, the source is first looked up is the Nix store and [substituters](@docroot@/command-ref/conf-file.md#conf-substituters), and only fetched if not available.
+        The `narHash` parameter can be used to substitute the source of the tree.
+        It also allows for verification of tree contents that may not be provided by the underlying transfer mechanism.
+        If `narHash` is set, the source is first looked up is the Nix store and [substituters](@docroot@/command-ref/conf-file.md#conf-substituters), and only fetched if not available.
 
-          A subset of the output attributes of `fetchTree` can be re-used for subsequent calls to `fetchTree` to produce the same result again.
-          That is, `fetchTree` is idempotent.
+      A subset of the output attributes of `fetchTree` can be re-used for subsequent calls to `fetchTree` to produce the same result again.
+      That is, `fetchTree` is idempotent.
 
-          Downloads are cached in `$XDG_CACHE_HOME/nix`.
-          The remote source is fetched from the network if both are true:
-          - A NAR hash is supplied and the corresponding store path is not [valid](@docroot@/glossary.md#gloss-validity), that is, not available in the store
+      Downloads are cached in `$XDG_CACHE_HOME/nix`.
+      The remote source is fetched from the network if both are true:
+      - A NAR hash is supplied and the corresponding store path is not [valid](@docroot@/glossary.md#gloss-validity), that is, not available in the store
 
-            > **Note**
+        > **Note**
+        >
+        > [Substituters](@docroot@/command-ref/conf-file.md#conf-substituters) are not used in fetching.
+
+      - There is no cache entry or the cache entry is older than [`tarball-ttl`](@docroot@/command-ref/conf-file.md#conf-tarball-ttl)
+
+      ## Source types
+
+      The following source types and associated input attributes are supported.
+
+      <!-- TODO: It would be soooo much more predictable to work with (and
+      document) if `fetchTree` was a curried call with the first parameter for
+      `type` or an attribute like `builtins.fetchTree.git`! -->
+
+      - `"file"`
+
+        Place a plain file into the Nix store.
+        This is similar to [`builtins.fetchurl`](@docroot@/language/builtins.md#builtins-fetchurl)
+
+        - `url` (String, required)
+
+          Supported protocols:
+
+          - `https`
+
+            > **Example**
             >
-            > [Substituters](@docroot@/command-ref/conf-file.md#conf-substituters) are not used in fetching.
+            > ```nix
+            > fetchTree {
+            >   type = "file";
+            >   url = "https://example.com/index.html";
+            > }
+            > ```
 
-          - There is no cache entry or the cache entry is older than [`tarball-ttl`](@docroot@/command-ref/conf-file.md#conf-tarball-ttl)
+          - `http`
 
-          ## Source types
+            Insecure HTTP transfer for legacy sources.
 
-          The following source types and associated input attributes are supported.
+            > **Warning**
+            >
+            > HTTP performs no encryption or authentication.
+            > Use a `narHash` known in advance to ensure the output has expected contents.
 
-          <!-- TODO: It would be soooo much more predictable to work with (and
-          document) if `fetchTree` was a curried call with the first parameter for
-          `type` or an attribute like `builtins.fetchTree.git`! -->
-        )");
+          - `file`
 
-        auto indentString = [](std::string const & str, std::string const & indent) {
-            std::string result;
-            std::istringstream stream(str);
-            std::string line;
-            bool first = true;
-            while (std::getline(stream, line)) {
-                if (!first)
-                    result += "\n";
-                result += indent + line;
-                first = false;
-            }
-            return result;
-        };
+            A file on the local file system.
 
-        for (const auto & [schemeName, scheme] : fetchers::getAllInputSchemes()) {
-            doc += "\n- `" + quoteString(schemeName, '"') + "`\n\n";
-            doc += indentString(scheme->schemeDescription(), "  ");
-            if (!doc.empty() && doc.back() != '\n')
-                doc += "\n";
+            > **Example**
+            >
+            > ```nix
+            > fetchTree {
+            >   type = "file";
+            >   url = "file:///home/eelco/nix/README.md";
+            > }
+            > ```
 
-            for (const auto & [attrName, attribute] : scheme->allowedAttrs()) {
-                doc += "\n  - `" + attrName + "` (" + attribute.type + ", "
-                       + (attribute.required ? "required" : "optional") + ")\n\n";
-                doc += indentString(stripIndentation(attribute.doc), "    ");
-                if (!doc.empty() && doc.back() != '\n')
-                    doc += "\n";
-            }
-        }
+      - `"git"`
 
-        doc += "\n" + stripIndentation(R"(
-          The following input types are still subject to change:
+        Fetch a Git tree and copy it to the Nix store.
+        This is similar to [`builtins.fetchGit`](@docroot@/language/builtins.md#builtins-fetchGit).
 
-          - `"path"`
-          - `"github"`
-          - `"gitlab"`
-          - `"sourcehut"`
-          - `"mercurial"`
+        - `allRefs` (Bool, optional)
 
-         *input* can also be a [URL-like reference](@docroot@/command-ref/new-cli/nix3-flake.md#flake-references).
-         The additional input types and the URL-like syntax requires the [`flakes` experimental feature](@docroot@/development/experimental-features.md#xp-feature-flakes) to be enabled.
+          By default, this has no effect. This becomes relevant only once `shallow` cloning is disabled.
+
+          Whether to fetch all references (eg. branches and tags) of the repository.
+          With this argument being true, it's possible to load a `rev` from *any* `ref`.
+          (Without setting this option, only `rev`s from the specified `ref` are supported).
+
+          Default: `false`
+
+        - `lastModified` (Integer, optional)
+
+          Unix timestamp of the fetched commit.
+
+          If set, pass through the value to the output attribute set.
+          Otherwise, generated from the fetched Git tree.
+
+        - `lfs` (Bool, optional)
+
+          Fetch any [Git LFS](https://git-lfs.com/) files.
+
+          Default: `false`
+
+        - `ref` (String, optional)
+
+          By default, this has no effect. This becomes relevant only once `shallow` cloning is disabled.
+
+          A [Git reference](https://git-scm.com/book/en/v2/Git-Internals-Git-References), such as a branch or tag name.
+
+          Default: `"HEAD"`
+
+        - `rev` (String, optional)
+
+          A Git revision; a commit hash.
+
+          Default: the tip of `ref`
+
+        - `revCount` (Integer, optional)
+
+          Number of revisions in the history of the Git repository before the fetched commit.
+
+          If set, pass through the value to the output attribute set.
+          Otherwise, generated from the fetched Git tree.
+
+        - `shallow` (Bool, optional)
+
+          Make a shallow clone when fetching the Git tree.
+          When this is enabled, the options `ref` and `allRefs` have no effect anymore.
+
+          Default: `true`
+
+        - `submodules` (Bool, optional)
+
+          Also fetch submodules if available.
+
+          Default: `false`
+
+        - `url` (String, required)
+
+          The URL formats supported are the same as for Git itself.
 
           > **Example**
           >
-          > Fetch a GitHub repository using the attribute set representation:
-          >
           > ```nix
-          > builtins.fetchTree {
-          >   type = "github";
-          >   owner = "NixOS";
-          >   repo = "nixpkgs";
-          >   rev = "ae2e6b3958682513d28f7d633734571fb18285dd";
-          > }
-          > ```
-          >
-          > This evaluates to the following attribute set:
-          >
-          > ```nix
-          > {
-          >   lastModified = 1686503798;
-          >   lastModifiedDate = "20230611171638";
-          >   narHash = "sha256-rA9RqKP9OlBrgGCPvfd5HVAXDOy8k2SmPtB/ijShNXc=";
-          >   outPath = "/nix/store/l5m6qlvfs9sdw14ja3qbzpglcjlb6j1x-source";
-          >   rev = "ae2e6b3958682513d28f7d633734571fb18285dd";
-          >   shortRev = "ae2e6b3";
+          > fetchTree {
+          >   type = "git";
+          >   url = "git@github.com:NixOS/nixpkgs.git";
           > }
           > ```
 
-          > **Example**
+          > **Note**
           >
-          > Fetch the same GitHub repository using the URL-like syntax:
-          >
-          >   ```nix
-          >   builtins.fetchTree "github:NixOS/nixpkgs/ae2e6b3958682513d28f7d633734571fb18285dd"
-          >   ```
-        )");
+          > If the URL points to a local directory, and no `ref` or `rev` is given, Nix only considers files added to the Git index, as listed by `git ls-files` but use the *current file contents* of the Git working directory.
 
-        return doc;
-    }(),
+      - `"tarball"`
+
+        Download a tar archive and extract it into the Nix store.
+        This has the same underlying implementation as [`builtins.fetchTarball`](@docroot@/language/builtins.md#builtins-fetchTarball)
+
+        - `url` (String, required)
+
+           > **Example**
+           >
+           > ```nix
+           > fetchTree {
+           >   type = "tarball";
+           >   url = "https://github.com/NixOS/nixpkgs/tarball/nixpkgs-23.11";
+           > }
+           > ```
+
+      The following input types are still subject to change:
+
+      - `"path"`
+      - `"github"`
+      - `"gitlab"`
+      - `"sourcehut"`
+      - `"mercurial"`
+
+     *input* can also be a [URL-like reference](@docroot@/command-ref/new-cli/nix3-flake.md#flake-references).
+     The additional input types and the URL-like syntax requires the [`flakes` experimental feature](@docroot@/development/experimental-features.md#xp-feature-flakes) to be enabled.
+
+      > **Example**
+      >
+      > Fetch a GitHub repository using the attribute set representation:
+      >
+      > ```nix
+      > builtins.fetchTree {
+      >   type = "github";
+      >   owner = "NixOS";
+      >   repo = "nixpkgs";
+      >   rev = "ae2e6b3958682513d28f7d633734571fb18285dd";
+      > }
+      > ```
+      >
+      > This evaluates to the following attribute set:
+      >
+      > ```nix
+      > {
+      >   lastModified = 1686503798;
+      >   lastModifiedDate = "20230611171638";
+      >   narHash = "sha256-rA9RqKP9OlBrgGCPvfd5HVAXDOy8k2SmPtB/ijShNXc=";
+      >   outPath = "/nix/store/l5m6qlvfs9sdw14ja3qbzpglcjlb6j1x-source";
+      >   rev = "ae2e6b3958682513d28f7d633734571fb18285dd";
+      >   shortRev = "ae2e6b3";
+      > }
+      > ```
+
+      > **Example**
+      >
+      > Fetch the same GitHub repository using the URL-like syntax:
+      >
+      >   ```nix
+      >   builtins.fetchTree "github:NixOS/nixpkgs/ae2e6b3958682513d28f7d633734571fb18285dd"
+      >   ```
+    )"_sds,
     .fun = prim_fetchTree,
     .experimentalFeature = Xp::FetchTree,
 });
@@ -515,7 +618,7 @@ static void prim_fetchurl(EvalState & state, const PosIdx pos, Value ** args, Va
 static RegisterPrimOp primop_fetchurl({
     .name = "__fetchurl",
     .args = {"arg"},
-    .doc = R"(
+    .doc = &R"(
       Download the specified URL and return the path of the downloaded file.
       `arg` can be either a string denoting the URL, or an attribute set with the following attributes:
 
@@ -529,7 +632,7 @@ static RegisterPrimOp primop_fetchurl({
         characters that are invalid for the store.
 
       Not available in [restricted evaluation mode](@docroot@/command-ref/conf-file.md#conf-restrict-eval).
-    )",
+    )"_sds,
     .fun = prim_fetchurl,
 });
 
@@ -541,7 +644,7 @@ static void prim_fetchTarball(EvalState & state, const PosIdx pos, Value ** args
 static RegisterPrimOp primop_fetchTarball({
     .name = "fetchTarball",
     .args = {"args"},
-    .doc = R"(
+    .doc = &R"(
       Download the specified URL, unpack it and return the path of the
       unpacked tree. The file must be a tape archive (`.tar`) compressed
       with `gzip`, `bzip2` or `xz`. If the tarball consists of a
@@ -579,7 +682,7 @@ static RegisterPrimOp primop_fetchTarball({
       ```
 
       Not available in [restricted evaluation mode](@docroot@/command-ref/conf-file.md#conf-restrict-eval).
-    )",
+    )"_sds,
     .fun = prim_fetchTarball,
 });
 
@@ -592,7 +695,7 @@ static void prim_fetchGit(EvalState & state, const PosIdx pos, Value ** args, Va
 static RegisterPrimOp primop_fetchGit({
     .name = "fetchGit",
     .args = {"args"},
-    .doc = R"(
+    .doc = &R"(
       Fetch a path from git. *args* can be a URL, in which case the HEAD
       of the repo at that URL is fetched. Otherwise, it can be an
       attribute with the following attributes (all except `url` optional):
@@ -795,7 +898,7 @@ static RegisterPrimOp primop_fetchGit({
       given, `fetchGit` uses the current content of the checked-out
       files, even if they are not committed or added to Git's index. It
       only considers files added to the Git repository, as listed by `git ls-files`.
-    )",
+    )"_sds,
     .fun = prim_fetchGit,
 });
 

src/libexpr/primops/fromTOML.cc

@@ -1,6 +1,6 @@
 #include "nix/expr/primops.hh"
 #include "nix/expr/eval-inline.hh"
-#include "nix/expr/static-string-data.hh"
+#include "nix/expr/string-data-static.hh"
 
 #include "expr-config-private.hh"
 
@@ -170,10 +170,10 @@ static void prim_fromTOML(EvalState & state, const PosIdx pos, Value ** args, Va
     }
 }
 
-static RegisterPrimOp primop_fromTOML(
-    {.name = "fromTOML",
-     .args = {"e"},
-     .doc = R"(
+static RegisterPrimOp primop_fromTOML({
+    .name = "fromTOML",
+    .args = {"e"},
+    .doc = &R"(
       Convert a TOML string to a Nix value. For example,
 
       ```nix
@@ -186,7 +186,8 @@ static RegisterPrimOp primop_fromTOML(
       ```
 
       returns the value `{ s = "a"; table = { y = 2; }; x = 1; }`.
-    )",
-     .fun = prim_fromTOML});
+    )"_sds,
+    .fun = prim_fromTOML,
+});
 
 } // namespace nix

src/libfetchers/fetchers.cc

@@ -26,9 +26,18 @@ void registerInputScheme(std::shared_ptr<InputScheme> && inputScheme)
         throw Error("Input scheme with name %s already registered", schemeName);
 }
 
-const InputSchemeMap & getAllInputSchemes()
+nlohmann::json dumpRegisterInputSchemeInfo()
 {
-    return inputSchemes();
+    using nlohmann::json;
+
+    auto res = json::object();
+
+    for (auto & [name, scheme] : inputSchemes()) {
+        auto & r = res[name] = json::object();
+        r["allowedAttrs"] = scheme->allowedAttrs();
+    }
+
+    return res;
 }
 
 Input Input::fromURL(const Settings & settings, const std::string & url, bool requireTree)

src/libfetchers/git.cc

@@ -194,183 +194,28 @@ struct GitInputScheme : InputScheme
         return "git";
     }
 
-    std::string schemeDescription() const override
+    StringSet allowedAttrs() const override
     {
-        return stripIndentation(R"(
-          Fetch a Git tree and copy it to the Nix store.
-          This is similar to [`builtins.fetchGit`](@docroot@/language/builtins.md#builtins-fetchGit).
-        )");
-    }
-
-    const std::map<std::string, AttributeInfo> & allowedAttrs() const override
-    {
-        static const std::map<std::string, AttributeInfo> attrs = {
-            {
-                "url",
-                {
-                    .type = "String",
-                    .required = true,
-                    .doc = R"(
-                      The URL formats supported are the same as for Git itself.
-
-                      > **Example**
-                      >
-                      > ```nix
-                      > fetchTree {
-                      >   type = "git";
-                      >   url = "git@github.com:NixOS/nixpkgs.git";
-                      > }
-                      > ```
-
-                      > **Note**
-                      >
-                      > If the URL points to a local directory, and no `ref` or `rev` is given, Nix only considers files added to the Git index, as listed by `git ls-files` but uses the *current file contents* of the Git working directory.
-                    )",
-                },
-            },
-            {
-                "ref",
-                {
-                    .type = "String",
-                    .required = false,
-                    .doc = R"(
-                      By default, this has no effect. This becomes relevant only once `shallow` cloning is disabled.
-
-                      A [Git reference](https://git-scm.com/book/en/v2/Git-Internals-Git-References), such as a branch or tag name.
-
-                      Default: `"HEAD"`
-                    )",
-                },
-            },
-            {
-                "rev",
-                {
-                    .type = "String",
-                    .required = false,
-                    .doc = R"(
-                      A Git revision; a commit hash.
-
-                      Default: the tip of `ref`
-                    )",
-                },
-            },
-            {
-                "shallow",
-                {
-                    .type = "Bool",
-                    .required = false,
-                    .doc = R"(
-                      Make a shallow clone when fetching the Git tree.
-                      When this is enabled, the options `ref` and `allRefs` have no effect anymore.
-
-                      Default: `true`
-                    )",
-                },
-            },
-            {
-                "submodules",
-                {
-                    .type = "Bool",
-                    .required = false,
-                    .doc = R"(
-                      Also fetch submodules if available.
-
-                      Default: `false`
-                    )",
-                },
-            },
-            {
-                "lfs",
-                {
-                    .type = "Bool",
-                    .required = false,
-                    .doc = R"(
-                      Fetch any [Git LFS](https://git-lfs.com/) files.
-
-                      Default: `false`
-                    )",
-                },
-            },
-            {
-                "exportIgnore",
-                {},
-            },
-            {
-                "lastModified",
-                {
-                    .type = "Integer",
-                    .required = false,
-                    .doc = R"(
-                      Unix timestamp of the fetched commit.
-
-                      If set, pass through the value to the output attribute set.
-                      Otherwise, generated from the fetched Git tree.
-                    )",
-                },
-            },
-            {
-                "revCount",
-                {
-                    .type = "Integer",
-                    .required = false,
-                    .doc = R"(
-                      Number of revisions in the history of the Git repository before the fetched commit.
-
-                      If set, pass through the value to the output attribute set.
-                      Otherwise, generated from the fetched Git tree.
-                    )",
-                },
-            },
-            {
-                "narHash",
-                {},
-            },
-            {
-                "allRefs",
-                {
-                    .type = "Bool",
-                    .required = false,
-                    .doc = R"(
-                      By default, this has no effect. This becomes relevant only once `shallow` cloning is disabled.
-
-                      Whether to fetch all references (eg. branches and tags) of the repository.
-                      With this argument being true, it's possible to load a `rev` from *any* `ref`.
-                      (Without setting this option, only `rev`s from the specified `ref` are supported).
-
-                      Default: `false`
-                    )",
-                },
-            },
-            {
-                "name",
-                {},
-            },
-            {
-                "dirtyRev",
-                {},
-            },
-            {
-                "dirtyShortRev",
-                {},
-            },
-            {
-                "verifyCommit",
-                {},
-            },
-            {
-                "keytype",
-                {},
-            },
-            {
-                "publicKey",
-                {},
-            },
-            {
-                "publicKeys",
-                {},
-            },
+        return {
+            "url",
+            "ref",
+            "rev",
+            "shallow",
+            "submodules",
+            "lfs",
+            "exportIgnore",
+            "lastModified",
+            "revCount",
+            "narHash",
+            "allRefs",
+            "name",
+            "dirtyRev",
+            "dirtyShortRev",
+            "verifyCommit",
+            "keytype",
+            "publicKey",
+            "publicKeys",
         };
-        return attrs;
     }
 
     std::optional<Input> inputFromAttrs(const Settings & settings, const Attrs & attrs) const override

src/libfetchers/github.cc

@@ -110,43 +110,18 @@ struct GitArchiveInputScheme : InputScheme
         return input;
     }
 
-    const std::map<std::string, AttributeInfo> & allowedAttrs() const override
+    StringSet allowedAttrs() const override
     {
-        static const std::map<std::string, AttributeInfo> attrs = {
-            {
-                "owner",
-                {},
-            },
-            {
-                "repo",
-                {},
-            },
-            {
-                "ref",
-                {},
-            },
-            {
-                "rev",
-                {},
-            },
-            {
-                "narHash",
-                {},
-            },
-            {
-                "lastModified",
-                {},
-            },
-            {
-                "host",
-                {},
-            },
-            {
-                "treeHash",
-                {},
-            },
+        return {
+            "owner",
+            "repo",
+            "ref",
+            "rev",
+            "narHash",
+            "lastModified",
+            "host",
+            "treeHash",
         };
-        return attrs;
     }
 
     std::optional<Input> inputFromAttrs(const fetchers::Settings & settings, const Attrs & attrs) const override
@@ -386,12 +361,6 @@ struct GitHubInputScheme : GitArchiveInputScheme
         return "github";
     }
 
-    std::string schemeDescription() const override
-    {
-        // TODO
-        return "";
-    }
-
     std::optional<std::pair<std::string, std::string>> accessHeaderFromToken(const std::string & token) const override
     {
         // Github supports PAT/OAuth2 tokens and HTTP Basic
@@ -473,12 +442,6 @@ struct GitLabInputScheme : GitArchiveInputScheme
         return "gitlab";
     }
 
-    std::string schemeDescription() const override
-    {
-        // TODO
-        return "";
-    }
-
     std::optional<std::pair<std::string, std::string>> accessHeaderFromToken(const std::string & token) const override
     {
         // Gitlab supports 4 kinds of authorization, two of which are
@@ -563,12 +526,6 @@ struct SourceHutInputScheme : GitArchiveInputScheme
         return "sourcehut";
     }
 
-    std::string schemeDescription() const override
-    {
-        // TODO
-        return "";
-    }
-
     std::optional<std::pair<std::string, std::string>> accessHeaderFromToken(const std::string & token) const override
     {
         // SourceHut supports both PAT and OAuth2. See

src/libfetchers/include/nix/fetchers/fetchers.hh

@@ -203,27 +203,14 @@ struct InputScheme
      */
     virtual std::string_view schemeName() const = 0;
 
-    /**
-     * Longform description of this scheme, for documentation purposes.
-     */
-    virtual std::string schemeDescription() const = 0;
-
-    // TODO remove these defaults
-    struct AttributeInfo
-    {
-        const char * type = "String";
-        bool required = true;
-        const char * doc = "";
-    };
-
     /**
      * Allowed attributes in an attribute set that is converted to an
-     * input, and documentation for each attribute.
+     * input.
      *
-     * `type` is not included from this map, because the `type` field is
+     * `type` is not included from this set, because the `type` field is
       parsed first to choose which scheme; `type` is always required.
      */
-    virtual const std::map<std::string, AttributeInfo> & allowedAttrs() const = 0;
+    virtual StringSet allowedAttrs() const = 0;
 
     virtual ParsedURL toURL(const Input & input) const;
 
@@ -276,12 +263,7 @@ struct InputScheme
 
 void registerInputScheme(std::shared_ptr<InputScheme> && fetcher);
 
-using InputSchemeMap = std::map<std::string_view, std::shared_ptr<InputScheme>>;
-
-/**
- * Use this for docs, not for finding a specific scheme
- */
-const InputSchemeMap & getAllInputSchemes();
+nlohmann::json dumpRegisterInputSchemeInfo();
 
 struct PublicKey
 {

src/libfetchers/indirect.cc

@@ -60,33 +60,14 @@ struct IndirectInputScheme : InputScheme
         return "indirect";
     }
 
-    std::string schemeDescription() const override
+    StringSet allowedAttrs() const override
     {
-        // TODO
-        return "";
-    }
-
-    const std::map<std::string, AttributeInfo> & allowedAttrs() const override
-    {
-        static const std::map<std::string, AttributeInfo> attrs = {
-            {
-                "id",
-                {},
-            },
-            {
-                "ref",
-                {},
-            },
-            {
-                "rev",
-                {},
-            },
-            {
-                "narHash",
-                {},
-            },
+        return {
+            "id",
+            "ref",
+            "rev",
+            "narHash",
         };
-        return attrs;
     }
 
     std::optional<Input> inputFromAttrs(const Settings & settings, const Attrs & attrs) const override

src/libfetchers/mercurial.cc

@@ -68,41 +68,16 @@ struct MercurialInputScheme : InputScheme
         return "hg";
     }
 
-    std::string schemeDescription() const override
+    StringSet allowedAttrs() const override
     {
-        // TODO
-        return "";
-    }
-
-    const std::map<std::string, AttributeInfo> & allowedAttrs() const override
-    {
-        static const std::map<std::string, AttributeInfo> attrs = {
-            {
-                "url",
-                {},
-            },
-            {
-                "ref",
-                {},
-            },
-            {
-                "rev",
-                {},
-            },
-            {
-                "revCount",
-                {},
-            },
-            {
-                "narHash",
-                {},
-            },
-            {
-                "name",
-                {},
-            },
+        return {
+            "url",
+            "ref",
+            "rev",
+            "revCount",
+            "narHash",
+            "name",
         };
-        return attrs;
     }
 
     std::optional<Input> inputFromAttrs(const Settings & settings, const Attrs & attrs) const override

src/libfetchers/path.cc

@@ -40,42 +40,20 @@ struct PathInputScheme : InputScheme
         return "path";
     }
 
-    std::string schemeDescription() const override
+    StringSet allowedAttrs() const override
     {
-        // TODO
-        return "";
-    }
-
-    const std::map<std::string, AttributeInfo> & allowedAttrs() const override
-    {
-        static const std::map<std::string, AttributeInfo> attrs = {
-            {
-                "path",
-                {},
-            },
+        return {
+            "path",
             /* Allow the user to pass in "fake" tree info
                attributes. This is useful for making a pinned tree work
                the same as the repository from which is exported (e.g.
                path:/nix/store/...-source?lastModified=1585388205&rev=b0c285...).
              */
-            {
-                "rev",
-                {},
-            },
-            {
-                "revCount",
-                {},
-            },
-            {
-                "lastModified",
-                {},
-            },
-            {
-                "narHash",
-                {},
-            },
+            "rev",
+            "revCount",
+            "lastModified",
+            "narHash",
         };
-        return attrs;
     }
 
     std::optional<Input> inputFromAttrs(const Settings & settings, const Attrs & attrs) const override

src/libfetchers/tarball.cc

@@ -278,7 +278,7 @@ struct CurlInputScheme : InputScheme
            HTTP request. Now that we've processed the Nix-specific
            attributes above, remove them so we don't also send them as
            part of the HTTP request. */
-        for (auto & [param, _] : allowedAttrs())
+        for (auto & param : allowedAttrs())
             url.query.erase(param);
 
         input.attrs.insert_or_assign("type", std::string{schemeName()});
@@ -286,83 +286,18 @@ struct CurlInputScheme : InputScheme
         return input;
     }
 
-    static const std::map<std::string, AttributeInfo> & allowedAttrsImpl()
+    StringSet allowedAttrs() const override
     {
-        static const std::map<std::string, AttributeInfo> attrs = {
-            {
-                "url",
-                {
-                    .type = "String",
-                    .required = true,
-                    .doc = R"(
-                      Supported protocols:
-
-                      - `https`
-
-                        > **Example**
-                        >
-                        > ```nix
-                        > fetchTree {
-                        >   type = "file";
-                        >   url = "https://example.com/index.html";
-                        > }
-                        > ```
-
-                      - `http`
-
-                        Insecure HTTP transfer for legacy sources.
-
-                        > **Warning**
-                        >
-                        > HTTP performs no encryption or authentication.
-                        > Use a `narHash` known in advance to ensure the output has expected contents.
-
-                      - `file`
-
-                        A file on the local file system.
-
-                        > **Example**
-                        >
-                        > ```nix
-                        > fetchTree {
-                        >   type = "file";
-                        >   url = "file:///home/eelco/nix/README.md";
-                        > }
-                        > ```
-                    )",
-                },
-            },
-            {
-                "narHash",
-                {},
-            },
-            {
-                "name",
-                {},
-            },
-            {
-                "unpack",
-                {},
-            },
-            {
-                "rev",
-                {},
-            },
-            {
-                "revCount",
-                {},
-            },
-            {
-                "lastModified",
-                {},
-            },
+        return {
+            "type",
+            "url",
+            "narHash",
+            "name",
+            "unpack",
+            "rev",
+            "revCount",
+            "lastModified",
         };
-        return attrs;
-    }
-
-    const std::map<std::string, AttributeInfo> & allowedAttrs() const override
-    {
-        return allowedAttrsImpl();
     }
 
     std::optional<Input> inputFromAttrs(const Settings & settings, const Attrs & attrs) const override
@@ -397,14 +332,6 @@ struct FileInputScheme : CurlInputScheme
         return "file";
     }
 
-    std::string schemeDescription() const override
-    {
-        return stripIndentation(R"(
-          Place a plain file into the Nix store.
-          This is similar to [`builtins.fetchurl`](@docroot@/language/builtins.md#builtins-fetchurl)
-        )");
-    }
-
     bool isValidURL(const ParsedURL & url, bool requireTree) const override
     {
         auto parsedUrlScheme = parseUrlScheme(url.scheme);
@@ -442,34 +369,6 @@ struct TarballInputScheme : CurlInputScheme
         return "tarball";
     }
 
-    std::string schemeDescription() const override
-    {
-        return stripIndentation(R"(
-          Download a tar archive and extract it into the Nix store.
-          This has the same underlying implementation as [`builtins.fetchTarball`](@docroot@/language/builtins.md#builtins-fetchTarball)
-        )");
-    }
-
-    const std::map<std::string, AttributeInfo> & allowedAttrs() const override
-    {
-        static const std::map<std::string, AttributeInfo> attrs = [] {
-            auto attrs = CurlInputScheme::allowedAttrsImpl();
-            // Override the "url" attribute to add tarball-specific example
-            attrs["url"].doc = R"(
-              > **Example**
-              >
-              > ```nix
-              > fetchTree {
-              >   type = "tarball";
-              >   url = "https://github.com/NixOS/nixpkgs/tarball/nixpkgs-23.11";
-              > }
-              > ```
-            )";
-            return attrs;
-        }();
-        return attrs;
-    }
-
     bool isValidURL(const ParsedURL & url, bool requireTree) const override
     {
         auto parsedUrlScheme = parseUrlScheme(url.scheme);

src/libflake/flake-primops.cc

@@ -19,6 +19,7 @@
 #include "nix/expr/eval-settings.hh"
 #include "nix/expr/symbol-table.hh"
 #include "nix/expr/value.hh"
+#include "nix/expr/string-data-static.hh"
 #include "nix/fetchers/attrs.hh"
 #include "nix/fetchers/fetchers.hh"
 #include "nix/util/configuration.hh"
@@ -62,7 +63,7 @@ PrimOp getFlake(const Settings & settings)
     return PrimOp{
         .name = "__getFlake",
         .args = {"args"},
-        .doc = R"(
+        .doc = &R"(
           Fetch a flake from a flake reference, and return its output attributes and some metadata. For example:
 
           ```nix
@@ -76,7 +77,7 @@ PrimOp getFlake(const Settings & settings)
           ```nix
           (builtins.getFlake "github:edolstra/dwarffs").rev
           ```
-        )",
+        )"_sds,
         .fun = prim_getFlake,
         .experimentalFeature = Xp::Flakes,
     };
@@ -104,7 +105,7 @@ static void prim_parseFlakeRef(EvalState & state, const PosIdx pos, Value ** arg
 nix::PrimOp parseFlakeRef({
     .name = "__parseFlakeRef",
     .args = {"flake-ref"},
-    .doc = R"(
+    .doc = &R"(
       Parse a flake reference, and return its exploded form.
 
       For example:
@@ -118,7 +119,7 @@ nix::PrimOp parseFlakeRef({
       ```nix
       { dir = "lib"; owner = "NixOS"; ref = "23.05"; repo = "nixpkgs"; type = "github"; }
       ```
-    )",
+    )"_sds,
     .fun = prim_parseFlakeRef,
     .experimentalFeature = Xp::Flakes,
 });
@@ -162,7 +163,7 @@ static void prim_flakeRefToString(EvalState & state, const PosIdx pos, Value **
 nix::PrimOp flakeRefToString({
     .name = "__flakeRefToString",
     .args = {"attrs"},
-    .doc = R"(
+    .doc = &R"(
       Convert a flake reference from attribute set format to URL format.
 
       For example:
@@ -178,7 +179,7 @@ nix::PrimOp flakeRefToString({
       ```nix
       "github:NixOS/nixpkgs/23.05?dir=lib"
       ```
-    )",
+    )"_sds,
     .fun = prim_flakeRefToString,
     .experimentalFeature = Xp::Flakes,
 });

src/libutil/include/nix/util/strings-inline.hh

@@ -114,69 +114,4 @@ std::string dropEmptyInitThenConcatStringsSep(const std::string_view sep, const
     return s;
 }
 
-/**
- * Consteval helper for the "constexpr 2-step" pattern.
- * Converts a std::string to a std::array<char, N> for stable compile-time storage.
- *
- * Usage:
- *   static constexpr auto str = stripIndentationConsteval(...);
- *   static constexpr auto arr = toArray<str.size()>(str);
- *   static constexpr std::string_view view{arr.data(), arr.size()};
- */
-template<size_t N>
-consteval std::string_view toArray(const std::string & s)
-{
-    static std::array<char, N> arr{};
-    for (size_t i = 0; i < N; ++i) {
-        arr[i] = s[i];
-    }
-    return {arr.data(), N};
-}
-
-constexpr std::string stripIndentationImpl(std::string_view s)
-{
-    size_t minIndent = 10000;
-    size_t curIndent = 0;
-    bool atStartOfLine = true;
-
-    for (auto & c : s) {
-        if (atStartOfLine && c == ' ')
-            curIndent++;
-        else if (c == '\n') {
-            if (atStartOfLine)
-                minIndent = std::max(minIndent, curIndent);
-            curIndent = 0;
-            atStartOfLine = true;
-        } else {
-            if (atStartOfLine) {
-                minIndent = std::min(minIndent, curIndent);
-                atStartOfLine = false;
-            }
-        }
-    }
-
-    std::string res;
-
-    size_t pos = 0;
-    while (pos < s.size()) {
-        auto eol = s.find('\n', pos);
-        if (eol == s.npos)
-            eol = s.size();
-        if (eol - pos > minIndent)
-            res.append(s.substr(pos + minIndent, eol - pos - minIndent));
-        res.push_back('\n');
-        pos = eol + 1;
-    }
-
-    return res;
-}
-
-template<std::string_view s>
-consteval std::string_view stripIndentationConsteval()
-{
-    constexpr auto stripped = stripIndentationImpl(s);
-    constexpr size_t n = stripped.size();
-    return toArray<n>(stripped);
-}
-
 } // namespace nix

src/libutil/include/nix/util/strings.hh

@@ -132,13 +132,6 @@ std::string optionalBracket(std::string_view prefix, const std::optional<T> & co
     return optionalBracket(prefix, std::string_view(*content), suffix);
 }
 
-/**
- * Remove common leading whitespace from the lines in the string
- * 's'. For example, if every line is indented by at least 3 spaces,
- * then we remove 3 spaces from the start of every line.
- */
-std::string stripIndentation(std::string_view s);
-
 /**
  * Hash implementation that can be used for zero-copy heterogenous lookup from
  * P1690R1[1] in unordered containers.

src/libutil/include/nix/util/util.hh

@@ -228,6 +228,13 @@ constexpr char treeLast[] = "└───";
 constexpr char treeLine[] = "│   ";
 constexpr char treeNull[] = "    ";
 
+/**
+ * Remove common leading whitespace from the lines in the string
+ * 's'. For example, if every line is indented by at least 3 spaces,
+ * then we remove 3 spaces from the start of every line.
+ */
+std::string stripIndentation(std::string_view s);
+
 /**
  * Get the prefix of 's' up to and excluding the next line break (LF
  * optionally preceded by CR), and the remainder following the line

src/libutil/strings.cc

@@ -152,8 +152,4 @@ std::string optionalBracket(std::string_view prefix, std::string_view content, s
     return result;
 }
 
-std::string stripIndentation(std::string_view s) {
-    return stripIndentationImpl(s);
-}
-
 } // namespace nix

src/libutil/util.cc

@@ -250,6 +250,44 @@ void ignoreExceptionExceptInterrupt(Verbosity lvl)
     }
 }
 
+std::string stripIndentation(std::string_view s)
+{
+    size_t minIndent = 10000;
+    size_t curIndent = 0;
+    bool atStartOfLine = true;
+
+    for (auto & c : s) {
+        if (atStartOfLine && c == ' ')
+            curIndent++;
+        else if (c == '\n') {
+            if (atStartOfLine)
+                minIndent = std::max(minIndent, curIndent);
+            curIndent = 0;
+            atStartOfLine = true;
+        } else {
+            if (atStartOfLine) {
+                minIndent = std::min(minIndent, curIndent);
+                atStartOfLine = false;
+            }
+        }
+    }
+
+    std::string res;
+
+    size_t pos = 0;
+    while (pos < s.size()) {
+        auto eol = s.find('\n', pos);
+        if (eol == s.npos)
+            eol = s.size();
+        if (eol - pos > minIndent)
+            res.append(s.substr(pos + minIndent, eol - pos - minIndent));
+        res.push_back('\n');
+        pos = eol + 1;
+    }
+
+    return res;
+}
+
 std::pair<std::string_view, std::string_view> getLine(std::string_view s)
 {
     auto newline = s.find('\n');

src/nix/main.cc

@@ -193,38 +193,20 @@ struct NixArgs : virtual MultiCommand, virtual MixCommonArgs, virtual RootArgs
 
     std::string dumpCli()
     {
-        using nlohmann::json;
-
-        auto res = json::object();
+        auto res = nlohmann::json::object();
 
         res["args"] = toJSON();
 
-        {
-            auto & stores = res["stores"] = json::object();
-            for (auto & [storeName, implem] : Implementations::registered()) {
-                auto & j = stores[storeName];
-                j["doc"] = implem.doc;
-                j["uri-schemes"] = implem.uriSchemes;
-                j["settings"] = implem.getConfig()->toJSON();
-                j["experimentalFeature"] = implem.experimentalFeature;
-            }
+        auto stores = nlohmann::json::object();
+        for (auto & [storeName, implem] : Implementations::registered()) {
+            auto & j = stores[storeName];
+            j["doc"] = implem.doc;
+            j["uri-schemes"] = implem.uriSchemes;
+            j["settings"] = implem.getConfig()->toJSON();
+            j["experimentalFeature"] = implem.experimentalFeature;
         }
-
-        {
-            auto & fetchers = res["fetchers"] = json::object();
-
-            for (const auto & [schemeName, scheme] : fetchers::getAllInputSchemes()) {
-                auto & s = fetchers[schemeName] = json::object();
-                s["description"] = scheme->schemeDescription();
-                auto & attrs = s["allowedAttrs"] = json::object();
-                for (auto & [fieldName, field] : scheme->allowedAttrs()) {
-                    auto & f = attrs[fieldName] = json::object();
-                    f["type"] = field.type;
-                    f["required"] = field.required;
-                    f["doc"] = stripIndentation(field.doc);
-                }
-            }
-        };
+        res["stores"] = std::move(stores);
+        res["fetchers"] = fetchers::dumpRegisterInputSchemeInfo();
 
         return res.dump();
     }
@@ -458,7 +440,7 @@ void mainWrapped(int argc, char ** argv)
             if (!primOp->doc)
                 continue;
             b["args"] = primOp->args;
-            b["doc"] = trim(stripIndentation(*primOp->doc));
+            b["doc"] = trim(stripIndentation(primOp->doc->view()));
             if (primOp->experimentalFeature)
                 b["experimental-feature"] = primOp->experimentalFeature;
             builtinsJson.emplace(state.symbols[builtin.name], std::move(b));
@@ -467,7 +449,7 @@ void mainWrapped(int argc, char ** argv)
             auto b = nlohmann::json::object();
             if (!info.doc)
                 continue;
-            b["doc"] = trim(stripIndentation(info.doc));
+            b["doc"] = trim(stripIndentation(info.doc->view()));
             b["type"] = showType(info.type, false);
             if (info.impureOnly)
                 b["impure-only"] = true;

src/nix/nix-env/user-env.cc

@@ -9,7 +9,7 @@
 #include "nix/expr/eval-inline.hh"
 #include "nix/store/profiles.hh"
 #include "nix/expr/print-ambiguous.hh"
-#include "nix/expr/static-string-data.hh"
+#include "nix/expr/string-data-static.hh"
 
 #include <limits>
 #include <sstream>