ValidPathInfo: make ca field a proper datatype

src/libexpr/get-drvs.cc

@@ -1,7 +1,7 @@
 #include "get-drvs.hh"
 #include "util.hh"
 #include "eval-inline.hh"
-#include "derivations.hh"
+#include "store-api.hh"
 
 #include <cstring>
 #include <regex>

src/libexpr/primops.cc

@@ -769,15 +769,14 @@ static void prim_derivationStrict(EvalState & state, const Pos & pos, Value * *
                 .nixCode = NixCode { .errPos = posDrvName }
             });
 
-        HashType ht = outputHashAlgo.empty() ? htUnknown : parseHashType(outputHashAlgo);
-
+        std::optional<HashType> ht = parseHashTypeOpt(outputHashAlgo);
         Hash h = newHashAllowEmpty(*outputHash, ht);
 
         auto outPath = state.store->makeFixedOutputPath(ingestionMethod, h, drvName);
         if (!jsonObject) drv.env["out"] = state.store->printStorePath(outPath);
         drv.outputs.insert_or_assign("out", DerivationOutput {
             .path = std::move(outPath),
-            .hash = DerivationOutputHash {
+            .hash = FixedOutputHash {
                 .method = ingestionMethod,
                 .hash = std::move(h),
             },
@@ -796,7 +795,7 @@ static void prim_derivationStrict(EvalState & state, const Pos & pos, Value * *
             drv.outputs.insert_or_assign(i,
                 DerivationOutput {
                     .path = StorePath::dummy,
-                    .hash = std::optional<DerivationOutputHash> {},
+                    .hash = std::optional<FixedOutputHash> {},
                 });
         }
 
@@ -808,7 +807,7 @@ static void prim_derivationStrict(EvalState & state, const Pos & pos, Value * *
             drv.outputs.insert_or_assign(i,
                 DerivationOutput {
                     .path = std::move(outPath),
-                    .hash = std::optional<DerivationOutputHash>(),
+                    .hash = std::optional<FixedOutputHash>(),
                 });
         }
     }
@@ -1006,8 +1005,8 @@ static void prim_findFile(EvalState & state, const Pos & pos, Value * * args, Va
 static void prim_hashFile(EvalState & state, const Pos & pos, Value * * args, Value & v)
 {
     string type = state.forceStringNoCtx(*args[0], pos);
-    HashType ht = parseHashType(type);
-    if (ht == htUnknown)
+    std::optional<HashType> ht = parseHashType(type);
+    if (!ht)
       throw Error({
           .hint = hintfmt("unknown hash type '%1%'", type),
           .nixCode = NixCode { .errPos = pos }
@@ -1016,7 +1015,7 @@ static void prim_hashFile(EvalState & state, const Pos & pos, Value * * args, Va
     PathSet context; // discarded
     Path p = state.coerceToPath(pos, *args[1], context);
 
-    mkString(v, hashFile(ht, state.checkSourcePath(p)).to_string(Base16, false), context);
+    mkString(v, hashFile(*ht, state.checkSourcePath(p)).to_string(Base16, false), context);
 }
 
 /* Read a directory (without . or ..) */
@@ -1943,8 +1942,8 @@ static void prim_stringLength(EvalState & state, const Pos & pos, Value * * args
 static void prim_hashString(EvalState & state, const Pos & pos, Value * * args, Value & v)
 {
     string type = state.forceStringNoCtx(*args[0], pos);
-    HashType ht = parseHashType(type);
-    if (ht == htUnknown)
+    std::optional<HashType> ht = parseHashType(type);
+    if (!ht)
         throw Error({
             .hint = hintfmt("unknown hash type '%1%'", type),
             .nixCode = NixCode { .errPos = pos }
@@ -1953,7 +1952,7 @@ static void prim_hashString(EvalState & state, const Pos & pos, Value * * args,
     PathSet context; // discarded
     string s = state.forceString(*args[1], context, pos);
 
-    mkString(v, hashString(ht, s).to_string(Base16, false), context);
+    mkString(v, hashString(*ht, s).to_string(Base16, false), context);
 }
 
 

src/libexpr/primops/context.cc

@@ -1,6 +1,6 @@
 #include "primops.hh"
 #include "eval-inline.hh"
-#include "derivations.hh"
+#include "store-api.hh"
 
 namespace nix {
 

src/libfetchers/fetchers.cc

@@ -36,7 +36,7 @@ std::unique_ptr<Input> inputFromAttrs(const Attrs & attrs)
         if (res) {
             if (auto narHash = maybeGetStrAttr(attrs, "narHash"))
                 // FIXME: require SRI hash.
-                res->narHash = newHashAllowEmpty(*narHash, htUnknown);
+                res->narHash = newHashAllowEmpty(*narHash, {});
             return res;
         }
     }

src/libfetchers/tarball.cc

@@ -70,7 +70,10 @@ DownloadFileResult downloadFile(
         ValidPathInfo info(store->makeFixedOutputPath(FileIngestionMethod::Flat, hash, name));
         info.narHash = hashString(htSHA256, *sink.s);
         info.narSize = sink.s->size();
-        info.ca = makeFixedOutputCA(FileIngestionMethod::Flat, hash);
+        info.ca = FixedOutputHash {
+            FileIngestionMethod::Flat,
+            hash,
+        };
         auto source = StringSource { *sink.s };
         store->addToStore(info, source, NoRepair, NoCheckSigs);
         storePath = std::move(info.path);
@@ -264,7 +267,7 @@ struct TarballInputScheme : InputScheme
 
         auto input = std::make_unique<TarballInput>(parseURL(getStrAttr(attrs, "url")));
         if (auto hash = maybeGetStrAttr(attrs, "hash"))
-            input->hash = newHashAllowEmpty(*hash, htUnknown);
+            input->hash = newHashAllowEmpty(*hash, {});
 
         return input;
     }

src/libstore/build.cc

@@ -3714,7 +3714,7 @@ void DerivationGoal::registerOutputs()
         /* Check that fixed-output derivations produced the right
            outputs (i.e., the content hash should match the specified
            hash). */
-        std::string ca;
+        std::optional<ContentAddress> ca;
 
         if (fixedOutput) {
 
@@ -3730,8 +3730,8 @@ void DerivationGoal::registerOutputs()
             /* Check the hash. In hash mode, move the path produced by
                the derivation to its content-addressed location. */
             Hash h2 = i.second.hash->method == FileIngestionMethod::Recursive
-                ? hashPath(i.second.hash->hash.type, actualPath).first
-                : hashFile(i.second.hash->hash.type, actualPath);
+                ? hashPath(*i.second.hash->hash.type, actualPath).first
+                : hashFile(*i.second.hash->hash.type, actualPath);
 
             auto dest = worker.store.makeFixedOutputPath(i.second.hash->method, h2, i.second.path.name());
 
@@ -3764,7 +3764,7 @@ void DerivationGoal::registerOutputs()
             else
                 assert(worker.store.parseStorePath(path) == dest);
 
-            ca = makeFixedOutputCA(i.second.hash->method, h2);
+            ca = FixedOutputHash { i.second.hash->method, h2 };
         }
 
         /* Get rid of all weird permissions.  This also checks that
@@ -3837,7 +3837,10 @@ void DerivationGoal::registerOutputs()
         info.ca = ca;
         worker.store.signPathInfo(info);
 
-        if (!info.references.empty()) info.ca.clear();
+        if (!info.references.empty()) {
+            // FIXME don't we have an experimental feature for fixed output with references?
+            info.ca = {};
+        }
 
         infos.emplace(i.first, std::move(info));
     }
@@ -4997,7 +5000,7 @@ bool Worker::pathContentsGood(const StorePath & path)
     if (!pathExists(store.printStorePath(path)))
         res = false;
     else {
-        HashResult current = hashPath(info->narHash.type, store.printStorePath(path));
+        HashResult current = hashPath(*info->narHash.type, store.printStorePath(path));
         Hash nullHash(htSHA256);
         res = info->narHash == nullHash || info->narHash == current.first;
     }

src/libstore/builtins/fetchurl.cc

@@ -63,9 +63,9 @@ void builtinFetchurl(const BasicDerivation & drv, const std::string & netrcData)
         for (auto hashedMirror : settings.hashedMirrors.get())
             try {
                 if (!hasSuffix(hashedMirror, "/")) hashedMirror += '/';
-                auto ht = parseHashType(getAttr("outputHashAlgo"));
+                auto ht = parseHashTypeOpt(getAttr("outputHashAlgo"));
                 auto h = Hash(getAttr("outputHash"), ht);
-                fetch(hashedMirror + printHashType(h.type) + "/" + h.to_string(Base16, false));
+                fetch(hashedMirror + printHashType(*h.type) + "/" + h.to_string(Base16, false));
                 return;
             } catch (Error & e) {
                 debug(e.what());

src/libstore/content-address.cc

@@ -0,0 +1,79 @@
+#include "content-address.hh"
+
+namespace nix {
+
+std::string FixedOutputHash::printMethodAlgo() const {
+    return makeFileIngestionPrefix(method) + printHashType(*hash.type);
+}
+
+std::string makeFileIngestionPrefix(const FileIngestionMethod m) {
+    switch (m) {
+    case FileIngestionMethod::Flat:
+        return "";
+    case FileIngestionMethod::Recursive:
+        return "r:";
+    default:
+        throw Error("impossible, caught both cases");
+    }
+}
+
+std::string makeFixedOutputCA(FileIngestionMethod method, const Hash & hash)
+{
+    return "fixed:"
+        + makeFileIngestionPrefix(method)
+        + hash.to_string(Base32, true);
+}
+
+// FIXME Put this somewhere?
+template<class... Ts> struct overloaded : Ts... { using Ts::operator()...; };
+template<class... Ts> overloaded(Ts...) -> overloaded<Ts...>;
+
+std::string renderContentAddress(ContentAddress ca) {
+    return std::visit(overloaded {
+        [](TextHash th) {
+            return "text:" + th.hash.to_string(Base32, true);
+        },
+        [](FixedOutputHash fsh) {
+            return makeFixedOutputCA(fsh.method, fsh.hash);
+        }
+    }, ca);
+}
+
+ContentAddress parseContentAddress(std::string_view rawCa) {
+    auto prefixSeparator = rawCa.find(':');
+    if (prefixSeparator != string::npos) {
+        auto prefix = string(rawCa, 0, prefixSeparator);
+        if (prefix == "text") {
+            auto hashTypeAndHash = rawCa.substr(prefixSeparator+1, string::npos);
+            Hash hash = Hash(string(hashTypeAndHash));
+            if (*hash.type != htSHA256) {
+                throw Error("parseContentAddress: the text hash should have type SHA256");
+            }
+            return TextHash { hash };
+        } else if (prefix == "fixed") {
+            // This has to be an inverse of makeFixedOutputCA
+            auto methodAndHash = rawCa.substr(prefixSeparator+1, string::npos);
+            if (methodAndHash.substr(0,2) == "r:") {
+                std::string_view hashRaw = methodAndHash.substr(2,string::npos);
+                return FixedOutputHash { FileIngestionMethod::Recursive, Hash(string(hashRaw)) };
+            } else {
+                std::string_view hashRaw = methodAndHash;
+                return FixedOutputHash { FileIngestionMethod::Flat, Hash(string(hashRaw)) };
+            }
+        } else {
+            throw Error("parseContentAddress: format not recognized; has to be text or fixed");
+        }
+    } else {
+        throw Error("Not a content address because it lacks an appropriate prefix");
+    }
+};
+
+std::optional<ContentAddress> parseContentAddressOpt(std::string_view rawCaOpt) {
+    return rawCaOpt == "" ? std::optional<ContentAddress> {} : parseContentAddress(rawCaOpt);
+};
+
+std::string renderContentAddress(std::optional<ContentAddress> ca) {
+    return ca ? renderContentAddress(*ca) : "";
+}
+
+}

src/libstore/content-address.hh

@@ -0,0 +1,56 @@
+#pragma once
+
+#include <variant>
+#include "hash.hh"
+
+namespace nix {
+
+enum struct FileIngestionMethod : uint8_t {
+    Flat = false,
+    Recursive = true
+};
+
+struct TextHash {
+    Hash hash;
+};
+
+/// Pair of a hash, and how the file system was ingested
+struct FixedOutputHash {
+    FileIngestionMethod method;
+    Hash hash;
+    std::string printMethodAlgo() const;
+};
+
+/*
+  We've accumulated several types of content-addressed paths over the years;
+  fixed-output derivations support multiple hash algorithms and serialisation
+  methods (flat file vs NAR). Thus, ‘ca’ has one of the following forms:
+
+  * ‘text:sha256:<sha256 hash of file contents>’: For paths
+    computed by makeTextPath() / addTextToStore().
+
+  * ‘fixed:<r?>:<ht>:<h>’: For paths computed by
+    makeFixedOutputPath() / addToStore().
+*/
+typedef std::variant<
+    TextHash, // for paths computed by makeTextPath() / addTextToStore
+    FixedOutputHash // for path computed by makeFixedOutputPath
+> ContentAddress;
+
+/* Compute the prefix to the hash algorithm which indicates how the files were
+   ingested. */
+std::string makeFileIngestionPrefix(const FileIngestionMethod m);
+
+/* Compute the content-addressability assertion (ValidPathInfo::ca)
+   for paths created by makeFixedOutputPath() / addToStore(). */
+std::string makeFixedOutputCA(FileIngestionMethod method, const Hash & hash);
+
+std::string renderContentAddress(ContentAddress ca);
+
+std::string renderContentAddress(std::optional<ContentAddress> ca);
+
+ContentAddress parseContentAddress(std::string_view rawCa);
+
+std::optional<ContentAddress> parseContentAddressOpt(std::string_view rawCaOpt);
+
+}

src/libstore/daemon.cc

@@ -652,7 +652,7 @@ static void performOp(TunnelLogger * logger, ref<Store> store,
             if (GET_PROTOCOL_MINOR(clientVersion) >= 16) {
                 to << info->ultimate
                    << info->sigs
-                   << info->ca;
+                   << renderContentAddress(info->ca);
             }
         } else {
             assert(GET_PROTOCOL_MINOR(clientVersion) >= 17);
@@ -710,7 +710,8 @@ static void performOp(TunnelLogger * logger, ref<Store> store,
         info.references = readStorePaths<StorePathSet>(*store, from);
         from >> info.registrationTime >> info.narSize >> info.ultimate;
         info.sigs = readStrings<StringSet>(from);
-        from >> info.ca >> repair >> dontCheckSigs;
+        info.ca = parseContentAddressOpt(readString(from));
+        from >> repair >> dontCheckSigs;
         if (!trusted && dontCheckSigs)
             dontCheckSigs = false;
         if (!trusted)

src/libstore/derivations.cc

@@ -8,11 +8,6 @@
 
 namespace nix {
 
-std::string DerivationOutputHash::printMethodAlgo() const {
-    return makeFileIngestionPrefix(method) + printHashType(hash.type);
-}
-
-
 const StorePath & BasicDerivation::findOutput(const string & id) const
 {
     auto i = outputs.find(id);
@@ -113,17 +108,15 @@ static DerivationOutput parseDerivationOutput(const Store & store, istringstream
     expect(str, ","); const auto hash = parseString(str);
     expect(str, ")");
 
-    std::optional<DerivationOutputHash> fsh;
+    std::optional<FixedOutputHash> fsh;
     if (hashAlgo != "") {
         auto method = FileIngestionMethod::Flat;
         if (string(hashAlgo, 0, 2) == "r:") {
             method = FileIngestionMethod::Recursive;
             hashAlgo = string(hashAlgo, 2);
         }
         const HashType hashType = parseHashType(hashAlgo);
-        if (hashType == htUnknown)
-            throw Error("unknown hash hashAlgorithm '%s'", hashAlgo);
-        fsh = DerivationOutputHash {
+        fsh = FixedOutputHash {
             .method = std::move(method),
             .hash = Hash(hash, hashType),
         };
@@ -413,17 +406,15 @@ static DerivationOutput readDerivationOutput(Source & in, const Store & store)
     auto hashAlgo = readString(in);
     const auto hash = readString(in);
 
-    std::optional<DerivationOutputHash> fsh;
+    std::optional<FixedOutputHash> fsh;
     if (hashAlgo != "") {
         auto method = FileIngestionMethod::Flat;
         if (string(hashAlgo, 0, 2) == "r:") {
             method = FileIngestionMethod::Recursive;
             hashAlgo = string(hashAlgo, 2);
         }
         const HashType hashType = parseHashType(hashAlgo);
-        if (hashType == htUnknown)
-            throw Error("unknown hash hashAlgorithm '%s'", hashAlgo);
-        fsh = DerivationOutputHash {
+        fsh = FixedOutputHash {
             .method = std::move(method),
             .hash = Hash(hash, hashType),
         };