From 331bf3e2613d2b1f730f3867dd32bd0dc1ca43a6 Mon Sep 17 00:00:00 2001
From: Eelco Dolstra <edolstra@gmail.com>
Date: Fri, 29 Nov 2024 16:55:27 +0100
Subject: [PATCH] Git fetcher: Calculate a fingerprint for dirty workdirs

This restores evaluation caching for dirty Git workdirs.
---
 src/libfetchers/git-utils.cc | 18 ++++++++++++++++--
 src/libfetchers/git-utils.hh |  7 ++++++-
 src/libfetchers/git.cc       | 30 +++++++++++++++++++++++++++---
 3 files changed, 49 insertions(+), 6 deletions(-)

diff --git a/src/libfetchers/git-utils.cc b/src/libfetchers/git-utils.cc
index 74e68fe12..bd5786857 100644
--- a/src/libfetchers/git-utils.cc
+++ b/src/libfetchers/git-utils.cc
@@ -437,7 +437,12 @@ struct GitRepoImpl : GitRepo, std::enable_shared_from_this<GitRepoImpl>
         {
             if (!(statusFlags & GIT_STATUS_INDEX_DELETED) &&
                 !(statusFlags & GIT_STATUS_WT_DELETED))
-                info.files.insert(CanonPath(path));
+                info.files.emplace(CanonPath(path),
+                    statusFlags == GIT_STATUS_CURRENT
+                    ? WorkdirInfo::State::Clean
+                    : WorkdirInfo::State::Dirty);
+            else
+                info.deletedFiles.insert(CanonPath(path));
             if (statusFlags != GIT_STATUS_CURRENT)
                 info.isDirty = true;
             return 0;
@@ -1202,6 +1207,15 @@ ref<SourceAccessor> GitRepoImpl::getAccessor(const Hash & rev, bool exportIgnore
     }
 }
 
+template<typename K, typename V>
+std::set<K> getKeys(const std::map<K, V> & c)
+{
+    std::set<K> res;
+    for (auto & i : c)
+        res.insert(i.first);
+    return res;
+}
+
 ref<SourceAccessor> GitRepoImpl::getAccessor(const WorkdirInfo & wd, bool exportIgnore, MakeNotAllowedError makeNotAllowedError)
 {
     auto self = ref<GitRepoImpl>(shared_from_this());
@@ -1214,7 +1228,7 @@ ref<SourceAccessor> GitRepoImpl::getAccessor(const WorkdirInfo & wd, bool export
         ? makeEmptySourceAccessor()
         : AllowListSourceAccessor::create(
             makeFSSourceAccessor(path),
-            std::set<CanonPath> { wd.files },
+            std::set<CanonPath> { getKeys(wd.files) },
             std::move(makeNotAllowedError)).cast<SourceAccessor>();
     if (exportIgnore)
         return make_ref<GitExportIgnoreSourceAccessor>(self, fileAccessor, std::nullopt);
diff --git a/src/libfetchers/git-utils.hh b/src/libfetchers/git-utils.hh
index f45b5a504..12cee5db1 100644
--- a/src/libfetchers/git-utils.hh
+++ b/src/libfetchers/git-utils.hh
@@ -55,9 +55,14 @@ struct GitRepo
            in the repo yet. */
         std::optional<Hash> headRev;
 
+        enum State { Clean, Dirty };
+
         /* All files in the working directory that are unchanged,
            modified or added, but excluding deleted files. */
-        std::set<CanonPath> files;
+        std::map<CanonPath, State> files;
+
+        /* The deleted files. */
+        std::set<CanonPath> deletedFiles;
 
         /* The submodules listed in .gitmodules of this workdir. */
         std::vector<Submodule> submodules;
diff --git a/src/libfetchers/git.cc b/src/libfetchers/git.cc
index a6883a2d3..eec134980 100644
--- a/src/libfetchers/git.cc
+++ b/src/libfetchers/git.cc
@@ -685,7 +685,7 @@ struct GitInputScheme : InputScheme
         if (getSubmodulesAttr(input))
             /* Create mountpoints for the submodules. */
             for (auto & submodule : repoInfo.workdirInfo.submodules)
-                repoInfo.workdirInfo.files.insert(submodule.path);
+                repoInfo.workdirInfo.files.emplace(submodule.path, GitRepo::WorkdirInfo::State::Clean);
 
         auto repo = GitRepo::openRepo(repoInfo.url, false, false);
 
@@ -793,10 +793,34 @@ struct GitInputScheme : InputScheme
 
     std::optional<std::string> getFingerprint(ref<Store> store, const Input & input) const override
     {
+        auto makeFingerprint = [&](const Hash & rev)
+        {
+            return rev.gitRev() + (getSubmodulesAttr(input) ? ";s" : "") + (getExportIgnoreAttr(input) ? ";e" : "");
+        };
+
         if (auto rev = input.getRev())
-            return rev->gitRev() + (getSubmodulesAttr(input) ? ";s" : "") + (getExportIgnoreAttr(input) ? ";e" : "");
-        else
+            return makeFingerprint(*rev);
+        else {
+            auto repoInfo = getRepoInfo(input);
+            if (repoInfo.isLocal && repoInfo.workdirInfo.headRev) {
+                /* Calculate a fingerprint that takes into account the
+                   deleted and modified/added files. */
+                HashSink hashSink{HashAlgorithm::SHA512};
+                for (auto & file : repoInfo.workdirInfo.files)
+                    if (file.second == GitRepo::WorkdirInfo::State::Dirty) {
+                        writeString("modified:", hashSink);
+                        writeString(file.first.abs(), hashSink);
+                        readFile(std::filesystem::path(repoInfo.url) + file.first.abs(), hashSink);
+                    }
+                for (auto & file : repoInfo.workdirInfo.deletedFiles) {
+                    writeString("deleted:", hashSink);
+                    writeString(file.abs(), hashSink);
+                }
+                return makeFingerprint(*repoInfo.workdirInfo.headRev)
+                    + ";d=" + hashSink.finish().first.to_string(HashFormat::Base16, false);
+            }
             return std::nullopt;
+        }
     }
 
     bool isLocked(const Input & input) const override