RepositoryConfig: Ensure consistency in reading blobs and trees

...with respect to rejecting invalid entries such as upwards
symlinks. Also ensure that valid trees are only checked once by
remebering known valid tress though marker files in local storage.

2 files changed, 18 insertions, 10 deletions

diff --git a/src/buildtool/execution_api/git/git_api.cpp b/src/buildtool/execution_api/git/git_api.cpp
index 231e717a..2c083e64 100644
--- a/src/buildtool/execution_api/git/git_api.cpp
+++ b/src/buildtool/execution_api/git/git_api.cpp
@@ -81,7 +81,8 @@ auto GitApi::RetrieveToPaths(
             }
         }
         else {
-            auto blob = repo_config_.ReadBlobFromGitCAS(info.digest.hash());
+            auto blob = repo_config_.ReadBlobFromGitCAS(
+                info.digest.hash(), /*is_symlink=*/IsSymlinkObject(info.type));
             if (not blob) {
                 return false;
             }
@@ -141,7 +142,8 @@ auto GitApi::RetrieveToFds(
             }
         }
         else {
-            auto blob = repo_config_.ReadBlobFromGitCAS(info.digest.hash());
+            auto blob = repo_config_.ReadBlobFromGitCAS(
+                info.digest.hash(), /*is_symlink=*/IsSymlinkObject(info.type));
             if (not blob) {
                 Logger::Log(LogLevel::Debug,
                             "Blob {} not known to git",
@@ -217,7 +219,9 @@ auto GitApi::RetrieveToCas(
             content = tree->RawData();
         }
         else {
-            content = repo_config_.ReadBlobFromGitCAS(info->digest.hash());
+            content = repo_config_.ReadBlobFromGitCAS(
+                info->digest.hash(),
+                /*is_symlink=*/IsSymlinkObject(info->type));
         }
         if (not content) {
             return false;
@@ -248,10 +252,14 @@ auto GitApi::RetrieveToCas(
 
 auto GitApi::RetrieveToMemory(Artifact::ObjectInfo const& artifact_info)
     const noexcept -> std::optional<std::string> {
-    return repo_config_.ReadBlobFromGitCAS(artifact_info.digest.hash());
+    return repo_config_.ReadBlobFromGitCAS(
+        artifact_info.digest.hash(),
+        /*is_symlink=*/IsSymlinkObject(artifact_info.type));
 }
 
 auto GitApi::IsAvailable(ArtifactDigest const& digest) const noexcept -> bool {
-    return repo_config_.ReadBlobFromGitCAS(digest.hash(), LogLevel::Trace)
-        .has_value();
+    return repo_config_
+        .ReadBlobFromGitCAS(
+            digest.hash(), /*is_symlink=*/false, LogLevel::Trace)
+        .has_value();  // allow invalid symlink if just to check existence
 }
diff --git a/src/buildtool/execution_api/utils/rehash_utils.cpp b/src/buildtool/execution_api/utils/rehash_utils.cpp
index 9e64a546..66d3a57f 100644
--- a/src/buildtool/execution_api/utils/rehash_utils.cpp
+++ b/src/buildtool/execution_api/utils/rehash_utils.cpp
@@ -285,10 +285,10 @@ auto RehashGitDigest(std::vector<Artifact::ObjectInfo> const& digests,
                      StorageConfig const& target_config,
                      RepositoryConfig const& repo_config)
     -> expected<std::vector<Artifact::ObjectInfo>, std::string> {
-    auto read = [&repo_config](
-                    ArtifactDigest const& digest,
-                    ObjectType /*type*/) -> std::optional<std::string> {
-        return repo_config.ReadBlobFromGitCAS(digest.hash());
+    auto read = [&repo_config](ArtifactDigest const& digest,
+                               ObjectType type) -> std::optional<std::string> {
+        return repo_config.ReadBlobFromGitCAS(
+            digest.hash(), /*is_symlink=*/IsSymlinkObject(type));
     };
     return RehashDigestImpl(digests,
                             source_config,