Merge remote-tracking branch 'upstream/main' into basic-sql

.github/PULL_REQUEST_TEMPLATE.md

@@ -1,7 +1,9 @@
 <!--
-Set the PR title to a meaningful commit message in imperative form. E.g.:
-
-clp-s: Don't add implicit wildcards ('*') at the beginning and the end of a query (fixes #390).
+Set the PR title to a meaningful commit message that:
+- follows the Conventional Commits specification (https://www.conventionalcommits.org).
+- is in imperative form.
+Example:
+fix: Don't add implicit wildcards ('*') at the beginning and the end of a query (fixes #390).
 -->
 
 # Description

.github/workflows/clp-core-build-macos.yaml

@@ -27,6 +27,9 @@ on:
       - "deps-tasks.yml"
       - "Taskfile.yml"
       - "tools/scripts/deps-download/**"
+  schedule:
+    # Run daily at 00:15 UTC (the 15 is to avoid periods of high load)
+    - cron: "15 0 * * *"
   workflow_dispatch:
 
 concurrency:

.github/workflows/clp-core-build.yaml

@@ -23,6 +23,9 @@ on:
       - "Taskfile.yml"
       - "tools/scripts/deps-download/**"
       - "!components/core/tools/scripts/lib_install/macos/**"
+  schedule:
+    # Run daily at 00:15 UTC (the 15 is to avoid periods of high load)
+    - cron: "15 0 * * *"
   workflow_dispatch:
 
 env:

.github/workflows/clp-docs.yaml

@@ -3,6 +3,9 @@ name: "clp-docs"
 on:
   pull_request:
   push:
+  schedule:
+    # Run daily at 00:15 UTC (the 15 is to avoid periods of high load)
+    - cron: "15 0 * * *"
   workflow_dispatch:
 
 concurrency:

.github/workflows/clp-execution-image-build.yaml

@@ -11,6 +11,9 @@ on:
       - ".github/actions/clp-execution-image-build/action.yaml"
       - ".github/workflows/clp-execution-image-build.yaml"
       - "tools/docker-images/**/*"
+  schedule:
+    # Run daily at 00:15 UTC (the 15 is to avoid periods of high load)
+    - cron: "15 0 * * *"
   workflow_dispatch:
 
 concurrency:

.github/workflows/clp-lint.yaml

@@ -4,7 +4,7 @@ on:
   pull_request:
   push:
   schedule:
-    # Run at midnight UTC every day with 15 minutes delay added to avoid high load periods
+    # Run daily at 00:15 UTC (the 15 is to avoid periods of high load)
     - cron: "15 0 * * *"
   workflow_dispatch:
 

.github/workflows/clp-pr-title-checks.yaml

@@ -0,0 +1,23 @@
+name: "clp-pr-title-checks"
+
+on:
+  pull_request_target:
+    types: ["edited", "opened", "reopened"]
+    branches: ["main"]
+
+concurrency:
+  group: "${{github.workflow}}-${{github.ref}}"
+
+  # Cancel in-progress jobs for efficiency
+  cancel-in-progress: true
+
+jobs:
+  conventional-commits:
+    permissions:
+      # For amannn/action-semantic-pull-request
+      pull-requests: "read"
+    runs-on: "ubuntu-latest"
+    steps:
+      - uses: "amannn/action-semantic-pull-request@v5"
+        env:
+          GITHUB_TOKEN: "${{secrets.GITHUB_TOKEN}}"

Taskfile.yml

@@ -228,10 +228,10 @@ tasks:
       - "client/src/**/*.css"
       - "client/src/**/*.jsx"
       - "client/src/webpack.config.js"
-      - "yscope-log-viewer/.babelrc"
-      - "yscope-log-viewer/customized-packages/**/*"
       - "yscope-log-viewer/package.json"
+      - "yscope-log-viewer/public/**/*"
       - "yscope-log-viewer/src/**/*"
+      - "yscope-log-viewer/tsconfig.json"
       - "yscope-log-viewer/webpack.common.js"
       - "yscope-log-viewer/webpack.prod.js"
     dir: "components/log-viewer-webui"
@@ -348,7 +348,7 @@ tasks:
       - "mkdir -p '{{.OUTPUT_TMP_DIR}}'"
       - >-
         curl --fail --location --show-error
-        "{{.URL_PREFIX}}/{{.TAR_NAME}}"
+        "{{trimSuffix "/" .URL_PREFIX}}/{{.TAR_NAME}}"
         --output "{{.TAR_PATH}}"
       - "tar xf '{{.TAR_PATH}}' --directory '{{.OUTPUT_TMP_DIR}}'"
       - "mv '{{.EXTRACTED_DIR}}' '{{.OUTPUT_DIR}}'"

components/core/cmake/Modules/FindMariaDBClient.cmake

@@ -20,6 +20,28 @@ include(cmake/Modules/FindLibraryDependencies.cmake)
 find_package(PkgConfig)
 pkg_check_modules(mariadbclient_PKGCONF QUIET "lib${mariadbclient_LIBNAME}")
 
+if(NOT mariadbclient_PKGCONF_FOUND AND APPLE)
+    execute_process(
+        COMMAND brew --prefix mariadb-connector-c
+        RESULT_VARIABLE mariadbclient_BREW_RESULT
+        OUTPUT_VARIABLE mariadbclient_MACOS_PREFIX
+    )
+    if(NOT mariadbclient_BREW_RESULT EQUAL 0)
+        message(
+            FATAL_ERROR
+            "pkg-config cannot find ${mariadbclient_LIBNAME} and mariadb-connector-c isn't"
+            " installed via Homebrew"
+        )
+    endif()
+    string(STRIP "${mariadbclient_MACOS_PREFIX}" mariadbclient_MACOS_PREFIX)
+    list(PREPEND CMAKE_PREFIX_PATH ${mariadbclient_MACOS_PREFIX})
+    pkg_check_modules(mariadbclient_PKGCONF QUIET "lib${mariadbclient_LIBNAME}")
+endif()
+
+if(NOT mariadbclient_PKGCONF_FOUND)
+    message(FATAL_ERROR "pkg-config cannot find ${mariadbclient_LIBNAME}")
+endif()
+
 # Set include directory
 find_path(MariaDBClient_INCLUDE_DIR mysql.h
         HINTS ${mariadbclient_PKGCONF_INCLUDEDIR}

components/core/src/clp/ffi/KeyValuePairLogEvent.cpp

@@ -153,20 +153,6 @@ node_type_matches_value_type(SchemaTree::Node::Type type, Value const& value) ->
         KeyValuePairLogEvent::NodeIdValuePairs const& node_id_value_pairs
 ) -> bool;
 
-/**
- * @param node_id_value_pairs
- * @param schema_tree
- * @return A result containing a bitmap where every bit corresponds to the ID of a node in the
- * schema tree, and the set bits correspond to the nodes in the subtree defined by all paths from
- * the root node to the nodes in `node_id_value_pairs`; or an error code indicating a failure:
- * - std::errc::result_out_of_range if a node ID in `node_id_value_pairs` doesn't exist in the
- *   schema tree.
- */
-[[nodiscard]] auto get_schema_subtree_bitmap(
-        KeyValuePairLogEvent::NodeIdValuePairs const& node_id_value_pairs,
-        SchemaTree const& schema_tree
-) -> OUTCOME_V2_NAMESPACE::std_result<vector<bool>>;
-
 /**
  * Inserts the given key-value pair into the JSON object (map).
  * @param node The schema tree node of the key to insert.
@@ -283,38 +269,6 @@ auto is_leaf_node(
     return true;
 }
 
-auto get_schema_subtree_bitmap(
-        KeyValuePairLogEvent::NodeIdValuePairs const& node_id_value_pairs,
-        SchemaTree const& schema_tree
-) -> OUTCOME_V2_NAMESPACE::std_result<vector<bool>> {
-    auto schema_subtree_bitmap{vector<bool>(schema_tree.get_size(), false)};
-    for (auto const& [node_id, val] : node_id_value_pairs) {
-        if (node_id >= schema_subtree_bitmap.size()) {
-            return std::errc::result_out_of_range;
-        }
-        schema_subtree_bitmap[node_id] = true;
-
-        // Iteratively mark the parents as true
-        auto optional_parent_id{schema_tree.get_node(node_id).get_parent_id()};
-        while (true) {
-            // Ideally, we'd use this if statement as the loop condition, but clang-tidy will
-            // complain about an unchecked `optional` access.
-            if (false == optional_parent_id.has_value()) {
-                // Reached the root
-                break;
-            }
-            auto const parent_id{optional_parent_id.value()};
-            if (schema_subtree_bitmap[parent_id]) {
-                // Parent already set by other child
-                break;
-            }
-            schema_subtree_bitmap[parent_id] = true;
-            optional_parent_id = schema_tree.get_node(parent_id).get_parent_id();
-        }
-    }
-    return schema_subtree_bitmap;
-}
-
 auto insert_kv_pair_into_json_obj(
         SchemaTree::Node const& node,
         std::optional<Value> const& optional_val,
@@ -393,6 +347,36 @@ auto KeyValuePairLogEvent::create(
     return KeyValuePairLogEvent{std::move(schema_tree), std::move(node_id_value_pairs), utc_offset};
 }
 
+auto KeyValuePairLogEvent::get_schema_subtree_bitmap(
+) const -> OUTCOME_V2_NAMESPACE::std_result<vector<bool>> {
+    auto schema_subtree_bitmap{vector<bool>(m_schema_tree->get_size(), false)};
+    for (auto const& [node_id, val] : m_node_id_value_pairs) {
+        if (node_id >= schema_subtree_bitmap.size()) {
+            return std::errc::result_out_of_range;
+        }
+        schema_subtree_bitmap[node_id] = true;
+
+        // Iteratively mark the parents as true
+        auto optional_parent_id{m_schema_tree->get_node(node_id).get_parent_id()};
+        while (true) {
+            // Ideally, we'd use this if statement as the loop condition, but clang-tidy will
+            // complain about an unchecked `optional` access.
+            if (false == optional_parent_id.has_value()) {
+                // Reached the root
+                break;
+            }
+            auto const parent_id{optional_parent_id.value()};
+            if (schema_subtree_bitmap[parent_id]) {
+                // Parent already set by other child
+                break;
+            }
+            schema_subtree_bitmap[parent_id] = true;
+            optional_parent_id = m_schema_tree->get_node(parent_id).get_parent_id();
+        }
+    }
+    return schema_subtree_bitmap;
+}
+
 auto KeyValuePairLogEvent::serialize_to_json(
 ) const -> OUTCOME_V2_NAMESPACE::std_result<nlohmann::json> {
     if (m_node_id_value_pairs.empty()) {
@@ -409,9 +393,7 @@ auto KeyValuePairLogEvent::serialize_to_json(
     // vector grows).
     std::stack<DfsIterator> dfs_stack;
 
-    auto const schema_subtree_bitmap_ret{
-            get_schema_subtree_bitmap(m_node_id_value_pairs, *m_schema_tree)
-    };
+    auto const schema_subtree_bitmap_ret{get_schema_subtree_bitmap()};
     if (schema_subtree_bitmap_ret.has_error()) {
         return schema_subtree_bitmap_ret.error();
     }

components/core/src/clp/ffi/KeyValuePairLogEvent.hpp

@@ -5,6 +5,7 @@
 #include <optional>
 #include <unordered_map>
 #include <utility>
+#include <vector>
 
 #include <json/single_include/nlohmann/json.hpp>
 #include <outcome/single-header/outcome.hpp>
@@ -60,6 +61,17 @@ class KeyValuePairLogEvent {
 
     [[nodiscard]] auto get_utc_offset() const -> UtcOffset { return m_utc_offset; }
 
+    /**
+     * @return A result containing a bitmap where every bit corresponds to the ID of a node in the
+     * schema tree, and the set bits correspond to the nodes in the subtree defined by all paths
+     * from the root node to the nodes in `node_id_value_pairs`; or an error code indicating a
+     * failure:
+     * - std::errc::result_out_of_range if a node ID in `node_id_value_pairs` doesn't exist in the
+     *   schema tree.
+     */
+    [[nodiscard]] auto get_schema_subtree_bitmap(
+    ) const -> OUTCOME_V2_NAMESPACE::std_result<std::vector<bool>>;
+
     /**
      * Serializes the log event into a `nlohmann::json` object.
      * @return A result containing the serialized JSON object or an error code indicating the

components/core/src/clp/ffi/ir_stream/Deserializer.hpp

@@ -14,7 +14,6 @@
 
 #include "../../ReaderInterface.hpp"
 #include "../../time_types.hpp"
-#include "../KeyValuePairLogEvent.hpp"
 #include "../SchemaTree.hpp"
 #include "decoding_methods.hpp"
 #include "ir_unit_deserialization_methods.hpp"
@@ -66,8 +65,8 @@ class Deserializer {
     /**
      * Deserializes the stream from the given reader up to and including the next log event IR unit.
      * @param reader
-     * @return std::errc::no_message_available if no tag bytes can be read to determine the next IR
-     * unit type.
+     * @return Forwards `deserialize_tag`s return values if no tag bytes can be read to determine
+     * the next IR unit type.
      * @return std::errc::protocol_not_supported if the IR unit type is not supported.
      * @return std::errc::operation_not_permitted if the deserializer already reached the end of
      * stream by deserializing an end-of-stream IR unit in the previous calls.
@@ -172,8 +171,8 @@ auto Deserializer<IrUnitHandler>::deserialize_next_ir_unit(ReaderInterface& read
     }
 
     encoded_tag_t tag{};
-    if (IRErrorCode::IRErrorCode_Success != deserialize_tag(reader, tag)) {
-        return std::errc::no_message_available;
+    if (auto const err{deserialize_tag(reader, tag)}; IRErrorCode::IRErrorCode_Success != err) {
+        return ir_error_code_to_errc(err);
     }
 
     auto const optional_ir_unit_type{get_ir_unit_type_from_tag(tag)};

components/core/src/clp/ffi/ir_stream/Serializer.hpp

@@ -14,7 +14,7 @@
 
 namespace clp::ffi::ir_stream {
 /**
- * A work-in-progress class for serializing log events into the kv-pair IR format.
+ * Class for serializing log events into the kv-pair IR format.
  *
  * This class:
  * - maintains all necessary internal data structures to track serialization state;

components/core/src/clp_s/JsonParser.cpp

@@ -21,7 +21,12 @@ JsonParser::JsonParser(JsonParserOption const& option)
     }
 
     if (false == m_timestamp_key.empty()) {
-        clp_s::StringUtils::tokenize_column_descriptor(m_timestamp_key, m_timestamp_column);
+        if (false
+            == clp_s::StringUtils::tokenize_column_descriptor(m_timestamp_key, m_timestamp_column))
+        {
+            SPDLOG_ERROR("Can not parse invalid timestamp key: \"{}\"", m_timestamp_key);
+            throw OperationFailed(ErrorCodeBadParam, __FILENAME__, __LINE__);
+        }
     }
 
     for (auto& file_path : option.file_paths) {

components/core/src/clp_s/TimestampDictionaryReader.cpp

@@ -44,7 +44,9 @@ void TimestampDictionaryReader::read_new_entries() {
         TimestampEntry entry;
         std::vector<std::string> tokens;
         entry.try_read_from_file(m_dictionary_decompressor);
-        StringUtils::tokenize_column_descriptor(entry.get_key_name(), tokens);
+        if (false == StringUtils::tokenize_column_descriptor(entry.get_key_name(), tokens)) {
+            throw OperationFailed(ErrorCodeCorrupt, __FILENAME__, __LINE__);
+        }
         m_entries.emplace_back(std::move(entry));
 
         // TODO: Currently, we only allow a single authoritative timestamp column at ingestion time,

components/core/src/clp_s/Utils.cpp

@@ -427,18 +427,34 @@ bool StringUtils::convert_string_to_double(std::string const& raw, double& conve
     return true;
 }
 
-void StringUtils::tokenize_column_descriptor(
+bool StringUtils::tokenize_column_descriptor(
         std::string const& descriptor,
         std::vector<std::string>& tokens
 ) {
-    // TODO: handle escaped . correctly
-    auto start = 0U;
-    auto end = descriptor.find('.');
-    while (end != std::string::npos) {
-        tokens.push_back(descriptor.substr(start, end - start));
-        start = end + 1;
-        end = descriptor.find('.', start);
+    // TODO: add support for unicode sequences e.g. \u263A
+    std::string cur_tok;
+    for (size_t cur = 0; cur < descriptor.size(); ++cur) {
+        if ('\\' == descriptor[cur]) {
+            ++cur;
+            if (cur >= descriptor.size()) {
+                return false;
+            }
+        } else if ('.' == descriptor[cur]) {
+            if (cur_tok.empty()) {
+                return false;
+            }
+            tokens.push_back(cur_tok);
+            cur_tok.clear();
+            continue;
+        }
+        cur_tok.push_back(descriptor[cur]);
     }
-    tokens.push_back(descriptor.substr(start));
+
+    if (cur_tok.empty()) {
+        return false;
+    }
+
+    tokens.push_back(cur_tok);
+    return true;
 }
 }  // namespace clp_s

components/core/src/clp_s/Utils.hpp

@@ -211,9 +211,9 @@ class StringUtils {
      * Converts a string column descriptor delimited by '.' into a list of tokens
      * @param descriptor
      * @param tokens
-     * @return the list of tokens pushed into the 'tokens' parameter
+     * @return true if the descriptor was tokenized successfully, false otherwise
      */
-    static void
+    [[nodiscard]] static bool
     tokenize_column_descriptor(std::string const& descriptor, std::vector<std::string>& tokens);
 
 private: