Merge remote-tracking branch 'upstream/main' into random-values-from-…

…file

.github/workflows/TestDebug.yml

@@ -0,0 +1,42 @@
+name: Test Debug
+on:
+  workflow_dispatch:
+  pull_request:
+  push:
+    branches:
+      - '**'
+      - '!main'
+
+jobs:
+  build-duckdb:
+    name: Build DuckDB
+    runs-on: ubuntu-latest
+    timeout-minutes: 120
+    outputs:
+      duckdb-hash: ${{ steps.find-hash.outputs.hash }}
+    env:
+      BUILD_ICU: 1
+      BUILD_JSON: 1
+      BUILD_TPCH: 1
+      BUILD_TPCDS: 1
+      BUILD_PARQUET: 1
+      BUILD_JEMALLOC: 1
+      CRASH_ON_ASSERT: 1
+
+    steps:
+      - name: Dependencies
+        shell: bash
+        run: sudo apt-get update -y -qq && sudo apt-get install -y -qq ninja-build ccache
+
+      - uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+          submodules: recursive
+
+      - name: Setup Ccache
+        uses: hendrikmuhs/ccache-action@main
+
+      - name: Build and Test
+        shell: bash
+        run: |
+          make test_debug

.github/workflows/test-fuzzer-ci-still-works.yml

@@ -10,88 +10,31 @@ on:
 jobs:
   build-duckdb:
     name: Build DuckDB
-    runs-on: ubuntu-latest
-    timeout-minutes: 120
-    outputs:
-      duckdb-hash: ${{ steps.find-hash.outputs.hash }}
-    env:
-      BUILD_ICU: 1
-      BUILD_JSON: 1
-      BUILD_TPCH: 1
-      BUILD_TPCDS: 1
-      BUILD_PARQUET: 1
-      BUILD_JEMALLOC: 1
-      CRASH_ON_ASSERT: 1
-      GEN: ninja
-
-    steps:
-      - name: Dependencies
-        shell: bash
-        run: sudo apt-get update -y -qq && sudo apt-get install -y -qq ninja-build ccache
-
-      - uses: actions/checkout@v3
-        with:
-          repository: duckdb/duckdb
-          fetch-depth: 0
-
-      - name: Setup Ccache
-        uses: hendrikmuhs/ccache-action@main
-
-      - id: find-hash
-        run: echo "::set-output name=hash::$(git rev-parse HEAD)"
-
-      - name: create build sqlsmith extension file
-        shell: bash
-        run: |
-          echo "duckdb_extension_load(sqlsmith 
-            GIT_URL https://github.com/duckdb/duckdb_sqlsmith 
-            GIT_TAG main 
-          )" > sqlsmith.cmake
-
-      - name: Build
-        shell: bash
-        run: |
-          EXTENSION_CONFIGS="sqlsmith.cmake" make debug
-
-      - uses: actions/upload-artifact@v3
-        with:
-          name: duckdb
-          path: build/debug/duckdb
+    uses: duckdblabs/duckdb-fuzzer-ci/.github/workflows/reusable_workflows/build_fuzzer.yml@main
+    with:
+      git_url: ${{ github.actor }}
+      git_tag: ${{ github.ref_name }}
+      timeout-minutes: 120
 
   fuzzer:
     name: Fuzzer
     needs:
     - build-duckdb
-    runs-on: ubuntu-latest
-    timeout-minutes: 20
     strategy:
       fail-fast: false
       matrix:
         fuzzer: [duckfuzz, sqlsmith, duckfuzz_functions]
         data: [emptyalltypes]
-
-    steps:
-      - uses: actions/checkout@v3
-        with:
-          path: duckdb_sqlsmith
-          fetch-depth: 0
-
-      - name: Download a single artifact
-        uses: actions/download-artifact@v3
-        with:
-          name: duckdb
-
-      - name: Fuzz
-        shell: bash
-        run: |
-            chmod +x duckdb
-            runtime="1 minute"
-            endtime=$(date -ud "$runtime" +%s)
-
-            cd duckdb_sqlsmith
-            while [[ $(date -u +%s) -le $endtime ]]
-            do
-                echo "Time Now: `date +%H:%M:%S`"
-                python3 scripts/run_fuzzer.py --no_checks --${{ matrix.fuzzer }} --${{ matrix.data }} --shell=../duckdb --dry --max_queries=10
-            done
-
+        enable_verification: [true, false]
+        exclude:
+          - enable_verification: true
+            fuzzer: sqlsmith
+          - enable_verification: true
+            fuzzer: duckfuzz_functions
+    uses: duckdblabs/duckdb-fuzzer-ci/.github/workflows/reusable_workflows/fuzz_duckdb.yml@main
+    with:
+      fuzzer: ${{ matrix.fuzzer }}
+      data: ${{ matrix.data }}
+      timeout-minutes: 20
+      max_queries: 10
+      enable_verification: ${{ matrix.enable_verification }}

scripts/reduce_sql.py

@@ -18,6 +18,7 @@
 SELECT * FROM reduce_sql_statement('${QUERY}');
 '''
 
+verbose = False
 
 class MultiStatementManager:
     delimiter = ';'
@@ -33,7 +34,8 @@ def __init__(self, multi_statement):
                 self.statements.append(stmt.strip() + ";")
 
     def is_multi_statement(sql_statement):
-        if len(sql_statement.split(';')) > 1:
+        splits = [x for x in sql_statement.split(';') if len(x.strip()) > 0]
+        if len(splits) > 1:
             return True
         return False
 
@@ -45,10 +47,12 @@ def sanitize_error(err):
     err = re.sub(r'Error: near line \d+: ', '', err)
     err = err.replace(os.getcwd() + '/', '')
     err = err.replace(os.getcwd(), '')
+    err = re.sub(r'LINE \d+:.*\n', '', err)
+    err = re.sub(r' *\^ *', '', err)
     if 'AddressSanitizer' in err:
         match = re.search(r'[ \t]+[#]0 ([A-Za-z0-9]+) ([^\n]+)', err).groups()[1]
         err = 'AddressSanitizer error ' + match
-    return err
+    return err.strip()
 
 
 def run_shell_command(shell, cmd):
@@ -69,7 +73,13 @@ def get_reduced_sql(shell, sql_query):
         raise Exception("Failed to reduce query")
     reduce_candidates = []
     for line in stdout.split('\n'):
-        reduce_candidates.append(line.strip('"').replace('""', '"'))
+        if len(line) <= 2:
+            continue
+        if line[0] == '"':
+            line = line[1:]
+        if line[len(line) - 1] == '"':
+            line = line[:len(line) - 1]
+        reduce_candidates.append(line.replace('""', '"'))
     return reduce_candidates[1:]
 
 
@@ -95,6 +105,15 @@ def reduce(sql_query, data_load, shell, error_msg, max_time_seconds=300):
                 print(sql_query)
                 print("=======================")
                 break
+            elif verbose:
+                print("Failed to reduce query")
+                print("=======================")
+                print(reduce_candidate)
+                print("=====Target error======")
+                print(error_msg)
+                print("=====Actual error======")
+                print(new_error)
+                print("=======================")
         if not found_new_candidate:
             break
     return sql_query
@@ -271,13 +290,16 @@ def reduce_query_log(queries, shell, data_load=[], max_time_seconds=300):
     parser.add_argument(
         '--max-time', dest='max_time', action='store', help='Maximum time in seconds to run the reducer', default=300
     )
+    parser.add_argument(
+        '--verbose', dest='verbose', action='store_true', help='Verbose output')
 
     args = parser.parse_args()
     print("Starting reduce process")
 
     shell = args.shell
     data_load = open(args.load).read()
     sql_query = open(args.exec).read()
+    verbose = args.verbose
     (stdout, stderr, returncode) = run_shell_command(shell, data_load + sql_query)
     expected_error = sanitize_error(stderr).strip()
     if len(expected_error) == 0:

scripts/run_fuzzer.py

@@ -13,9 +13,10 @@
 db = None
 shell = None
 perform_checks = True
-dry = False
+no-git-checks = False
 max_queries = 1000
 verification = False
+
 for param in sys.argv:
     if param == '--sqlsmith':
         fuzzer = 'sqlsmith'
@@ -31,16 +32,16 @@
         db = 'emptyalltypes'
     elif param == '--no_checks':
         perform_checks = False
-    elif param == '--enable_verification':
-        verification = True
+    elif param.startswith('--enable_verification'):
+        verification = param.replace('--enable_verification=', '').lower() == 'true'
     elif param.startswith('--shell='):
         shell = param.replace('--shell=', '')
     elif param.startswith('--seed='):
         seed = int(param.replace('--seed=', ''))
     elif param.startswith('--max_queries='):
         max_queries = int(param.replace('--max_queries=', ''))
-    elif param.startswith('--dry'):
-        dry = True
+    elif param.startswith('--no-git-checks'):
+        no-git-checks = param.replace('--no-git-checks=', '').lower() == 'true'
 
 if fuzzer is None:
     print("Unrecognized fuzzer to run, expected e.g. --sqlsmith or --duckfuzz")
@@ -103,11 +104,14 @@ def run_shell_command(cmd):
 
 # first get a list of all github issues, and check if we can still reproduce them
 
-if dry:
+if no-git-checks:
     current_errors = []
 else:
     current_errors = fuzzer_helper.extract_github_issues(shell, perform_checks)
 
+# Don't go on and fuzz if perform checks = true
+if perform_checks:
+    exit(0)
 
 last_query_log_file = 'sqlsmith.log'
 complete_log_file = 'sqlsmith.complete.log'
@@ -118,6 +122,7 @@ def run_shell_command(cmd):
 =========================================='''
 )
 
+
 load_script = create_db_script(db)
 fuzzer_name = get_fuzzer_name(fuzzer)
 fuzzer = (
@@ -212,5 +217,5 @@ def run_shell_command(cmd):
 print(f"{error_msg}")
 print(f"================MARKER====================")
 
-if not dry:
+if not no-git-checks:
     fuzzer_helper.file_issue(cmd, error_msg, fuzzer_name, seed, git_hash)

src/include/statement_simplifier.hpp

@@ -16,6 +16,7 @@ class SQLStatement;
 class SelectStatement;
 class InsertStatement;
 class UpdateStatement;
+class PrepareStatement;
 class DeleteStatement;
 class TableRef;
 class SelectNode;
@@ -41,6 +42,7 @@ class StatementSimplifier {
 	void Simplify(SelectStatement &stmt);
 	void Simplify(InsertStatement &stmt);
 	void Simplify(UpdateStatement &stmt);
+	void Simplify(PrepareStatement &stmt);
 	void Simplify(DeleteStatement &stmt);
 
 	void Simplification();

src/statement_simplifier.cpp

@@ -8,6 +8,7 @@
 #include "duckdb/parser/expression/list.hpp"
 #include "duckdb/parser/statement/delete_statement.hpp"
 #include "duckdb/parser/statement/insert_statement.hpp"
+#include "duckdb/parser/statement/prepare_statement.hpp"
 #include "duckdb/parser/statement/update_statement.hpp"
 #include "duckdb/parser/statement/select_statement.hpp"
 #endif
@@ -308,6 +309,7 @@ void StatementSimplifier::SimplifyExpression(duckdb::unique_ptr<ParsedExpression
 			SimplifyChildExpression(expr, case_check.then_expr);
 			SimplifyChildExpression(expr, case_check.when_expr);
 		}
+		SimplifyList(op.case_checks, false);
 		break;
 	}
 	case ExpressionClass::CAST: {
@@ -423,6 +425,10 @@ void StatementSimplifier::Simplify(UpdateSetInfo &info) {
 	}
 }
 
+void StatementSimplifier::Simplify(PrepareStatement &stmt) {
+	Simplify(*stmt.statement);
+}
+
 void StatementSimplifier::Simplify(UpdateStatement &stmt) {
 	Simplify(stmt.cte_map);
 	SimplifyOptional(stmt.from_table);
@@ -445,6 +451,9 @@ void StatementSimplifier::Simplify(SQLStatement &stmt) {
 	case StatementType::DELETE_STATEMENT:
 		Simplify(stmt.Cast<DeleteStatement>());
 		break;
+	case StatementType::PREPARE_STATEMENT:
+		Simplify(stmt.Cast<PrepareStatement>());
+		break;
 	default:
 		throw InvalidInputException("Expected a single SELECT, INSERT or UPDATE statement");
 	}

test/sql/call_fuzzyduck.test

@@ -0,0 +1,11 @@
+# name: test/sql/call_fuzzyduck.test
+# description: Test fuzzyduck
+# group: [fuzzyduck]
+
+require sqlsmith
+
+statement ok
+call fuzzyduck(max_queries=2, verbose_output=1, log='sqlsmith.log', complete_log='sqlsmith.complete.log', enable_verification=True);
+
+statement ok
+call fuzzyduck(max_queries=2, verbose_output=1, log='__TEST_DIR__/logs.txt', complete_log='__TEST_DIR__/clog.txt', enable_verification=false);

test/sql/sql_reduce.test

@@ -1,4 +1,4 @@
-# name: test/sql_reduce.test
+# name: test/sql/sql_reduce.test
 # description: Test reduce SQL statement
 # group: [sqlsmith]
 
@@ -77,3 +77,16 @@ DELETE FROM a WHERE (i < 5000)
 DELETE FROM a WHERE (i >= 2000)
 DELETE FROM a WHERE (i AND (i < 5000))
 DELETE FROM a WHERE NULL
+
+query I
+SELECT * FROM reduce_sql_statement('PREPARE v1 AS SELECT a, b FROM tbl') ORDER BY 1
+----
+PREPARE v1 AS SELECT NULL, b FROM tbl
+PREPARE v1 AS SELECT NULL, b FROM tbl
+PREPARE v1 AS SELECT a FROM tbl
+PREPARE v1 AS SELECT a, NULL FROM tbl
+PREPARE v1 AS SELECT a, NULL FROM tbl
+PREPARE v1 AS SELECT a, b
+PREPARE v1 AS SELECT a, b FROM tbl
+PREPARE v1 AS SELECT a, b FROM tbl
+PREPARE v1 AS SELECT b FROM tbl