@eagleoutice/flowr 2.2.6 → 2.2.8

package/README.md

@@ -48,7 +48,7 @@ It offers a wide variety of features, for example:
     
     ```shell
     $ docker run -it --rm eagleoutice/flowr # or npm run flowr 
-    flowR repl using flowR v2.2.5, R v4.4.0 (r-shell engine)
+    flowR repl using flowR v2.2.7, R v4.4.0 (r-shell engine)
     R> :slicer test/testfiles/example.R --criterion "11@sum"
     ```
     
@@ -95,7 +95,7 @@ It offers a wide variety of features, for example:
         
 
 * 🚀 **fast data- and control-flow graphs**\
-  Within just <i><span title="This measurement is automatically fetched from the latest benchmark!">117.7 ms</i></span> (as of Feb 17, 2025), 
+  Within just <i><span title="This measurement is automatically fetched from the latest benchmark!">118 ms</i></span> (as of Feb 19, 2025), 
   _flowR_ can analyze the data- and control-flow of the average real-world R script. See the [benchmarks](https://flowr-analysis.github.io/flowr/wiki/stats/benchmark) for more information,
   and consult the [wiki pages](https://github.com/flowr-analysis/flowr/wiki/Dataflow-Graph) for more details on the dataflow graph.
 
@@ -131,7 +131,7 @@ It offers a wide variety of features, for example:
     
     ```shell
     $ docker run -it --rm eagleoutice/flowr # or npm run flowr 
-    flowR repl using flowR v2.2.5, R v4.4.0 (r-shell engine)
+    flowR repl using flowR v2.2.7, R v4.4.0 (r-shell engine)
     R> :dataflow* test/testfiles/example.R
     ```
     
@@ -375,7 +375,7 @@ It offers a wide variety of features, for example:
         52 -->|"argument"| 50
     ```
     	
-    (The analysis required _24.34 ms_ (including parse and normalize, using the [r-shell](https://github.com/flowr-analysis/flowr/wiki/Engines) engine) within the generation environment.)
+    (The analysis required _23.15 ms_ (including parse and normalize, using the [r-shell](https://github.com/flowr-analysis/flowr/wiki/Engines) engine) within the generation environment.)
     
     
     

package/cli/flowr-main-options.js

@@ -111,6 +111,12 @@ exports.flowrMainOptionDefinitions = [
         description: 'The path to the tree-sitter WASM binary to use. Defaults to the path specified by the tree-sitter package.',
         multiple: false
     },
+    {
+        name: 'engine.tree-sitter.lax',
+        type: Boolean,
+        description: 'Use the lax parser for parsing R code (allowing for syntax errors).',
+        multiple: false
+    },
     {
         name: 'default-engine',
         type: String,

package/cli/flowr.d.ts

@@ -20,6 +20,7 @@ export interface FlowrCliOptions {
     'engine.tree-sitter.disabled': boolean;
     'engine.tree-sitter.wasm-path': string | undefined;
     'engine.tree-sitter.tree-sitter-wasm-path': string | undefined;
+    'engine.tree-sitter.lax': boolean;
 }
 export declare const optionHelp: ({
     header: string;

package/cli/flowr.js

@@ -79,7 +79,8 @@ if (!options['engine.tree-sitter.disabled']) {
     (0, config_1.amendConfig)({ engines: [{
                 type: 'tree-sitter',
                 wasmPath: options['engine.tree-sitter.wasm-path'],
-                treeSitterWasmPath: options['engine.tree-sitter.tree-sitter-wasm-path']
+                treeSitterWasmPath: options['engine.tree-sitter.tree-sitter-wasm-path'],
+                lax: options['engine.tree-sitter.lax']
             }] });
 }
 if (options['default-engine']) {

package/config.d.ts

@@ -61,6 +61,10 @@ export interface TreeSitterEngineConfig extends MergeableRecord {
      * The path to the tree-sitter WASM binary to use. If this is undefined, the path specified by the tree-sitter package will be used.
      */
     readonly treeSitterWasmPath?: string;
+    /**
+     * Whether to use the lax parser for parsing R code (allowing for syntax errors). If this is undefined, the strict parser will be used.
+     */
+    readonly lax?: boolean;
 }
 export interface RShellEngineConfig extends MergeableRecord {
     readonly type: 'r-shell';

package/config.js

@@ -59,7 +59,8 @@ exports.flowrConfigFileSchema = joi_1.default.object({
     engines: joi_1.default.array().items(joi_1.default.alternatives(joi_1.default.object({
         type: joi_1.default.string().required().valid('tree-sitter').description('Use the tree sitter engine.'),
         wasmPath: joi_1.default.string().optional().description('The path to the tree-sitter-r WASM binary to use. If this is undefined, this uses the default path.'),
-        treeSitterWasmPath: joi_1.default.string().optional().description('The path to the tree-sitter WASM binary to use. If this is undefined, this uses the default path.')
+        treeSitterWasmPath: joi_1.default.string().optional().description('The path to the tree-sitter WASM binary to use. If this is undefined, this uses the default path.'),
+        lax: joi_1.default.boolean().optional().description('Whether to use the lax parser for parsing R code (allowing for syntax errors). If this is undefined, the strict parser will be used.')
     }).description('The configuration for the tree sitter engine.'), joi_1.default.object({
         type: joi_1.default.string().required().valid('r-shell').description('Use the R shell engine.'),
         rPath: joi_1.default.string().optional().description('The path to the R executable to use. If this is undefined, this uses the default path.')

package/dataflow/environments/resolve-by-name.d.ts

@@ -27,7 +27,7 @@ export declare function trackAliasesInGraph(id: NodeId, graph: DataflowGraph, id
  * Convenience function using the variable resolver as specified within the configuration file
  * In the future we may want to have this set once at the start of the analysis
  *
- * @see {@link resolve} - for a more general approach which "evaluates" a node based on value resolve
+ * @see {@link resolveIdToValue} - for a more general approach which "evaluates" a node based on value resolve
  */
 export declare function resolveValueOfVariable(identifier: Identifier | undefined, environment: REnvironmentInformation, idMap?: AstIdMap): unknown[] | undefined;
 export interface ResolveInfo {
@@ -43,9 +43,10 @@ export interface ResolveInfo {
 /**
  * Generalized {@link resolveValueOfVariable} function which evaluates a node based on the value resolve
  *
- * @param id         - The node id or node to resolve
+ * @param id          - The node id or node to resolve
  * @param environment - The current environment used for name resolution
- * @param idMap      - The id map to resolve the node if given as an id
- * @param full       - Whether to track variables
+ * @param graph       - The graph to resolve in
+ * @param idMap       - The id map to resolve the node if given as an id
+ * @param full        - Whether to track variables
  */
-export declare function resolve(id: NodeId | RNodeWithParent, { environment, graph, idMap, full }: ResolveInfo): unknown[] | undefined;
+export declare function resolveIdToValue(id: NodeId | RNodeWithParent, { environment, graph, idMap, full }: ResolveInfo): unknown[] | undefined;

package/dataflow/environments/resolve-by-name.js

@@ -7,7 +7,7 @@ exports.getAliases = getAliases;
 exports.trackAliasInEnvironments = trackAliasInEnvironments;
 exports.trackAliasesInGraph = trackAliasesInGraph;
 exports.resolveValueOfVariable = resolveValueOfVariable;
-exports.resolve = resolve;
+exports.resolveIdToValue = resolveIdToValue;
 const environment_1 = require("./environment");
 const logic_1 = require("../../util/logic");
 const identifier_1 = require("./identifier");
@@ -189,6 +189,20 @@ function trackAliasInEnvironments(identifier, use, idMap) {
     }
     return values;
 }
+function isNestedInLoop(node, ast) {
+    const parent = node?.info.parent;
+    if (node === undefined || !parent) {
+        return false;
+    }
+    const parentNode = ast.get(parent);
+    if (parentNode === undefined) {
+        return false;
+    }
+    if (parentNode.type === type_1.RType.WhileLoop || parentNode.type === type_1.RType.RepeatLoop) {
+        return true;
+    }
+    return isNestedInLoop(parentNode, ast);
+}
 function trackAliasesInGraph(id, graph, idMap) {
     idMap ??= graph.idMap;
     (0, assert_1.guard)(idMap !== undefined, 'The ID map is required to get the lineage of a node');
@@ -198,6 +212,7 @@ function trackAliasesInGraph(id, graph, idMap) {
     const clean = (0, environment_1.initializeCleanEnvironments)();
     const cleanFingerprint = (0, fingerprint_1.envFingerprint)(clean);
     queue.add(id, clean, cleanFingerprint, false);
+    let forceBot = false;
     const resultIds = [];
     while (queue.nonEmpty()) {
         const { id, baseEnvironment } = queue.next();
@@ -206,18 +221,36 @@ function trackAliasesInGraph(id, graph, idMap) {
             continue;
         }
         const [vertex, outgoingEdges] = res;
+        const cds = vertex.controlDependencies;
+        for (const cd of cds ?? []) {
+            const target = graph.idMap?.get(cd.id);
+            if (target === undefined) {
+                continue;
+            }
+            if (target.type === type_1.RType.WhileLoop || target.type === type_1.RType.RepeatLoop) {
+                forceBot = true;
+                break;
+            }
+        }
+        if (!forceBot && (cds?.length === 0 && isNestedInLoop(idMap.get(id), idMap))) {
+            forceBot = true;
+        }
+        if (forceBot) {
+            break;
+        }
         if (vertex.tag === vertex_1.VertexType.Value) {
             resultIds.push(id);
             continue;
         }
         // travel all read and defined-by edges
         for (const [targetId, edge] of outgoingEdges) {
-            if ((0, edge_1.edgeIncludesType)(edge.types, edge_1.EdgeType.Reads | edge_1.EdgeType.DefinedBy | edge_1.EdgeType.DefinedByOnCall)) {
+            // currently, they have to be exact!
+            if (edge.types === edge_1.EdgeType.Reads || edge.types === edge_1.EdgeType.DefinedBy || edge.types === edge_1.EdgeType.DefinedByOnCall) {
                 queue.add(targetId, baseEnvironment, cleanFingerprint, false);
             }
         }
     }
-    if (resultIds.length === 0) {
+    if (forceBot || resultIds.length === 0) {
         return undefined;
     }
     const values = [];
@@ -233,7 +266,7 @@ function trackAliasesInGraph(id, graph, idMap) {
  * Convenience function using the variable resolver as specified within the configuration file
  * In the future we may want to have this set once at the start of the analysis
  *
- * @see {@link resolve} - for a more general approach which "evaluates" a node based on value resolve
+ * @see {@link resolveIdToValue} - for a more general approach which "evaluates" a node based on value resolve
  */
 function resolveValueOfVariable(identifier, environment, idMap) {
     const resolve = (0, config_1.getConfig)().solver.variables;
@@ -247,12 +280,13 @@ function resolveValueOfVariable(identifier, environment, idMap) {
 /**
  * Generalized {@link resolveValueOfVariable} function which evaluates a node based on the value resolve
  *
- * @param id         - The node id or node to resolve
+ * @param id          - The node id or node to resolve
  * @param environment - The current environment used for name resolution
- * @param idMap      - The id map to resolve the node if given as an id
- * @param full       - Whether to track variables
+ * @param graph       - The graph to resolve in
+ * @param idMap       - The id map to resolve the node if given as an id
+ * @param full        - Whether to track variables
  */
-function resolve(id, { environment, graph, idMap, full }) {
+function resolveIdToValue(id, { environment, graph, idMap, full }) {
     idMap ??= graph?.idMap;
     const node = typeof id === 'object' ? id : idMap?.get(id);
     if (node === undefined) {

package/documentation/print-engines-wiki.js

@@ -39,6 +39,7 @@ are exposed with some command line options (e.g., when using the docker image of
 - ${(0, doc_cli_option_1.getCliLongOptionOf)('flowr', 'engine.r-shell.disabled', false)} to disable the ${(0, doc_types_1.shortLink)(shell_1.RShell.name, types.info)} engine
 - ${(0, doc_cli_option_1.getCliLongOptionOf)('flowr', 'engine.r-shell.r-path', false)} (which is the canonical version of ${(0, doc_cli_option_1.getCliLongOptionOf)('flowr', 'r-path')})
 - ${(0, doc_cli_option_1.getCliLongOptionOf)('flowr', 'engine.tree-sitter.disabled', false)} to disable the ${(0, doc_types_1.shortLink)(tree_sitter_executor_1.TreeSitterExecutor.name, types.info)} engine
+- ${(0, doc_cli_option_1.getCliLongOptionOf)('flowr', 'engine.tree-sitter.lax', false)} to use lax parsing with tree-sitter
 - ${(0, doc_cli_option_1.getCliLongOptionOf)('flowr', 'engine.tree-sitter.wasm-path', false)} pass the path to the wasm of the r grammar of tree-sitter (see [below](#tree-sitter))
 - ${(0, doc_cli_option_1.getCliLongOptionOf)('flowr', 'engine.tree-sitter.tree-sitter-wasm-path', false)} pass the path to the wasm of tree-sitter (see [below](#tree-sitter))
 - ${(0, doc_cli_option_1.getCliLongOptionOf)('flowr', 'default-engine', false)} to set the default engine to use

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@eagleoutice/flowr",
-  "version": "2.2.6",
+  "version": "2.2.8",
   "description": "Static Dataflow Analyzer and Program Slicer for the R Programming Language",
   "types": "dist/src/index.d.ts",
   "repository": {

package/queries/catalog/dependencies-query/dependencies-query-executor.js

@@ -76,7 +76,7 @@ function executeDependenciesQuery(data, queries) {
         nodeId: id,
         functionName: vertex.name,
         // write functions that don't have argIndex are assumed to write to stdout
-        destination: value ?? (linkedIds?.length ? dependencies_query_format_1.Unknown : 'stdout'),
+        destination: value ?? 'stdout',
         lexemeOfArgument: getLexeme(value, argId),
         linkedIds: linkedIds?.length ? linkedIds : undefined
     }));
@@ -160,7 +160,7 @@ function collectValuesFromLinks(args, data, linkedIds) {
             }
         }
     }
-    return map;
+    return map.size ? map : undefined;
 }
 function hasCharacterOnly(data, vertex, idMap) {
     if (!vertex.args || vertex.args.length === 0 || !idMap) {
@@ -193,7 +193,7 @@ function resolveBasedOnConfig(data, vertex, argument, environment, idMap, resolv
             full = false;
         }
     }
-    return (0, resolve_by_name_1.resolve)(argument, { environment, graph: data.dataflow.graph, full });
+    return (0, resolve_by_name_1.resolveIdToValue)(argument, { environment, graph: data.dataflow.graph, full });
 }
 function unwrapRValue(value) {
     if (value === undefined) {

package/queries/catalog/dependencies-query/dependencies-query-format.js

@@ -18,7 +18,7 @@ var DependencyInfoLinkConstraint;
 // these lists are originally based on https://github.com/duncantl/CodeDepends/blob/7fd96dfee16b252e5f642c77a7ababf48e9326f8/R/codeTypes.R
 exports.LibraryFunctions = [
     { name: 'library', argIdx: 0, argName: 'package', resolveValue: 'library' },
-    { name: 'require', argIdx: 0, argName: 'package', resolveValue: true },
+    { name: 'require', argIdx: 0, argName: 'package', resolveValue: 'library' },
     { name: 'loadNamespace', argIdx: 0, argName: 'package', resolveValue: true },
     { name: 'attachNamespace', argIdx: 0, argName: 'ns', resolveValue: true },
     { name: 'attach', argIdx: 0, argName: 'what', resolveValue: true },

package/queries/catalog/resolve-value-query/resolve-value-query-executor.js

@@ -18,9 +18,9 @@ function executeResolveValueQuery({ dataflow: { graph }, ast }, queries) {
         }
         const values = query.criteria
             .map(criteria => (0, parse_1.slicingCriterionToId)(criteria, ast.idMap))
-            .flatMap(ident => (0, resolve_by_name_1.resolve)(ident, { graph, full: true, idMap: ast.idMap }));
+            .flatMap(ident => (0, resolve_by_name_1.resolveIdToValue)(ident, { graph, full: true, idMap: ast.idMap }) ?? []);
         results[key] = {
-            values: [...new Set(values)]
+            values: values ? [...new Set(values)] : []
         };
     }
     return {

package/r-bridge/lang-4.x/tree-sitter/tree-sitter-executor.d.ts

@@ -10,7 +10,12 @@ export declare class TreeSitterExecutor implements SyncParser<Parser.Tree> {
     readonly name = "tree-sitter";
     readonly parser: Parser;
     private static language;
-    static initTreeSitter(): Promise<void>;
+    /**
+     * Initializes the underlying tree-sitter parser. This only needs to be called once globally.
+     * @param overrideWasmPath - The path to the tree-sitter-r wasm file, which takes precedence over the config and default paths if set.
+     * @param overrideTreeSitterWasmPath - The path to the tree-sitter wasm file, which takes precedence over the config and default paths if set.
+     */
+    static initTreeSitter(overrideWasmPath?: string, overrideTreeSitterWasmPath?: string): Promise<void>;
     constructor();
     rVersion(): Promise<string | 'unknown' | 'none'>;
     treeSitterVersion(): number;

package/r-bridge/lang-4.x/tree-sitter/tree-sitter-executor.js

@@ -7,8 +7,10 @@ exports.TreeSitterExecutor = exports.DEFAULT_TREE_SITTER_WASM_PATH = exports.DEF
 const web_tree_sitter_1 = __importDefault(require("web-tree-sitter"));
 const fs_1 = __importDefault(require("fs"));
 const config_1 = require("../../../config");
+const log_1 = require("../../../util/log");
 exports.DEFAULT_TREE_SITTER_R_WASM_PATH = `${__dirname}/tree-sitter-r.wasm`;
 exports.DEFAULT_TREE_SITTER_WASM_PATH = `${__dirname}/tree-sitter.wasm`;
+const wasmLog = log_1.log.getSubLogger({ name: 'tree-sitter-wasm' });
 /**
  * Synchronous and (way) faster alternative to the {@link RShell} using tree-sitter.
  */
@@ -16,9 +18,13 @@ class TreeSitterExecutor {
     name = 'tree-sitter';
     parser;
     static language;
-    static async initTreeSitter() {
-        const config = (0, config_1.getEngineConfig)('tree-sitter');
-        const treeSitterWasmPath = config?.treeSitterWasmPath ?? exports.DEFAULT_TREE_SITTER_WASM_PATH;
+    /**
+     * Initializes the underlying tree-sitter parser. This only needs to be called once globally.
+     * @param overrideWasmPath - The path to the tree-sitter-r wasm file, which takes precedence over the config and default paths if set.
+     * @param overrideTreeSitterWasmPath - The path to the tree-sitter wasm file, which takes precedence over the config and default paths if set.
+     */
+    static async initTreeSitter(overrideWasmPath, overrideTreeSitterWasmPath) {
+        const treeSitterWasmPath = overrideTreeSitterWasmPath ?? (0, config_1.getEngineConfig)('tree-sitter')?.treeSitterWasmPath ?? exports.DEFAULT_TREE_SITTER_WASM_PATH;
         // noinspection JSUnusedGlobalSymbols - this is used by emscripten, see https://emscripten.org/docs/api_reference/module.html
         await web_tree_sitter_1.default.init({
             locateFile: (path, prefix) => {
@@ -27,9 +33,12 @@ class TreeSitterExecutor {
                     return treeSitterWasmPath;
                 }
                 return prefix + path;
-            }
+            },
+            onAbort: (s) => wasmLog.error(`Tree-sitter wasm aborted: ${s}`),
+            print: (s) => wasmLog.debug(s),
+            printErr: (s) => wasmLog.error(s)
         });
-        const wasmPath = config?.wasmPath ?? exports.DEFAULT_TREE_SITTER_R_WASM_PATH;
+        const wasmPath = overrideWasmPath ?? (0, config_1.getEngineConfig)('tree-sitter')?.wasmPath ?? exports.DEFAULT_TREE_SITTER_R_WASM_PATH;
         TreeSitterExecutor.language = await web_tree_sitter_1.default.Language.load(wasmPath);
     }
     constructor() {

package/r-bridge/lang-4.x/tree-sitter/tree-sitter-normalize.d.ts

@@ -1,3 +1,8 @@
 import type { RExpressionList } from '../ast/model/nodes/r-expression-list';
 import type { Tree } from 'web-tree-sitter';
+/**
+ * @param tree - The tree to normalize
+ */
 export declare function normalizeTreeSitterTreeToAst(tree: Tree): RExpressionList;
+export declare function makeTreeSitterLax(): void;
+export declare function makeTreeSitterStrict(): void;

package/r-bridge/lang-4.x/tree-sitter/tree-sitter-normalize.js

@@ -1,6 +1,8 @@
 "use strict";
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.normalizeTreeSitterTreeToAst = normalizeTreeSitterTreeToAst;
+exports.makeTreeSitterLax = makeTreeSitterLax;
+exports.makeTreeSitterStrict = makeTreeSitterStrict;
 const normalizer_data_1 = require("../ast/parser/main/normalizer-data");
 const tree_sitter_types_1 = require("./tree-sitter-types");
 const type_1 = require("../ast/model/type");
@@ -10,20 +12,39 @@ const normalize_meta_1 = require("../ast/parser/main/normalize-meta");
 const arrays_1 = require("../../../util/arrays");
 const r_function_call_1 = require("../ast/model/nodes/r-function-call");
 const strings_1 = require("../../../util/strings");
+const config_1 = require("../../../config");
+const log_1 = require("../../../util/log");
+/**
+ * @param tree - The tree to normalize
+ */
 function normalizeTreeSitterTreeToAst(tree) {
+    const lax = (0, config_1.getEngineConfig)('tree-sitter')?.lax;
+    if (lax) {
+        makeTreeSitterLax();
+    }
+    else {
+        makeTreeSitterStrict();
+    }
     const root = convertTreeNode(tree.rootNode);
     if (root.type !== type_1.RType.ExpressionList) {
         throw new normalizer_data_1.ParseError(`expected root to resolve to an expression list, got a ${root.type}`);
     }
     return root;
 }
-function nonErrorChildren(node) {
-    if (node.hasError) {
-        return [];
-    }
-    else {
-        return node.children;
-    }
+function nonErrorChildrenStrict(node) {
+    return node.hasError ? [] : node.children;
+}
+function nonErrorChildrenLax(node) {
+    return node.hasError ? node.children.filter(n => n.type !== tree_sitter_types_1.TreeSitterType.Error) : node.children;
+}
+let nonErrorChildren = nonErrorChildrenStrict;
+function makeTreeSitterLax() {
+    log_1.log.info('[Tree-Sitter] Lax parsing active');
+    nonErrorChildren = nonErrorChildrenLax;
+}
+function makeTreeSitterStrict() {
+    log_1.log.info('[Tree-Sitter] Strict parsing active');
+    nonErrorChildren = nonErrorChildrenStrict;
 }
 function convertTreeNode(node) {
     // generally, the grammar source file dictates what children a node has in what order:

package/util/version.js

@@ -3,7 +3,7 @@ Object.defineProperty(exports, "__esModule", { value: true });
 exports.flowrVersion = flowrVersion;
 const semver_1 = require("semver");
 // this is automatically replaced with the current version by release-it
-const version = '2.2.6';
+const version = '2.2.8';
 function flowrVersion() {
     return new semver_1.SemVer(version);
 }