@eagleoutice/flowr 2.2.15 → 2.3.0

package/cli/script-core/statistics-helper-core.js

@@ -20,6 +20,7 @@ const assert_1 = require("../../util/assert");
 const time_1 = require("../../util/text/time");
 const tar_1 = require("tar");
 const ansi_1 = require("../../util/text/ansi");
+const config_1 = require("../../config");
 function compressFolder(folder, target) {
     return (0, tar_1.create)({
         gzip: true,
@@ -35,7 +36,7 @@ function compressFolder(folder, target) {
         console.error('Error during compression:', e);
     });
 }
-async function getStatsForSingleFile(options) {
+async function getStatsForSingleFile(options, config) {
     if (options['no-ansi']) {
         log_1.log.info('disabling ansi colors');
         (0, ansi_1.setFormatter)(ansi_1.voidFormatter);
@@ -50,15 +51,15 @@ async function getStatsForSingleFile(options) {
     }
     // assume correct
     const processedFeatures = new Set(options.features);
-    const shell = new shell_1.RShell();
+    const shell = new shell_1.RShell((0, config_1.getEngineConfig)(config, 'r-shell'));
     (0, statistics_file_1.initFileProvider)(options['output-dir']);
     await shell.obtainTmpDir();
-    const stats = await (0, statistics_1.extractUsageStatistics)(shell, () => { }, processedFeatures, (0, statistics_1.staticRequests)({ request: 'file', content: options.input }), options['root-dir']);
+    const stats = await (0, statistics_1.extractUsageStatistics)(shell, config, () => { }, processedFeatures, (0, statistics_1.staticRequests)({ request: 'file', content: options.input }), options['root-dir']);
     // console.warn(`skipped ${stats.meta.failedRequests.length} requests due to errors (run with logs to get more info)`)
     if (stats.outputs.size === 1) {
         if (options['dump-json']) {
             const [, output] = [...stats.outputs.entries()][0];
-            const cfg = (0, extract_cfg_1.extractCfg)(output.normalize, output.dataflow.graph);
+            const cfg = (0, extract_cfg_1.extractCfg)(output.normalize, config, output.dataflow.graph);
             statistics_file_1.statisticsFileProvider.append('output-json', 'parse', await (0, print_1.printStepResult)(_00_parse_1.PARSE_WITH_R_SHELL_STEP, output.parse, 2 /* StepOutputFormat.Json */));
             statistics_file_1.statisticsFileProvider.append('output-json', 'normalize', await (0, print_1.printStepResult)(_10_normalize_1.NORMALIZE, output.normalize, 2 /* StepOutputFormat.Json */));
             statistics_file_1.statisticsFileProvider.append('output-json', 'dataflow', await (0, print_1.printStepResult)(_20_dataflow_1.STATIC_DATAFLOW, output.dataflow, 2 /* StepOutputFormat.Json */));

package/cli/slicer-app.js

@@ -14,6 +14,7 @@ const slicer_1 = require("../benchmark/slicer");
 const print_1 = require("../benchmark/stats/print");
 const magic_comments_1 = require("../reconstruct/auto-select/magic-comments");
 const auto_select_defaults_1 = require("../reconstruct/auto-select/auto-select-defaults");
+const config_1 = require("../config");
 const options = (0, script_1.processCommandLineArgs)('slicer', ['input', 'criterion'], {
     subtitle: 'Slice R code based on a given slicing criterion',
     examples: [
@@ -28,9 +29,10 @@ async function getSlice() {
     const slicer = new slicer_1.BenchmarkSlicer('r-shell');
     (0, assert_1.guard)(options.input !== undefined, 'input must be given');
     (0, assert_1.guard)(options.criterion !== undefined, 'a slicing criterion must be given');
+    const config = (0, config_1.getConfig)();
     await slicer.init(options['input-is-text']
         ? { request: 'text', content: options.input.replaceAll('\\n', '\n') }
-        : { request: 'file', content: options.input }, options['no-magic-comments'] ? auto_select_defaults_1.doNotAutoSelect : (0, magic_comments_1.makeMagicCommentHandler)(auto_select_defaults_1.doNotAutoSelect));
+        : { request: 'file', content: options.input }, config, options['no-magic-comments'] ? auto_select_defaults_1.doNotAutoSelect : (0, magic_comments_1.makeMagicCommentHandler)(auto_select_defaults_1.doNotAutoSelect));
     let mappedSlices = [];
     let reconstruct = undefined;
     const doSlicing = options.criterion.trim() !== '';
@@ -58,7 +60,7 @@ async function getSlice() {
     const { stats, normalize, parse, tokenMap, dataflow } = slicer.finish();
     const mappedCriteria = mappedSlices.map(c => `    ${c.criterion} => ${c.id} (${JSON.stringify(normalize.idMap.get(c.id)?.location)})`).join('\n');
     log_1.log.info(`Mapped criteria:\n${mappedCriteria}`);
-    const sliceStatsAsString = (0, print_1.stats2string)(await (0, process_1.summarizeSlicerStats)(stats));
+    const sliceStatsAsString = (0, print_1.stats2string)(await (0, process_1.summarizeSlicerStats)(stats, undefined, (0, config_1.getEngineConfig)(config, 'r-shell')));
     if (options.api) {
         const output = {
             tokenMap,

package/cli/statistics-app.js

@@ -2,6 +2,7 @@
 Object.defineProperty(exports, "__esModule", { value: true });
 const script_1 = require("./common/script");
 const statistics_core_1 = require("./script-core/statistics-core");
+const config_1 = require("../config");
 const scriptOptions = (0, script_1.processCommandLineArgs)('stats', [], {
     subtitle: 'Given input files or folders, this will collect usage statistics for the given features and write them to a file',
     examples: [
@@ -11,5 +12,5 @@ const scriptOptions = (0, script_1.processCommandLineArgs)('stats', [], {
         '{bold --help}'
     ]
 });
-void (0, statistics_core_1.flowrScriptGetStats)(scriptOptions);
+void (0, statistics_core_1.flowrScriptGetStats)(scriptOptions, (0, config_1.getConfig)());
 //# sourceMappingURL=statistics-app.js.map

package/cli/statistics-helper-app.js

@@ -2,6 +2,7 @@
 Object.defineProperty(exports, "__esModule", { value: true });
 const script_1 = require("./common/script");
 const statistics_helper_core_1 = require("./script-core/statistics-helper-core");
+const config_1 = require("../config");
 const scriptOptions = (0, script_1.processCommandLineArgs)('stats-helper', [], {
     subtitle: 'Given a single input file, this will collect usage statistics for the given features and write them to a file',
     examples: [
@@ -9,5 +10,5 @@ const scriptOptions = (0, script_1.processCommandLineArgs)('stats-helper', [], {
         '{bold --help}'
     ]
 });
-void (0, statistics_helper_core_1.getStatsForSingleFile)(scriptOptions);
+void (0, statistics_helper_core_1.getStatsForSingleFile)(scriptOptions, (0, config_1.getConfig)());
 //# sourceMappingURL=statistics-helper-app.js.map

package/config.d.ts

@@ -2,7 +2,7 @@ import type { MergeableRecord } from './util/objects';
 import Joi from 'joi';
 import type { BuiltInDefinitions } from './dataflow/environments/built-in-config';
 import type { KnownParser } from './r-bridge/parser';
-import type { DeepPartial } from 'ts-essentials';
+import type { DeepWritable } from 'ts-essentials';
 export declare enum VariableResolve {
     /** Don't resolve constants at all */
     Disabled = "disabled",
@@ -47,7 +47,7 @@ export interface FlowrLaxSourcingOptions extends MergeableRecord {
     /**
      * Additionally search in these paths
      */
-    readonly searchPath: readonly string[];
+    readonly searchPath: string[];
     /**
      * Allow to drop the first or all parts of the sourced path,
      * if it is relative.
@@ -77,7 +77,7 @@ export interface FlowrLaxSourcingOptions extends MergeableRecord {
      * - `foo bar.R` (original name)
      * - `main.R` (replaced with main.R)
      * - `foo_bar.R` (replaced spaces)
-     * - `foo-bar.R` (replaced spaces and oo)
+     * - `faa-bar.R` (replaced spaces and oo)
      */
     readonly applyReplacements?: Record<string, string>[];
 }
@@ -103,7 +103,7 @@ export interface FlowrConfigOptions extends MergeableRecord {
      * The engines to use for interacting with R code. Currently, supports {@link TreeSitterEngineConfig} and {@link RShellEngineConfig}.
      * An empty array means all available engines will be used.
      */
-    readonly engines: readonly EngineConfig[];
+    readonly engines: EngineConfig[];
     /**
      * The default engine to use for interacting with R code. If this is undefined, an arbitrary engine from {@link engines} will be used.
      */
@@ -145,6 +145,37 @@ export interface FlowrConfigOptions extends MergeableRecord {
             readonly threshold?: number;
         };
     };
+    /**
+     * Configuration options for abstract interpretation
+     */
+    readonly abstractInterpretation: {
+        /**
+         * The configuration of the shape inference for data frames
+         */
+        readonly dataFrame: {
+            /**
+             * The maximum number of columns names to infer for data frames before over-approximating the column names to top
+             */
+            readonly maxColNames: number;
+            /**
+             * The threshold for the number of visitations of a node at which widening should be performed to ensure the termination of the fixpoint iteration
+             */
+            readonly wideningThreshold: number;
+            /**
+             * Configuration options for reading data frame shapes from loaded external data files, such as CSV files
+             */
+            readonly readLoadedData: {
+                /**
+                 * Whether data frame shapes should be extracted from loaded external data files, such as CSV files
+                 */
+                readonly readExternalFiles: boolean;
+                /**
+                 * The maximum number of lines to read when extracting data frame shapes from loaded files, such as CSV files
+                 */
+                readonly maxReadLines: number;
+            };
+        };
+    };
 }
 export interface TreeSitterEngineConfig extends MergeableRecord {
     readonly type: 'tree-sitter';
@@ -174,12 +205,14 @@ export type KnownEngines = {
 };
 export declare const defaultConfigOptions: FlowrConfigOptions;
 export declare const flowrConfigFileSchema: Joi.ObjectSchema<any>;
-export declare function setConfigFile(file: string | undefined, workingDirectory?: string, forceLoad?: boolean): void;
 export declare function parseConfig(jsonString: string): FlowrConfigOptions | undefined;
-export declare function setConfig(config: FlowrConfigOptions): void;
-export declare function amendConfig(amendment: DeepPartial<FlowrConfigOptions>): void;
-export declare function getConfig(): FlowrConfigOptions;
-export declare function getEngineConfig<T extends EngineConfig['type']>(engine: T): EngineConfig & {
+/**
+ * Creates a new flowr config that has the updated values.
+ */
+export declare function amendConfig(config: FlowrConfigOptions, amendmentFunc: (config: DeepWritable<FlowrConfigOptions>) => FlowrConfigOptions): FlowrConfigOptions;
+export declare function cloneConfig(config: FlowrConfigOptions): FlowrConfigOptions;
+export declare function getConfig(configFile?: string, configWorkingDirectory?: string): FlowrConfigOptions;
+export declare function getEngineConfig<T extends EngineConfig['type']>(config: FlowrConfigOptions, engine: T): EngineConfig & {
     type: T;
 } | undefined;
-export declare function isOverPointerAnalysisThreshold(count: number): boolean;
+export declare function isOverPointerAnalysisThreshold(config: FlowrConfigOptions, count: number): boolean;

package/config.js

@@ -4,10 +4,9 @@ var __importDefault = (this && this.__importDefault) || function (mod) {
 };
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.flowrConfigFileSchema = exports.defaultConfigOptions = exports.DropPathsOption = exports.InferWorkingDirectory = exports.VariableResolve = void 0;
-exports.setConfigFile = setConfigFile;
 exports.parseConfig = parseConfig;
-exports.setConfig = setConfig;
 exports.amendConfig = amendConfig;
+exports.cloneConfig = cloneConfig;
 exports.getConfig = getConfig;
 exports.getEngineConfig = getEngineConfig;
 exports.isOverPointerAnalysisThreshold = isOverPointerAnalysisThreshold;
@@ -82,6 +81,16 @@ exports.defaultConfigOptions = {
         slicer: {
             threshold: 50
         }
+    },
+    abstractInterpretation: {
+        dataFrame: {
+            maxColNames: 50,
+            wideningThreshold: 4,
+            readLoadedData: {
+                readExternalFiles: true,
+                maxReadLines: 1e6
+            }
+        }
     }
 };
 exports.flowrConfigFileSchema = joi_1.default.object({
@@ -121,21 +130,18 @@ exports.flowrConfigFileSchema = joi_1.default.object({
         slicer: joi_1.default.object({
             threshold: joi_1.default.number().optional().description('The maximum number of iterations to perform on a single function call during slicing.')
         }).optional().description('The configuration for the slicer.')
-    }).description('How to resolve constants, constraints, cells, ...')
+    }).description('How to resolve constants, constraints, cells, ...'),
+    abstractInterpretation: joi_1.default.object({
+        dataFrame: joi_1.default.object({
+            maxColNames: joi_1.default.number().min(0).description('The maximum number of columns names to infer for data frames before over-approximating the column names to top.'),
+            wideningThreshold: joi_1.default.number().min(1).description('The threshold for the number of visitations of a node at which widening should be performed to ensure the termination of the fixpoint iteration.'),
+            readLoadedData: joi_1.default.object({
+                readExternalFiles: joi_1.default.boolean().description('Whether data frame shapes should be extracted from loaded external files, such as CSV files.'),
+                maxReadLines: joi_1.default.number().min(1).description('The maximum number of lines to read when extracting data frame shapes from loaded files, such as CSV files.')
+            }).description('Configuration options for reading data frame shapes from loaded external data files, such as CSV files.')
+        }).description('The configuration of the shape inference for data frames.')
+    }).description('The configuration options for abstract interpretation.')
 }).description('The configuration file format for flowR.');
-// we don't load from a config file at all by default unless setConfigFile is called
-let configFile = undefined;
-let configWorkingDirectory = process.cwd();
-let currentConfig = undefined;
-function setConfigFile(file, workingDirectory = process.cwd(), forceLoad = false) {
-    configFile = file;
-    configWorkingDirectory = workingDirectory;
-    // reset the config so it gets reloaded
-    currentConfig = undefined;
-    if (forceLoad) {
-        getConfig();
-    }
-}
 function parseConfig(jsonString) {
     try {
         const parsed = JSON.parse(jsonString);
@@ -153,46 +159,44 @@ function parseConfig(jsonString) {
         log_1.log.error(`Failed to parse config ${jsonString}: ${e.message}`);
     }
 }
-function setConfig(config) {
-    currentConfig = config;
+/**
+ * Creates a new flowr config that has the updated values.
+ */
+function amendConfig(config, amendmentFunc) {
+    return amendmentFunc(cloneConfig(config));
 }
-function amendConfig(amendment) {
-    setConfig((0, objects_1.deepMergeObject)(getConfig(), amendment));
-    log_1.log.trace(`Amending config with ${JSON.stringify(amendment)}, resulting in ${JSON.stringify(getConfig())}}`);
+function cloneConfig(config) {
+    return JSON.parse(JSON.stringify(config));
 }
-function getConfig() {
-    // lazy-load the config based on the current settings
-    if (currentConfig === undefined) {
-        try {
-            setConfig(loadConfigFromFile(configFile, configWorkingDirectory));
-        }
-        catch (e) {
-            log_1.log.error(`Failed to load config: ${e.message}`);
-            setConfig(exports.defaultConfigOptions);
-        }
+function getConfig(configFile, configWorkingDirectory = process.cwd()) {
+    try {
+        return loadConfigFromFile(configFile, configWorkingDirectory);
+    }
+    catch (e) {
+        log_1.log.error(`Failed to load config: ${e.message}`);
+        return exports.defaultConfigOptions;
     }
-    return currentConfig;
 }
-function getEngineConfig(engine) {
-    const config = getConfig().engines;
-    if (!config.length) {
+function getEngineConfig(config, engine) {
+    const engines = config.engines;
+    if (!engines.length) {
         return defaultEngineConfigs[engine];
     }
     else {
-        return config.find(e => e.type == engine);
+        return engines.find(e => e.type == engine);
     }
 }
-function getPointerAnalysisThreshold() {
-    const config = getConfig().solver.pointerTracking;
-    if (typeof config === 'object') {
-        return config.maxIndexCount;
+function getPointerAnalysisThreshold(config) {
+    const pointerTracking = config.solver.pointerTracking;
+    if (typeof pointerTracking === 'object') {
+        return pointerTracking.maxIndexCount;
     }
     else {
-        return config ? 'unlimited' : 'disabled';
+        return pointerTracking ? 'unlimited' : 'disabled';
     }
 }
-function isOverPointerAnalysisThreshold(count) {
-    const threshold = getPointerAnalysisThreshold();
+function isOverPointerAnalysisThreshold(config, count) {
+    const threshold = getPointerAnalysisThreshold(config);
     return threshold !== 'unlimited' && (threshold === 'disabled' || count > threshold);
 }
 function loadConfigFromFile(configFile, workingDirectory) {

package/control-flow/cfg-dead-code.js

@@ -10,6 +10,7 @@ const general_1 = require("../dataflow/eval/values/general");
 const r_value_1 = require("../dataflow/eval/values/r-value");
 class CfgConditionalDeadCodeRemoval extends semantic_cfg_guided_visitor_1.SemanticCfgGuidedVisitor {
     cachedConditions = new Map();
+    cachedStatements = new Map();
     getValue(id) {
         const has = this.cachedConditions.get(id);
         if (has) {
@@ -18,6 +19,14 @@ class CfgConditionalDeadCodeRemoval extends semantic_cfg_guided_visitor_1.Semant
         this.visitNode(id);
         return this.cachedConditions.get(id) ?? logic_1.Ternary.Maybe;
     }
+    isUnconditionalJump(id) {
+        const has = this.cachedStatements.get(id);
+        if (has) {
+            return has;
+        }
+        this.visitNode(id);
+        return this.cachedStatements.get(id) ?? false;
+    }
     unableToCalculateValue(id) {
         this.cachedConditions.set(id, logic_1.Ternary.Maybe);
     }
@@ -36,11 +45,16 @@ class CfgConditionalDeadCodeRemoval extends semantic_cfg_guided_visitor_1.Semant
                         this.config.controlFlow.graph.removeEdge(from, target);
                     }
                 }
+                else if (edge.label === 0 /* CfgEdgeType.Fd */) {
+                    if (this.isUnconditionalJump(target)) {
+                        this.config.controlFlow.graph.removeEdge(from, target);
+                    }
+                }
             }
         }
     }
     handleValuesFor(id, valueId) {
-        const values = (0, general_1.valueSetGuard)((0, alias_tracking_1.resolveIdToValue)(valueId, { graph: this.config.dfg, full: true, idMap: this.config.normalizedAst.idMap }));
+        const values = (0, general_1.valueSetGuard)((0, alias_tracking_1.resolveIdToValue)(valueId, { graph: this.config.dfg, full: true, idMap: this.config.normalizedAst.idMap, resolve: this.config.flowrConfig.solver.variables }));
         if (values === undefined || values.elements.length !== 1 || values.elements[0].type != 'logical' || !(0, r_value_1.isValue)(values.elements[0].value)) {
             this.unableToCalculateValue(id);
             return;
@@ -56,12 +70,40 @@ class CfgConditionalDeadCodeRemoval extends semantic_cfg_guided_visitor_1.Semant
         }
         this.handleValuesFor(id, typeof data.condition === 'object' ? data.condition.nodeId : data.condition);
     }
+    getBoolArgValue(data) {
+        if (data.call.args.length !== 1 || data.call.args[0] === r_function_call_1.EmptyArgument) {
+            return undefined;
+        }
+        const values = (0, general_1.valueSetGuard)((0, alias_tracking_1.resolveIdToValue)(data.call.args[0].nodeId, { graph: this.config.dfg, full: true, idMap: this.config.normalizedAst.idMap, resolve: this.config.flowrConfig.solver.variables }));
+        if (values === undefined || values.elements.length !== 1 || values.elements[0].type != 'logical' || !(0, r_value_1.isValue)(values.elements[0].value)) {
+            return undefined;
+        }
+        return Boolean(values.elements[0].value);
+    }
+    handleFunctionCall(data) {
+        switch (data.call.origin[0]) {
+            case 'builtin:return':
+            case 'builtin:stop':
+                this.cachedStatements.set(data.call.id, true);
+                break;
+            case 'builtin:stopifnot': {
+                const arg = this.getBoolArgValue(data);
+                if (arg !== undefined) {
+                    this.cachedStatements.set(data.call.id, !arg);
+                }
+                break;
+            }
+        }
+    }
     onIfThenElseCall(data) {
         this.handleWithCondition(data);
     }
     onWhileLoopCall(data) {
         this.handleWithCondition(data);
     }
+    onDefaultFunctionCall(data) {
+        this.handleFunctionCall(data);
+    }
 }
 /** Breaks unsatisfiable control dependencies */
 function cfgAnalyzeDeadCode(cfg, info) {
@@ -73,7 +115,8 @@ function cfgAnalyzeDeadCode(cfg, info) {
         controlFlow: cfg,
         normalizedAst: info.ast,
         dfg: info.dfg,
-        defaultVisitingOrder: 'forward'
+        flowrConfig: info.config,
+        defaultVisitingOrder: 'forward',
     });
     visitor.start();
     return cfg;

package/control-flow/cfg-simplification.d.ts

@@ -2,9 +2,11 @@ import type { ControlFlowInformation } from './control-flow-graph';
 import type { NormalizedAst } from '../r-bridge/lang-4.x/ast/model/processing/decorate';
 import type { DataflowGraph } from '../dataflow/graph/graph';
 import { cfgAnalyzeDeadCode } from './cfg-dead-code';
+import type { FlowrConfigOptions } from '../config';
 export interface CfgPassInfo {
     ast?: NormalizedAst;
     dfg?: DataflowGraph;
+    config: FlowrConfigOptions;
 }
 export type CfgSimplificationPass = (cfg: ControlFlowInformation, info: CfgPassInfo) => ControlFlowInformation;
 export declare const CfgSimplificationPasses: {

package/control-flow/control-flow-graph.d.ts

@@ -67,6 +67,8 @@ export interface CfgBasicBlockVertex extends CfgBaseVertex {
  */
 export type CfgSimpleVertex = CfgStatementVertex | CfgExpressionVertex | CfgBasicBlockVertex | CfgMidMarkerVertex | CfgEndMarkerVertex;
 export declare function equalVertex(a: CfgSimpleVertex, b: CfgSimpleVertex): boolean;
+export declare function isMarkerVertex(vertex: CfgSimpleVertex): vertex is CfgMidMarkerVertex | CfgEndMarkerVertex;
+export declare function getVertexRootId(vertex: CfgSimpleVertex): NodeId;
 interface CfgFlowDependencyEdge extends MergeableRecord {
     label: CfgEdgeType.Fd;
 }

package/control-flow/control-flow-graph.js

@@ -3,6 +3,8 @@ Object.defineProperty(exports, "__esModule", { value: true });
 exports.ControlFlowGraph = exports.CfgVertexType = void 0;
 exports.edgeTypeToString = edgeTypeToString;
 exports.equalVertex = equalVertex;
+exports.isMarkerVertex = isMarkerVertex;
+exports.getVertexRootId = getVertexRootId;
 exports.emptyControlFlowInformation = emptyControlFlowInformation;
 const assert_1 = require("../util/assert");
 var CfgVertexType;
@@ -43,6 +45,12 @@ function equalVertex(a, b) {
     }
     return true;
 }
+function isMarkerVertex(vertex) {
+    return vertex.type === CfgVertexType.MidMarker || vertex.type === CfgVertexType.EndMarker;
+}
+function getVertexRootId(vertex) {
+    return isMarkerVertex(vertex) ? vertex.root : vertex.id;
+}
 /**
  * This class represents the control flow graph of an R program.
  * The control flow may be hierarchical when confronted with function definitions (see {@link CfgSimpleVertex} and {@link CFG#rootVertexIds|rootVertexIds()}).

package/control-flow/dfg-cfg-guided-visitor.d.ts

@@ -1,4 +1,4 @@
-import type { CfgExpressionVertex, CfgStatementVertex, ControlFlowInformation } from './control-flow-graph';
+import { type CfgBasicBlockVertex, type CfgEndMarkerVertex, type CfgExpressionVertex, type CfgSimpleVertex, type CfgStatementVertex, type ControlFlowInformation } from './control-flow-graph';
 import type { NodeId } from '../r-bridge/lang-4.x/ast/model/processing/node-id';
 import type { DataflowGraphVertexArgument, DataflowGraphVertexFunctionCall, DataflowGraphVertexFunctionDefinition, DataflowGraphVertexUse, DataflowGraphVertexValue, DataflowGraphVertexVariableDefinition } from '../dataflow/graph/vertex';
 import type { BasicCfgGuidedVisitorConfiguration } from './basic-cfg-guided-visitor';
@@ -6,6 +6,7 @@ import { BasicCfgGuidedVisitor } from './basic-cfg-guided-visitor';
 import type { DataflowGraph } from '../dataflow/graph/graph';
 export interface DataflowCfgGuidedVisitorConfiguration<ControlFlow extends ControlFlowInformation = ControlFlowInformation, Dfg extends DataflowGraph = DataflowGraph> extends BasicCfgGuidedVisitorConfiguration<ControlFlow> {
     readonly dfg: Dfg;
+    readonly defaultVisitingType?: 'entry' | 'exit';
 }
 /**
  * This visitor extends on the {@link BasicCfgGuidedVisitor} by dispatching visitors based on the dataflow graph.
@@ -19,11 +20,12 @@ export declare class DataflowAwareCfgGuidedVisitor<ControlFlow extends ControlFl
     protected getDataflowGraph(id: NodeId): DataflowGraphVertexArgument | undefined;
     protected onStatementNode(node: CfgStatementVertex): void;
     protected onExpressionNode(node: CfgExpressionVertex): void;
-    private onExprOrStmtNode;
+    protected onEndMarkerNode(node: CfgEndMarkerVertex): void;
+    protected visitDataflowNode(node: Exclude<CfgSimpleVertex, CfgBasicBlockVertex>): void;
     /**
      * called for every cfg vertex that has no corresponding dataflow vertex.
      */
-    protected visitUnknown(_vertex: CfgStatementVertex | CfgExpressionVertex): void;
+    protected visitUnknown(_vertex: Exclude<CfgSimpleVertex, CfgBasicBlockVertex>): void;
     protected visitValue(_val: DataflowGraphVertexValue): void;
     protected visitVariableUse(_use: DataflowGraphVertexUse): void;
     protected visitVariableDefinition(_def: DataflowGraphVertexVariableDefinition): void;

package/control-flow/dfg-cfg-guided-visitor.js

@@ -1,6 +1,7 @@
 "use strict";
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.DataflowAwareCfgGuidedVisitor = void 0;
+const control_flow_graph_1 = require("./control-flow-graph");
 const vertex_1 = require("../dataflow/graph/vertex");
 const basic_cfg_guided_visitor_1 = require("./basic-cfg-guided-visitor");
 const assert_1 = require("../util/assert");
@@ -18,14 +19,24 @@ class DataflowAwareCfgGuidedVisitor extends basic_cfg_guided_visitor_1.BasicCfgG
     }
     onStatementNode(node) {
         super.onStatementNode(node);
-        this.onExprOrStmtNode(node);
+        if (this.config.defaultVisitingType !== 'exit' || node.end === undefined) {
+            this.visitDataflowNode(node);
+        }
     }
     onExpressionNode(node) {
         super.onExpressionNode(node);
-        this.onExprOrStmtNode(node);
+        if (this.config.defaultVisitingType !== 'exit' || node.end === undefined) {
+            this.visitDataflowNode(node);
+        }
+    }
+    onEndMarkerNode(node) {
+        super.onEndMarkerNode(node);
+        if (this.config.defaultVisitingType === 'exit') {
+            this.visitDataflowNode(node);
+        }
     }
-    onExprOrStmtNode(node) {
-        const dfgVertex = this.getDataflowGraph(node.id);
+    visitDataflowNode(node) {
+        const dfgVertex = this.getDataflowGraph((0, control_flow_graph_1.getVertexRootId)(node));
         if (!dfgVertex) {
             this.visitUnknown(node);
             return;

package/control-flow/extract-cfg.d.ts

@@ -3,18 +3,20 @@ import type { NormalizedAst, ParentInformation } from '../r-bridge/lang-4.x/ast/
 import type { DataflowGraph } from '../dataflow/graph/graph';
 import type { ControlFlowInformation } from './control-flow-graph';
 import type { CfgSimplificationPassName } from './cfg-simplification';
+import type { FlowrConfigOptions } from '../config';
 /**
  * Given a normalized AST this approximates the control flow graph of the program.
- * This few is different from the computation of the dataflow graph and may differ,
+ * This view is different from the computation of the dataflow graph and may differ,
  * especially because it focuses on intra-procedural analysis.
  *
  * @param ast             - the normalized AST
+ * @param config          - the Flowr config
  * @param graph           - additional dataflow facts to consider by the control flow extraction
  * @param simplifications - a list of simplification passes to apply to the control flow graph
  *
  * @see {@link extractSimpleCfg} - for a simplified version of this function
  */
-export declare function extractCfg<Info = ParentInformation>(ast: NormalizedAst<Info & ParentInformation>, graph?: DataflowGraph, simplifications?: readonly CfgSimplificationPassName[]): ControlFlowInformation;
+export declare function extractCfg<Info = ParentInformation>(ast: NormalizedAst<Info & ParentInformation>, config: FlowrConfigOptions, graph?: DataflowGraph, simplifications?: readonly CfgSimplificationPassName[]): ControlFlowInformation;
 /**
  * Simplified version of {@link extractCfg} that is much quicker, but much simpler!
  */

package/control-flow/extract-cfg.js

@@ -53,17 +53,18 @@ function dataflowCfgFolds(dataflowGraph) {
 }
 /**
  * Given a normalized AST this approximates the control flow graph of the program.
- * This few is different from the computation of the dataflow graph and may differ,
+ * This view is different from the computation of the dataflow graph and may differ,
  * especially because it focuses on intra-procedural analysis.
  *
  * @param ast             - the normalized AST
+ * @param config          - the Flowr config
  * @param graph           - additional dataflow facts to consider by the control flow extraction
  * @param simplifications - a list of simplification passes to apply to the control flow graph
  *
  * @see {@link extractSimpleCfg} - for a simplified version of this function
  */
-function extractCfg(ast, graph, simplifications) {
-    return (0, cfg_simplification_1.simplifyControlFlowInformation)((0, fold_1.foldAst)(ast.ast, graph ? dataflowCfgFolds(graph) : cfgFolds), { ast, dfg: graph }, simplifications);
+function extractCfg(ast, config, graph, simplifications) {
+    return (0, cfg_simplification_1.simplifyControlFlowInformation)((0, fold_1.foldAst)(ast.ast, graph ? dataflowCfgFolds(graph) : cfgFolds), { ast, dfg: graph, config }, simplifications);
 }
 /**
  * Simplified version of {@link extractCfg} that is much quicker, but much simpler!

package/control-flow/semantic-cfg-guided-visitor.d.ts

@@ -14,7 +14,9 @@ import type { NoInfo, RNode } from '../r-bridge/lang-4.x/ast/model/model';
 import type { RSymbol } from '../r-bridge/lang-4.x/ast/model/nodes/r-symbol';
 import type { BuiltInProcessorMapper } from '../dataflow/environments/built-in';
 import type { RExpressionList } from '../r-bridge/lang-4.x/ast/model/nodes/r-expression-list';
+import type { FlowrConfigOptions } from '../config';
 export interface SemanticCfgGuidedVisitorConfiguration<OtherInfo = NoInfo, ControlFlow extends ControlFlowInformation = ControlFlowInformation, Ast extends NormalizedAst<OtherInfo> = NormalizedAst<OtherInfo>, Dfg extends DataflowGraph = DataflowGraph> extends DataflowCfgGuidedVisitorConfiguration<ControlFlow, Dfg>, SyntaxCfgGuidedVisitorConfiguration<OtherInfo, ControlFlow, Ast> {
+    readonly flowrConfig: FlowrConfigOptions;
 }
 /**
  * This visitor extends on the {@link DataflowAwareCfgGuidedVisitor} by dispatching visitors for separate function calls as well,
@@ -43,7 +45,7 @@ export declare class SemanticCfgGuidedVisitor<OtherInfo = NoInfo, ControlFlow ex
     /**
      * A helper function to get the normalized AST node for the given id or fail if it does not exist.
      */
-    protected getNormalizedAst(id: NodeId): RNode<OtherInfo & ParentInformation> | undefined;
+    protected getNormalizedAst(id: NodeId | undefined): RNode<OtherInfo & ParentInformation> | undefined;
     /**
      * See {@link DataflowAwareCfgGuidedVisitor#visitValue} for the base implementation.
      * This now dispatches the value to the appropriate event handler based on its type.
@@ -126,7 +128,11 @@ export declare class SemanticCfgGuidedVisitor<OtherInfo = NoInfo, ControlFlow ex
      * A helper function to request the {@link getOriginInDfg|origins} of the given node.
      */
     protected getOrigins(id: NodeId): Origin[] | undefined;
-    /** Called for every occurrence of a `NULL` in the program. */
+    /**
+     * Called for every occurrence of a `NULL` in the program.
+     *
+     * For other symbols that are not referenced as a variable, see {@link SemanticCfgGuidedVisitor#onSymbolConstant|`onSymbolConstant`}.
+     */
     protected onNullConstant(_data: {
         vertex: DataflowGraphVertexValue;
         node: RSymbol<OtherInfo & ParentInformation, 'NULL'>;
@@ -158,6 +164,18 @@ export declare class SemanticCfgGuidedVisitor<OtherInfo = NoInfo, ControlFlow ex
         vertex: DataflowGraphVertexValue;
         node: RLogical;
     }): void;
+    /**
+     * Called for every constant symbol value in the program.
+     *
+     * For example, `foo` in `library(foo)` or `a` in `l$a`. This most likely happens as part of non-standard-evaluation, i.e., the symbol is not evaluated to a value,
+     * but used as a symbol in and of itself.
+     *
+     * Please note, that due to its special behaviors, `NULL` is handled in {@link SemanticCfgGuidedVisitor#onNullConstant|`onNullConstant`} and not here.
+     */
+    protected onSymbolConstant(_data: {
+        vertex: DataflowGraphVertexValue;
+        node: RSymbol;
+    }): void;
     /**
      * Called for every variable that is read within the program.
      * You can use {@link getOrigins} to get the origins of the variable.