@eagleoutice/flowr 2.2.11 → 2.2.12

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 v2.2.10, R v4.4.0 (r-shell engine)
+    flowR repl using flowR v2.2.11, R v4.4.3 (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!">119.1 ms</span></i> (as of Feb 23, 2025), 
+  Within just <i><span title="This measurement is automatically fetched from the latest benchmark!">117.9 ms</span></i> (as of Mar 2, 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 v2.2.10, R v4.4.0 (r-shell engine)
+    flowR repl using flowR v2.2.11, R v4.4.3 (r-shell engine)
     R> :dataflow* test/testfiles/example.R
     ```
     
@@ -377,7 +377,7 @@ It offers a wide variety of features, for example:
     ```
     
     	
-    (The analysis required _21.37 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 _22.14 ms_ (including parse and normalize, using the [r-shell](https://github.com/flowr-analysis/flowr/wiki/Engines) engine) within the generation environment.)
     
     
     

package/benchmark/slicer.d.ts

@@ -43,6 +43,13 @@ export interface BenchmarkSingleSliceStats extends MergeableRecord {
     /** the final code, as the result of the 'reconstruct' step */
     code: ReconstructionResult;
 }
+/**
+ * The type of sampling strategy to use when slicing all possible variables.
+ *
+ * - `'random'`: Randomly select the given number of slicing criteria.
+ * - `'equidistant'`: Select the given number of slicing criteria in an equidistant manner.
+ */
+export type SamplingStrategy = 'random' | 'equidistant';
 export declare class BenchmarkSlicer {
     /** Measures all data recorded *once* per slicer (complete setup up to the dataflow graph creation) */
     private readonly commonMeasurements;
@@ -62,38 +69,58 @@ export declare class BenchmarkSlicer {
      * Initialize the slicer on the given request.
      * Can only be called once for each instance.
      */
-    init(request: RParseRequestFromFile | RParseRequestFromText, autoSelectIf?: AutoSelectPredicate): Promise<void>;
+    init(request: RParseRequestFromFile | RParseRequestFromText, autoSelectIf?: AutoSelectPredicate, threshold?: number): Promise<void>;
     private calculateStatsAfterInit;
     /**
-   * Slice for the given {@link SlicingCriteria}.
-   * @see SingleSlicingCriterion
-   *
-   * @returns The per slice stats retrieved for this slicing criteria
-   */
+     * Counts the number of stored indices in the dataflow graph created by the pointer analysis.
+     */
+    private countStoredVertexIndices;
+    /**
+     * Counts the number of stored indices in the dataflow graph created by the pointer analysis.
+     */
+    private countStoredEnvIndices;
+    /**
+     * Counts the number of stored indices in the passed definitions.
+     */
+    private countStoredIndices;
+    /**
+     * Recursively counts the number of indices and sub-indices in the given collection.
+     */
+    private countIndices;
+    /**
+     * Slice for the given {@link SlicingCriteria}.
+     * @see SingleSlicingCriterion
+     *
+     * @returns The per slice stats retrieved for this slicing criteria
+     */
     slice(...slicingCriteria: SlicingCriteria): Promise<BenchmarkSingleSliceStats>;
     /** Bridging the gap between the new internal and the old names for the benchmarking */
     private measureCommonStep;
     private measureSliceStep;
     private guardActive;
     /**
-   * Call {@link slice} for all slicing criteria that match the given filter.
-   * See {@link collectAllSlicingCriteria} for details.
-   * <p>
-   * the `report` function will be called *before* each *individual* slice is performed.
-   *
-   * @returns The number of slices that were produced
-   *
-   * @see collectAllSlicingCriteria
-   * @see SlicingCriteriaFilter
-   */
-    sliceForAll(filter: SlicingCriteriaFilter, report?: (current: number, total: number, allCriteria: SlicingCriteria[]) => void, sampleRandom?: number): Promise<number>;
+     * Call {@link slice} for all slicing criteria that match the given filter.
+     * See {@link collectAllSlicingCriteria} for details.
+     * <p>
+     * the `report` function will be called *before* each *individual* slice is performed.
+     *
+     * @returns The number of slices that were produced
+     *
+     * @see collectAllSlicingCriteria
+     * @see SlicingCriteriaFilter
+     */
+    sliceForAll(filter: SlicingCriteriaFilter, report?: (current: number, total: number, allCriteria: SlicingCriteria[]) => void, options?: {
+        sampleCount?: number;
+        maxSliceCount?: number;
+        sampleStrategy?: SamplingStrategy;
+    }): Promise<number>;
     /**
-   * Retrieves the final stats and closes the shell session.
-   * Can be called multiple times to retrieve the stored stats, but will only close the session once (the first time).
-   */
+     * Retrieves the final stats and closes the shell session.
+     * Can be called multiple times to retrieve the stored stats, but will only close the session once (the first time).
+     */
     finish(): BenchmarkSlicerStats;
     /**
-   * Only call in case of an error - if the session must be closed and the benchmark itself is to be considered failed/dead.
-   */
+     * Only call in case of an error - if the session must be closed and the benchmark itself is to be considered failed/dead.
+     */
     ensureSessionClosed(): void;
 }

package/benchmark/slicer.js

@@ -22,6 +22,8 @@ const size_of_1 = require("./stats/size-of");
 const shell_1 = require("../r-bridge/shell");
 const tree_sitter_types_1 = require("../r-bridge/lang-4.x/tree-sitter/tree-sitter-types");
 const tree_sitter_executor_1 = require("../r-bridge/lang-4.x/tree-sitter/tree-sitter-executor");
+const vertex_1 = require("../dataflow/graph/vertex");
+const arrays_1 = require("../util/arrays");
 /**
  * The logger to be used for benchmarking as a global object.
  */
@@ -49,7 +51,7 @@ class BenchmarkSlicer {
      * Initialize the slicer on the given request.
      * Can only be called once for each instance.
      */
-    async init(request, autoSelectIf) {
+    async init(request, autoSelectIf, threshold) {
         (0, assert_1.guard)(this.stats === undefined, 'cannot initialize the slicer twice');
         // we know these are in sync so we just cast to one of them
         this.parser = await this.commonMeasurements.measure('initialize R session', async () => {
@@ -64,7 +66,8 @@ class BenchmarkSlicer {
         this.executor = (0, default_pipelines_1.createSlicePipeline)(this.parser, {
             request: { ...request },
             criterion: [],
-            autoSelectIf
+            autoSelectIf,
+            threshold,
         });
         this.loadedXml = (await this.measureCommonStep('parse', 'retrieve AST from R code')).parsed;
         this.normalizedAst = await this.measureCommonStep('normalize', 'normalize R AST');
@@ -127,6 +130,9 @@ class BenchmarkSlicer {
             }
             return false;
         });
+        const storedVertexIndices = this.countStoredVertexIndices();
+        const storedEnvIndices = this.countStoredEnvIndices();
+        const overwrittenIndices = storedVertexIndices - storedEnvIndices;
         const split = loadedContent.split('\n');
         const nonWhitespace = (0, strings_1.withoutWhitespace)(loadedContent).length;
         this.stats = {
@@ -150,7 +156,10 @@ class BenchmarkSlicer {
                 numberOfEdges: numberOfEdges,
                 numberOfCalls: numberOfCalls,
                 numberOfFunctionDefinitions: numberOfDefinitions,
-                sizeOfObject: (0, size_of_1.getSizeOfDfGraph)(this.dataflow.graph)
+                sizeOfObject: (0, size_of_1.getSizeOfDfGraph)(this.dataflow.graph),
+                storedVertexIndices: storedVertexIndices,
+                storedEnvIndices: storedEnvIndices,
+                overwrittenIndices: overwrittenIndices,
             },
             // these are all properly initialized in finish()
             commonMeasurements: new Map(),
@@ -161,11 +170,52 @@ class BenchmarkSlicer {
         };
     }
     /**
-   * Slice for the given {@link SlicingCriteria}.
-   * @see SingleSlicingCriterion
-   *
-   * @returns The per slice stats retrieved for this slicing criteria
-   */
+     * Counts the number of stored indices in the dataflow graph created by the pointer analysis.
+     */
+    countStoredVertexIndices() {
+        return this.countStoredIndices(this.dataflow?.out.map(ref => ref) ?? []);
+    }
+    /**
+     * Counts the number of stored indices in the dataflow graph created by the pointer analysis.
+     */
+    countStoredEnvIndices() {
+        return this.countStoredIndices(this.dataflow?.environment.current.memory.values()
+            ?.flatMap(def => def)
+            .map(def => def) ?? []);
+    }
+    /**
+     * Counts the number of stored indices in the passed definitions.
+     */
+    countStoredIndices(definitions) {
+        let numberOfIndices = 0;
+        for (const reference of definitions) {
+            if (reference.indicesCollection) {
+                numberOfIndices += this.countIndices(reference.indicesCollection);
+            }
+        }
+        return numberOfIndices;
+    }
+    /**
+     * Recursively counts the number of indices and sub-indices in the given collection.
+     */
+    countIndices(collection) {
+        let numberOfIndices = 0;
+        for (const indices of collection ?? []) {
+            for (const index of indices.indices) {
+                numberOfIndices++;
+                if ((0, vertex_1.isParentContainerIndex)(index)) {
+                    numberOfIndices += this.countIndices(index.subIndices);
+                }
+            }
+        }
+        return numberOfIndices;
+    }
+    /**
+     * Slice for the given {@link SlicingCriteria}.
+     * @see SingleSlicingCriterion
+     *
+     * @returns The per slice stats retrieved for this slicing criteria
+     */
     async slice(...slicingCriteria) {
         exports.benchmarkLogger.trace(`try to slice for criteria ${JSON.stringify(slicingCriteria)}`);
         this.guardActive();
@@ -227,23 +277,33 @@ class BenchmarkSlicer {
         (0, assert_1.guard)(this.stats !== undefined && !this.finished, 'need to call init before, and can not do after finish!');
     }
     /**
-   * Call {@link slice} for all slicing criteria that match the given filter.
-   * See {@link collectAllSlicingCriteria} for details.
-   * <p>
-   * the `report` function will be called *before* each *individual* slice is performed.
-   *
-   * @returns The number of slices that were produced
-   *
-   * @see collectAllSlicingCriteria
-   * @see SlicingCriteriaFilter
-   */
-    async sliceForAll(filter, report = () => { }, sampleRandom = -1) {
+     * Call {@link slice} for all slicing criteria that match the given filter.
+     * See {@link collectAllSlicingCriteria} for details.
+     * <p>
+     * the `report` function will be called *before* each *individual* slice is performed.
+     *
+     * @returns The number of slices that were produced
+     *
+     * @see collectAllSlicingCriteria
+     * @see SlicingCriteriaFilter
+     */
+    async sliceForAll(filter, report = () => { }, options = {}) {
+        const { sampleCount, maxSliceCount, sampleStrategy } = { sampleCount: -1, maxSliceCount: -1, sampleStrategy: 'random', ...options };
         this.guardActive();
         let count = 0;
-        const allCriteria = [...(0, collect_all_1.collectAllSlicingCriteria)(this.normalizedAst.ast, filter)];
-        if (sampleRandom > 0) {
-            allCriteria.sort(() => Math.random() - 0.5);
-            allCriteria.length = Math.min(allCriteria.length, sampleRandom);
+        let allCriteria = [...(0, collect_all_1.collectAllSlicingCriteria)(this.normalizedAst.ast, filter)];
+        // Cancel slicing if the number of slices exceeds the limit
+        if (maxSliceCount > 0 && allCriteria.length > maxSliceCount) {
+            return -allCriteria.length;
+        }
+        if (sampleCount > 0) {
+            if (sampleStrategy === 'equidistant') {
+                allCriteria = (0, arrays_1.equidistantSampling)(allCriteria, sampleCount, 'ceil');
+            }
+            else {
+                allCriteria.sort(() => Math.random() - 0.5);
+                allCriteria.length = Math.min(allCriteria.length, sampleCount);
+            }
         }
         for (const slicingCriteria of allCriteria) {
             report(count, allCriteria.length, allCriteria);
@@ -253,9 +313,9 @@ class BenchmarkSlicer {
         return count;
     }
     /**
-   * Retrieves the final stats and closes the shell session.
-   * Can be called multiple times to retrieve the stored stats, but will only close the session once (the first time).
-   */
+     * Retrieves the final stats and closes the shell session.
+     * Can be called multiple times to retrieve the stored stats, but will only close the session once (the first time).
+     */
     finish() {
         (0, assert_1.guard)(this.stats !== undefined, 'need to call init before finish');
         if (!this.finished) {
@@ -291,8 +351,8 @@ class BenchmarkSlicer {
         };
     }
     /**
-   * Only call in case of an error - if the session must be closed and the benchmark itself is to be considered failed/dead.
-   */
+     * Only call in case of an error - if the session must be closed and the benchmark itself is to be considered failed/dead.
+     */
     ensureSessionClosed() {
         this.parser?.close();
     }

package/benchmark/stats/print.js

@@ -140,11 +140,14 @@ Input:
   Normalized R tokens (w/o comments):  ${pad(stats.input.numberOfNormalizedTokensNoComments)}
 
 Dataflow:
-  Number of nodes:            ${pad(stats.dataflow.numberOfNodes)}
-  Number of edges:            ${pad(stats.dataflow.numberOfEdges)}
-  Number of calls:            ${pad(stats.dataflow.numberOfCalls)}
-  Number of function defs:    ${pad(stats.dataflow.numberOfFunctionDefinitions)}
-  Size of graph:              ${convertNumberToNiceBytes(stats.dataflow.sizeOfObject)}`;
+  Number of nodes:               ${pad(stats.dataflow.numberOfNodes)}
+  Number of edges:               ${pad(stats.dataflow.numberOfEdges)}
+  Number of calls:               ${pad(stats.dataflow.numberOfCalls)}
+  Number of function defs:       ${pad(stats.dataflow.numberOfFunctionDefinitions)}
+  Number of stored Vtx indices:  ${pad(stats.dataflow.storedVertexIndices)}
+  Number of stored Env indices:  ${pad(stats.dataflow.storedEnvIndices)}
+  Number of overwritten indices: ${pad(stats.dataflow.overwrittenIndices)}
+  Size of graph:                 ${convertNumberToNiceBytes(stats.dataflow.sizeOfObject)}`;
 }
 function ultimateStats2String(stats) {
     const slice = stats.totalSlices > 0 ? `Slice summary for:
@@ -195,11 +198,14 @@ Input:
   Normalized R tokens (w/o comments):  ${formatSummarizedMeasure(stats.input.numberOfNormalizedTokensNoComments)}
 
 Dataflow:
-  Number of nodes:            ${formatSummarizedMeasure(stats.dataflow.numberOfNodes)}
-  Number of edges:            ${formatSummarizedMeasure(stats.dataflow.numberOfEdges)}
-  Number of calls:            ${formatSummarizedMeasure(stats.dataflow.numberOfCalls)}
-  Number of function defs:    ${formatSummarizedMeasure(stats.dataflow.numberOfFunctionDefinitions)}
-  Size of graph:              ${formatSummarizedMeasure(stats.dataflow.sizeOfObject, convertNumberToNiceBytes)}
+  Number of nodes:               ${formatSummarizedMeasure(stats.dataflow.numberOfNodes)}
+  Number of edges:               ${formatSummarizedMeasure(stats.dataflow.numberOfEdges)}
+  Number of calls:               ${formatSummarizedMeasure(stats.dataflow.numberOfCalls)}
+  Number of function defs:       ${formatSummarizedMeasure(stats.dataflow.numberOfFunctionDefinitions)}
+  Number of stored Vtx indices:  ${formatSummarizedMeasure(stats.dataflow.storedVertexIndices)}
+  Number of stored Env indices:  ${formatSummarizedMeasure(stats.dataflow.storedEnvIndices)}
+  Number of overwritten indices: ${formatSummarizedMeasure(stats.dataflow.overwrittenIndices)}
+  Size of graph:                 ${formatSummarizedMeasure(stats.dataflow.sizeOfObject, convertNumberToNiceBytes)}
 `;
 }
 function reduction2String(title, reduction) {

package/benchmark/stats/size-of.js

@@ -65,6 +65,23 @@ function getSizeOfDfGraph(df) {
         });
         verts.push(vertex);
     }
-    return (0, object_sizeof_1.default)([...verts, ...df.edges()]);
+    return safeSizeOf([...verts, ...df.edges()]);
+}
+/**
+ * Calculates the size of an array in bytes.
+ *
+ * @param array - The array to calculate the size of.
+ * @returns The size of the array in bytes.
+ */
+function safeSizeOf(array) {
+    const size = (0, object_sizeof_1.default)(array);
+    if (typeof size === 'number') {
+        return size;
+    }
+    // the sizeOf method returns an error object, when the size could not be calculated
+    // in this case, we split the array in half and calculate the size of each half recursively
+    const chunkSize = Math.ceil(array.length / 2);
+    // subtract 1, because of the separate stringification of the array
+    return safeSizeOf(array.slice(0, chunkSize)) + safeSizeOf(array.slice(chunkSize)) - 1;
 }
 //# sourceMappingURL=size-of.js.map

package/benchmark/stats/stats.d.ts

@@ -37,6 +37,9 @@ export interface SlicerStatsDataflow<T = number> {
     numberOfCalls: T;
     numberOfFunctionDefinitions: T;
     sizeOfObject: T;
+    storedVertexIndices: T;
+    storedEnvIndices: T;
+    overwrittenIndices: T;
 }
 /**
  * Please note, that these measurement can be negative as there is no guarantee that the memory usage will increase

package/benchmark/summarizer/second-phase/process.js

@@ -85,7 +85,10 @@ function summarizeAllSummarizedStats(stats) {
             numberOfFunctionDefinitions: (0, summarizer_1.summarizeMeasurement)(dataflows.map(d => d.numberOfFunctionDefinitions)),
             numberOfCalls: (0, summarizer_1.summarizeMeasurement)(dataflows.map(d => d.numberOfCalls)),
             numberOfEdges: (0, summarizer_1.summarizeMeasurement)(dataflows.map(d => d.numberOfEdges)),
-            sizeOfObject: (0, summarizer_1.summarizeMeasurement)(dataflows.map(d => d.sizeOfObject))
+            sizeOfObject: (0, summarizer_1.summarizeMeasurement)(dataflows.map(d => d.sizeOfObject)),
+            storedVertexIndices: (0, summarizer_1.summarizeMeasurement)(dataflows.map(d => d.storedVertexIndices)),
+            storedEnvIndices: (0, summarizer_1.summarizeMeasurement)(dataflows.map(d => d.storedEnvIndices)),
+            overwrittenIndices: (0, summarizer_1.summarizeMeasurement)(dataflows.map(d => d.overwrittenIndices)),
         }
     };
 }
@@ -125,7 +128,10 @@ function summarizeAllUltimateStats(stats) {
             numberOfFunctionDefinitions: (0, process_1.summarizeSummarizedMeasurement)(stats.map(s => s.dataflow.numberOfFunctionDefinitions)),
             numberOfCalls: (0, process_1.summarizeSummarizedMeasurement)(stats.map(s => s.dataflow.numberOfCalls)),
             numberOfEdges: (0, process_1.summarizeSummarizedMeasurement)(stats.map(s => s.dataflow.numberOfEdges)),
-            sizeOfObject: (0, process_1.summarizeSummarizedMeasurement)(stats.map(s => s.dataflow.sizeOfObject))
+            sizeOfObject: (0, process_1.summarizeSummarizedMeasurement)(stats.map(s => s.dataflow.sizeOfObject)),
+            storedVertexIndices: (0, process_1.summarizeSummarizedMeasurement)(stats.map(s => s.dataflow.storedVertexIndices)),
+            storedEnvIndices: (0, process_1.summarizeSummarizedMeasurement)(stats.map(s => s.dataflow.storedEnvIndices)),
+            overwrittenIndices: (0, process_1.summarizeSummarizedMeasurement)(stats.map(s => s.dataflow.overwrittenIndices)),
         }
     };
 }

package/cli/benchmark-app.d.ts

@@ -9,4 +9,9 @@ export interface BenchmarkCliOptions {
     limit?: number;
     runs?: number;
     parser: KnownParserName;
+    'enable-pointer-tracking': boolean;
+    'max-file-slices': number;
+    threshold?: number;
+    'per-file-time-limit'?: number;
+    'sampling-strategy': string;
 }

package/cli/benchmark-app.js

@@ -42,10 +42,17 @@ async function benchmark() {
     console.log(`Using ${options.parallel} parallel executors`);
     // we do not use the limit argument to be able to pick the limit randomly
     const files = [];
-    for (const input of options.input) {
-        for await (const file of (0, files_1.allRFiles)(input)) {
-            files.push({ request: file, baseDir: input });
-        }
+    const firstFile = options.input[0];
+    // Check whether input is single JSON file containing a list of paths
+    if (options.input.length === 1 && fs_1.default.statSync(firstFile).isFile() && firstFile.endsWith('.json')) {
+        console.log('Input is a single JSON file. Assuming it contains a list of files to process');
+        const content = fs_1.default.readFileSync(firstFile, 'utf8');
+        const paths = JSON.parse(content);
+        const baseDir = findCommonBaseDir(paths);
+        await collectFiles(files, paths, () => baseDir);
+    }
+    else {
+        await collectFiles(files, options.input, (p) => p);
     }
     if (options.limit) {
         log_1.log.info(`limiting to ${options.limit} files`);
@@ -59,18 +66,54 @@ async function benchmark() {
         '--file-id', `${i}`,
         '--output', path_1.default.join(options.output, path_1.default.relative(f.baseDir, `${f.request.content}.json`)),
         '--slice', options.slice, ...verboseAdd,
-        '--parser', options.parser
+        '--parser', options.parser,
+        ...(options['enable-pointer-tracking'] ? ['--enable-pointer-tracking'] : []),
+        '--max-slices', `${options['max-file-slices']}`,
+        ...(options.threshold ? ['--threshold', `${options.threshold}`] : []),
+        '--sampling-strategy', options['sampling-strategy'],
     ]);
     const runs = options.runs ?? 1;
     for (let i = 1; i <= runs; i++) {
         console.log(`Run ${i} of ${runs}`);
         const pool = new parallel_1.LimitedThreadPool(`${__dirname}/benchmark-helper-app`, 
         // we reverse here "for looks", since the helper pops from the end, and we want file ids to be ascending :D
-        args.map(a => [...a, '--run-num', `${i}`]).reverse(), limit, options.parallel);
+        args.map(a => [...a, '--run-num', `${i}`]).reverse(), limit, options.parallel, options['per-file-time-limit']);
         await pool.run();
         const stats = pool.getStats();
         console.log(`Run ${i} of ${runs}: Benchmarked ${stats.counter} files, skipped ${stats.skipped.length} files due to errors`);
     }
 }
+/**
+ * Collect all R files from the given paths.
+ *
+ * @param files - list of files to append to
+ * @param paths - list of paths to search for R files
+ * @param getBaseDir - function to get the base directory of a path
+ */
+async function collectFiles(files, paths, getBaseDir) {
+    for (const input of paths) {
+        for await (const file of (0, files_1.allRFiles)(input)) {
+            files.push({ request: file, baseDir: getBaseDir(input) });
+        }
+    }
+}
+/**
+ * Find the common base directory of a list of paths.
+ *
+ * @param paths - list of paths
+ * @returns the common base directory
+ */
+function findCommonBaseDir(paths) {
+    const baseDirs = paths.map(f => path_1.default.dirname(f));
+    return baseDirs.reduce((acc, dir) => {
+        const split = dir.split(path_1.default.sep);
+        const accSplit = acc.split(path_1.default.sep);
+        let i = 0;
+        while (i < split.length && i < accSplit.length && split[i] === accSplit[i]) {
+            i++;
+        }
+        return split.slice(0, i).join(path_1.default.sep);
+    }, baseDirs[0]);
+}
 void benchmark();
 //# sourceMappingURL=benchmark-app.js.map

package/cli/benchmark-helper-app.d.ts

@@ -8,4 +8,8 @@ export interface SingleBenchmarkCliOptions {
     slice: string;
     output?: string;
     parser: KnownParserName;
+    'enable-pointer-tracking': boolean;
+    'max-slices': number;
+    threshold?: number;
+    'sampling-strategy': string;
 }

package/cli/benchmark-helper-app.js

@@ -11,6 +11,7 @@ const script_1 = require("./common/script");
 const slicer_1 = require("../benchmark/slicer");
 const all_variables_1 = require("../slicing/criterion/filters/all-variables");
 const path_1 = __importDefault(require("path"));
+const config_1 = require("../config");
 const options = (0, script_1.processCommandLineArgs)('benchmark-helper', [], {
     subtitle: 'Will slice for all possible variables, signal by exit code if slicing was successful, and can be run standalone',
     examples: [
@@ -23,6 +24,7 @@ if (options.verbose) {
 }
 const numberRegex = /^\d+$/;
 (0, assert_1.guard)(options.slice === 'all' || options.slice === 'no' || numberRegex.test(options.slice), 'slice must be either all, no, or a number');
+(0, assert_1.guard)(options['sampling-strategy'] === 'random' || options['sampling-strategy'] === 'equidistant', 'sample-strategy must be either random or equidistant');
 async function benchmark() {
     // we do not use the limit argument to be able to pick the limit randomly
     (0, assert_1.guard)(options.input !== undefined, 'No input file given');
@@ -31,18 +33,31 @@ async function benchmark() {
     // prefix for printing to console, includes file id and run number if present
     const prefix = `[${options.input}${options['file-id'] !== undefined ? ` (file ${options['file-id']}, run ${options['run-num']})` : ''}]`;
     console.log(`${prefix} Appending output to ${options.output}`);
-    fs_1.default.mkdirSync(path_1.default.parse(options.output).dir, { recursive: true });
+    const directory = path_1.default.parse(options.output).dir;
+    // ensure the directory exists if path contains one
+    if (directory !== '') {
+        fs_1.default.mkdirSync(directory, { recursive: true });
+    }
+    // Enable pointer analysis if requested, otherwise disable it
+    if (options['enable-pointer-tracking']) {
+        (0, config_1.amendConfig)({ solver: { ...(0, config_1.getConfig)().solver, pointerTracking: true, } });
+    }
+    else {
+        (0, config_1.amendConfig)({ solver: { ...(0, config_1.getConfig)().solver, pointerTracking: false, } });
+    }
     // ensure the file exists
     const fileStat = fs_1.default.statSync(options.input);
     (0, assert_1.guard)(fileStat.isFile(), `File ${options.input} does not exist or is no file`);
     const request = { request: 'file', content: options.input };
+    const maxSlices = options['max-slices'] ?? -1;
     const slicer = new slicer_1.BenchmarkSlicer(options.parser);
     try {
-        await slicer.init(request);
+        await slicer.init(request, undefined, options.threshold);
         // ${escape}1F${escape}1G${escape}2K for line reset
         if (options.slice === 'all') {
-            const count = await slicer.sliceForAll(all_variables_1.DefaultAllVariablesFilter, (i, total, arr) => console.log(`${prefix} Slicing ${i + 1}/${total} [${JSON.stringify(arr[i])}]`));
+            const count = await slicer.sliceForAll(all_variables_1.DefaultAllVariablesFilter, (i, total, arr) => console.log(`${prefix} Slicing ${i + 1}/${total} [${JSON.stringify(arr[i])}]`), { maxSliceCount: maxSlices });
             console.log(`${prefix} Completed Slicing`);
+            (0, assert_1.guard)(count >= 0, `Number of slices exceeded limit of ${maxSlices} with ${-count} slices, skipping in count`);
             (0, assert_1.guard)(count > 0, `No possible slices found for ${options.input}, skipping in count`);
         }
         else if (options.slice === 'no') {
@@ -51,8 +66,9 @@ async function benchmark() {
         else {
             const limit = parseInt(options.slice);
             console.log(`${prefix} Slicing up to ${limit} possible slices`);
-            const count = await slicer.sliceForAll(all_variables_1.DefaultAllVariablesFilter, (i, total, arr) => console.log(`${prefix} Slicing ${i + 1}/${total} [${JSON.stringify(arr[i])}]`), limit);
+            const count = await slicer.sliceForAll(all_variables_1.DefaultAllVariablesFilter, (i, total, arr) => console.log(`${prefix} Slicing ${i + 1}/${total} [${JSON.stringify(arr[i])}]`), { sampleCount: limit, maxSliceCount: maxSlices, sampleStrategy: options['sampling-strategy'] });
             console.log(`${prefix} Completed Slicing`);
+            (0, assert_1.guard)(count >= 0, `Number of slices exceeded limit of ${maxSlices} with ${-count} slices, skipping in count`);
             (0, assert_1.guard)(count > 0, `No possible slices found for ${options.input}, skipping in count`);
         }
         const { stats } = slicer.finish();

package/cli/common/options.js

@@ -16,11 +16,16 @@ exports.benchmarkOptions = [
     { name: 'help', alias: 'h', type: Boolean, description: 'Print this usage guide' },
     { name: 'limit', alias: 'l', type: Number, description: 'Limit the number of files to process (if given, this will choose these files randomly and add the chosen names to the output' },
     { name: 'runs', alias: 'r', type: Number, description: 'The amount of benchmark runs that should be done, out of which an average will be calculated' },
-    { name: 'input', alias: 'i', type: String, description: 'Pass a folder or file as src to read from', multiple: true, defaultOption: true, defaultValue: [], typeLabel: '{underline files/folders}' },
+    { name: 'input', alias: 'i', type: String, description: 'Pass a folder or file as src to read from. Alternatively, pass a single JSON file that contains a list of paths.', multiple: true, defaultOption: true, defaultValue: [], typeLabel: '{underline files/folders}' },
     { name: 'parallel', alias: 'p', type: String, description: 'Number of parallel executors (defaults to {italic max(cpu.count-1, 1)})', defaultValue: Math.max(os_1.default.cpus().length - 1, 1), typeLabel: '{underline number}' },
     { name: 'slice', alias: 's', type: String, description: 'Automatically slice for *all* variables (default) or *no* slicing and only parsing/dataflow construction. Numbers will indicate: sample X random slices from all.', defaultValue: 'all', typeLabel: '{underline all/no}' },
-    { name: 'output', alias: 'o', type: String, description: `Directory to write all the measurements to in a per-file-basis (defaults to {italic benchmark-${StartTimeString}})`, defaultValue: `benchmark-${StartTimeString}`, typeLabel: '{underline file}' },
-    { name: 'parser', type: String, description: 'The parser to use for the benchmark', defaultValue: 'r-shell', typeLabel: '{underline parser}' }
+    { name: 'output', alias: 'o', type: String, description: `Folder to write all the measurements to in a per-file-basis (defaults to {italic benchmark-${StartTimeString}})`, defaultValue: `benchmark-${StartTimeString}`, typeLabel: '{underline folder}' },
+    { name: 'parser', type: String, description: 'The parser to use for the benchmark', defaultValue: 'r-shell', typeLabel: '{underline parser}' },
+    { name: 'enable-pointer-tracking', type: Boolean, description: 'Run dataflow analysis with pointer tracking', defaultValue: false },
+    { name: 'max-file-slices', type: Number, description: 'If file has more than passed number of slices, the file is not processed', defaultValue: -1, typeLabel: '{underline number}' },
+    { name: 'threshold', alias: 't', type: Number, description: 'How many re-visits of the same node are ok?', defaultValue: undefined, typeLabel: '{underline number}' },
+    { name: 'per-file-time-limit', type: Number, description: 'Time limit in milliseconds to process single file (disabled by default)', defaultValue: undefined, typeLabel: '{underline number}' },
+    { name: 'sampling-strategy', type: String, description: 'Which strategy to use, when sampling is enabled', defaultValue: 'random', typeLabel: '{underline random/edquidistant}' },
 ];
 exports.benchmarkHelperOptions = [
     { name: 'verbose', alias: 'v', type: Boolean, description: 'Run with verbose logging [do not use for the real benchmark as this affects the time measurements, but only to find errors]' },
@@ -30,7 +35,11 @@ exports.benchmarkHelperOptions = [
     { name: 'run-num', alias: 'r', type: Number, description: 'The n-th time that the file with the given file-id is being benchmarked' },
     { name: 'slice', alias: 's', type: String, description: 'Automatically slice for *all* variables (default) or *no* slicing and only parsing/dataflow construction. Numbers will indicate: sample X random slices from all.', defaultValue: 'all', typeLabel: '{underline all/no}' },
     { name: 'output', alias: 'o', type: String, description: 'File to write the measurements to (appends a single line in JSON format)', typeLabel: '{underline file}' },
-    { name: 'parser', type: String, description: 'The parser to use for the benchmark', defaultValue: 'r-shell', typeLabel: '{underline parser}' }
+    { name: 'parser', type: String, description: 'The parser to use for the benchmark', defaultValue: 'r-shell', typeLabel: '{underline parser}' },
+    { name: 'enable-pointer-tracking', type: Boolean, description: 'Run dataflow analysis with pointer tracking', defaultValue: false },
+    { name: 'max-slices', type: Number, description: 'If file has more than passed number of slices, the file is not processed', defaultValue: -1, typeLabel: '{underline number}' },
+    { name: 'threshold', alias: 't', type: Number, description: 'How many re-visits of the same node are ok?', defaultValue: undefined, typeLabel: '{underline number}' },
+    { name: 'sampling-strategy', type: String, description: 'Which strategy to use, when sampling is enabled', defaultValue: 'random', typeLabel: '{underline random/edquidistant}' },
 ];
 exports.exportQuadsOptions = [
     { name: 'verbose', alias: 'v', type: Boolean, description: 'Run with verbose logging' },

package/cli/repl/commands/repl-commands.js

@@ -81,6 +81,8 @@ const _commands = {
     'normalize*': repl_normalize_1.normalizeStarCommand,
     'dataflow': repl_dataflow_1.dataflowCommand,
     'dataflow*': repl_dataflow_1.dataflowStarCommand,
+    'dataflowsimple': repl_dataflow_1.dataflowSimplifiedCommand,
+    'dataflowsimple*': repl_dataflow_1.dataflowSimpleStarCommand,
     'controlflow': repl_cfg_1.controlflowCommand,
     'controlflow*': repl_cfg_1.controlflowStarCommand,
     'lineage': repl_lineage_1.lineageCommand,

package/cli/repl/commands/repl-dataflow.d.ts

@@ -1,3 +1,5 @@
 import type { ReplCommand } from './repl-main';
 export declare const dataflowCommand: ReplCommand;
 export declare const dataflowStarCommand: ReplCommand;
+export declare const dataflowSimplifiedCommand: ReplCommand;
+export declare const dataflowSimpleStarCommand: ReplCommand;