[Query] Find Exceptions of Functions

README.md

@@ -41,7 +41,7 @@ It offers a wide variety of features, for example:
                ╰ Path `/root/x.txt` at 1.1-23
            ╰ Metadata: totalReads: 1, totalUnknown: 0, totalWritesBeforeAlways: 0, totalValid: 0, searchTimeMs: 0, processTimeMs: 0
        ╰ Seeded Randomness (seeded-randomness):
-           ╰ Metadata: consumerCalls: 0, callsWithFunctionProducers: 0, callsWithAssignmentProducers: 0, callsWithNonConstantProducers: 0, callsWithOtherBranchProducers: 0, searchTimeMs: 1, processTimeMs: 0
+           ╰ Metadata: consumerCalls: 0, callsWithFunctionProducers: 0, callsWithAssignmentProducers: 0, callsWithNonConstantProducers: 0, callsWithOtherBranchProducers: 0, searchTimeMs: 0, processTimeMs: 0
        ╰ Absolute Paths (absolute-file-paths):
            ╰ certain:
                ╰ Path `/root/x.txt` at 1.1-23
@@ -51,7 +51,7 @@ It offers a wide variety of features, for example:
        ╰ Naming Convention (naming-convention):
            ╰ Metadata: numMatches: 0, numBreak: 0, searchTimeMs: 0, processTimeMs: 0
        ╰ Network Functions (network-functions):
-           ╰ Metadata: totalCalls: 0, totalFunctionDefinitions: 0, searchTimeMs: 0, processTimeMs: 0
+           ╰ Metadata: totalCalls: 0, totalFunctionDefinitions: 0, searchTimeMs: 0, processTimeMs: 1
        ╰ Dataframe Access Validation (dataframe-access-validation):
            ╰ Metadata: numOperations: 0, numAccesses: 0, totalAccessed: 0, searchTimeMs: 0, processTimeMs: 0
        ╰ Dead Code (dead-code):
@@ -80,9 +80,9 @@ It offers a wide variety of features, for example:
 
     _Results (prettified and summarized):_
 
-    Query: **linter** (3 ms)\
+    Query: **linter** (4 ms)\
        ╰ **Deprecated Functions** (deprecated-functions):\
-    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;╰ _Metadata_: <code>totalCalls: 0, totalFunctionDefinitions: 0, searchTimeMs: 0, processTimeMs: 0</code>\
+    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;╰ _Metadata_: <code>totalCalls: 0, totalFunctionDefinitions: 0, searchTimeMs: 2, processTimeMs: 0</code>\
     &nbsp;&nbsp;&nbsp;╰ **File Path Validity** (file-path-validity):\
     &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;╰ certain:\
     &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;╰ Path `/root/x.txt` at 1.1-23\
@@ -92,24 +92,24 @@ It offers a wide variety of features, for example:
     &nbsp;&nbsp;&nbsp;╰ **Absolute Paths** (absolute-file-paths):\
     &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;╰ certain:\
     &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;╰ Path `/root/x.txt` at 1.1-23\
-    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;╰ _Metadata_: <code>totalConsidered: 1, totalUnknown: 0, searchTimeMs: 1, processTimeMs: 0</code>\
+    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;╰ _Metadata_: <code>totalConsidered: 1, totalUnknown: 0, searchTimeMs: 0, processTimeMs: 0</code>\
     &nbsp;&nbsp;&nbsp;╰ **Unused Definitions** (unused-definitions):\
     &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;╰ _Metadata_: <code>totalConsidered: 0, searchTimeMs: 0, processTimeMs: 0</code>\
     &nbsp;&nbsp;&nbsp;╰ **Naming Convention** (naming-convention):\
     &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;╰ _Metadata_: <code>numMatches: 0, numBreak: 0, searchTimeMs: 0, processTimeMs: 0</code>\
     &nbsp;&nbsp;&nbsp;╰ **Network Functions** (network-functions):\
     &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;╰ _Metadata_: <code>totalCalls: 0, totalFunctionDefinitions: 0, searchTimeMs: 0, processTimeMs: 0</code>\
     &nbsp;&nbsp;&nbsp;╰ **Dataframe Access Validation** (dataframe-access-validation):\
-    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;╰ _Metadata_: <code>numOperations: 0, numAccesses: 0, totalAccessed: 0, searchTimeMs: 0, processTimeMs: 0</code>\
+    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;╰ _Metadata_: <code>numOperations: 0, numAccesses: 0, totalAccessed: 0, searchTimeMs: 0, processTimeMs: 1</code>\
     &nbsp;&nbsp;&nbsp;╰ **Dead Code** (dead-code):\
     &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;╰ _Metadata_: <code>consideredNodes: 5, searchTimeMs: 0, processTimeMs: 0</code>\
     &nbsp;&nbsp;&nbsp;╰ **Useless Loops** (useless-loop):\
     &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;╰ _Metadata_: <code>numOfUselessLoops: 0, searchTimeMs: 0, processTimeMs: 0</code>\
-    _All queries together required ≈3 ms (1ms accuracy, total 3 ms)_
+    _All queries together required ≈4 ms (1ms accuracy, total 4 ms)_
 
     <details> <summary style="color:gray">Show Detailed Results as Json</summary>
 
-    The analysis required _3.3 ms_ (including parsing and normalization and the query) within the generation environment.
+    The analysis required _4.3 ms_ (including parsing and normalization and the query) within the generation environment.
 
     In general, the JSON contains the Ids of the nodes in question as they are present in the normalized AST or the dataflow graph of flowR.
     Please consult the [Interface](https://github.com/flowr-analysis/flowr/wiki/Interface) wiki page for more information on how to get those.
@@ -126,7 +126,7 @@ It offers a wide variety of features, for example:
             ".meta": {
               "totalCalls": 0,
               "totalFunctionDefinitions": 0,
-              "searchTimeMs": 0,
+              "searchTimeMs": 2,
               "processTimeMs": 0
             }
           },
@@ -180,7 +180,7 @@ It offers a wide variety of features, for example:
             ".meta": {
               "totalConsidered": 1,
               "totalUnknown": 0,
-              "searchTimeMs": 1,
+              "searchTimeMs": 0,
               "processTimeMs": 0
             }
           },
@@ -217,7 +217,7 @@ It offers a wide variety of features, for example:
               "numAccesses": 0,
               "totalAccessed": 0,
               "searchTimeMs": 0,
-              "processTimeMs": 0
+              "processTimeMs": 1
             }
           },
           "dead-code": {
@@ -238,11 +238,11 @@ It offers a wide variety of features, for example:
           }
         },
         ".meta": {
-          "timing": 3
+          "timing": 4
         }
       },
       ".meta": {
-        "timing": 3
+        "timing": 4
       }
     }
     ```
@@ -321,7 +321,7 @@ It offers a wide variety of features, for example:
     N <- 10
     for(i in 1:(N-1)) sum <- sum + i + w
     sum
-    All queries together required ≈3 ms (1ms accuracy, total 3 ms)
+    All queries together required ≈4 ms (1ms accuracy, total 4 ms)
     ```
 
 
@@ -401,6 +401,7 @@ It offers a wide variety of features, for example:
 
     ```text
     https://mermaid.live/view#base64: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
+    Copied mermaid url to clipboard (dataflow: 0ms).
     ```
 
 
@@ -696,7 +697,7 @@ It offers a wide variety of features, for example:
     ```
 
 
-    (The analysis required _2.3 ms_ (including parse and normalize, using the [tree-sitter](https://github.com/flowr-analysis/flowr/wiki/Engines) engine) within the generation environment.)
+    (The analysis required _1.4 ms_ (including parse and normalize, using the [tree-sitter](https://github.com/flowr-analysis/flowr/wiki/Engines) engine) within the generation environment.)
 
 
 

src/dataflow/fn/exceptions-of-function.ts

@@ -0,0 +1,58 @@
+import type { NodeId } from '../../r-bridge/lang-4.x/ast/model/processing/node-id';
+import type { CallGraph } from '../graph/call-graph';
+import { VertexType } from '../graph/vertex';
+import type { ControlDependency } from '../info';
+import { ExitPointType } from '../info';
+import type { BuiltInMappingName } from '../environments/built-in';
+import { isBuiltIn } from '../environments/built-in';
+
+const CatchHandlers: ReadonlySet<string> = new Set<BuiltInMappingName>(['builtin:try']);
+export interface ExceptionPoint {
+	id:   NodeId;
+	cds?: readonly ControlDependency[] | undefined;
+}
+/**
+ * Collect exception sources of a function in the call graph.
+ * This returns the `NodeId`s of functions that may throw exceptions when called by the given function.
+ * Please be aware, that these are restricted to functions known by flowR.
+ */
+export function calculateExceptionsOfFunction(id: NodeId, graph: CallGraph): ExceptionPoint[] {
+	const collectedExceptions:  ExceptionPoint[] = [];
+	const visited = new Set<NodeId>();
+	const toVisit: NodeId[] = [id];
+
+	while(toVisit.length > 0) {
+		const currentId = toVisit.pop() as NodeId;
+		if(visited.has(currentId)) {
+			continue;
+		}
+		visited.add(currentId);
+
+		const vtx = graph.getVertex(currentId);
+		if(!vtx) {
+			continue;
+		}
+
+		if(isBuiltIn(currentId)) {
+			continue;
+		}
+
+		if(vtx.tag === VertexType.FunctionDefinition) {
+			for(const e of vtx.exitPoints.filter(e => e.type === ExitPointType.Error)) {
+				if(!collectedExceptions.find(x => x.id === e.nodeId)) {
+					collectedExceptions.push({ id: e.nodeId, cds: e.controlDependencies });
+				}
+			}
+		} else if(vtx.tag === VertexType.FunctionCall && vtx.origin !== 'unnamed' && vtx.origin.some(c => CatchHandlers.has(c))) {
+			// skip the try-catch handlers as they catch all exceptions within their body
+			continue;
+		}
+
+		const outEdges = graph.outgoingEdges(currentId) ?? [];
+		for(const [out] of outEdges) {
+			toVisit.push(out);
+		}
+	}
+
+	return Array.from(collectedExceptions);
+}

src/dataflow/internal/process/functions/call/built-in/built-in-try-catch.ts

@@ -1,6 +1,6 @@
 import type { DataflowProcessorInformation } from '../../../../../processor';
-import type { ExitPoint, DataflowInformation } from '../../../../../info';
-import { ExitPointType } from '../../../../../info';
+import type { ControlDependency, DataflowInformation, ExitPoint } from '../../../../../info';
+import { ExitPointType, happensInEveryBranch } from '../../../../../info';
 import { processKnownFunctionCall } from '../known-call-handling';
 import type { ParentInformation } from '../../../../../../r-bridge/lang-4.x/ast/model/processing/decorate';
 import {
@@ -11,10 +11,13 @@ import type { RSymbol } from '../../../../../../r-bridge/lang-4.x/ast/model/node
 import type { NodeId } from '../../../../../../r-bridge/lang-4.x/ast/model/processing/node-id';
 import { dataflowLogger } from '../../../../../logger';
 import { pMatch } from '../../../../linker';
+import type { DataflowGraphVertexInfo } from '../../../../../graph/vertex';
 import { VertexType } from '../../../../../graph/vertex';
 import { tryUnpackNoNameArg } from '../argument/unpack-argument';
 import type { DataflowGraph } from '../../../../../graph/graph';
 import { RType } from '../../../../../../r-bridge/lang-4.x/ast/model/type';
+import { isUndefined } from '../../../../../../util/assert';
+import { EdgeType } from '../../../../../graph/edge';
 
 
 function getArgsOfName(argMaps: Map<NodeId, string>, name: string): Set<NodeId> {
@@ -61,43 +64,86 @@ export function processTryCatch<OtherInfo>(
 	const errorArg = getArgsOfName(argMaps, 'error');
 	const finallyArg = getArgsOfName(argMaps, 'finally');
 	// only take those exit points from the block
+	// check whether blockArg has *always* happening exceptions, if so we do not constrain the error handler
+	const blockErrorExitPoints: (ControlDependency | undefined)[] = [];
+	const errorExitPoints: ExitPoint[] = [];
 	(info.exitPoints as ExitPoint[]) = res.processedArguments.flatMap(arg => {
 		if(!arg) {
 			return [];
 		}
 		// this calls error and finally args
-		if(finallyArg.has(arg.entryPoint) || errorArg.has(arg.entryPoint)) {
-			return handleFdefAsCalled(arg.entryPoint, info.graph, arg.exitPoints);
+		if(finallyArg.has(arg.entryPoint)) {
+			return handleFdefAsCalled(arg.entryPoint, info.graph, arg.exitPoints, undefined);
+		} else if(errorArg.has(arg.entryPoint)) {
+			errorExitPoints.push(...getExitPoints(info.graph.getVertex(arg.entryPoint), info.graph) ?? []);
 		}
 		if(!blockArg.has(arg.entryPoint)) {
 			// not killing other args
 			return arg.exitPoints;
 		}
+		blockErrorExitPoints.push(...arg.exitPoints.filter(ep => ep.type === ExitPointType.Error).flatMap(a => a.controlDependencies));
 		return arg.exitPoints.filter(ep => ep.type !== ExitPointType.Error);
 	});
-
-
+	if(errorExitPoints.length > 0) {
+		if(happensInEveryBranch(blockErrorExitPoints.some(isUndefined) ? undefined : blockErrorExitPoints as ControlDependency[])) {
+			(info.exitPoints as ExitPoint[]).push(...errorExitPoints);
+		} else {
+			(info.exitPoints as ExitPoint[]).push(...constrainExitPoints(errorExitPoints, blockArg));
+		}
+	}
+	for(const e of errorArg) {
+		info.graph.addEdge(rootId, e, EdgeType.Calls);
+	}
+	for(const f of finallyArg) {
+		info.graph.addEdge(rootId, f, EdgeType.Calls);
+	}
+	for(const e of info.exitPoints) {
+		if(e.type !== ExitPointType.Error) {
+			info.graph.addEdge(rootId, e.nodeId, EdgeType.Returns);
+		}
+	}
 	return info;
 }
 
-function handleFdefAsCalled(nodeId: NodeId, graph: DataflowGraph, def: readonly ExitPoint[]): readonly ExitPoint[] {
-	const v = graph.getVertex(nodeId);
-	if(!v) {
-		return def;
+function getExitPoints(vertex: DataflowGraphVertexInfo | undefined, graph: DataflowGraph): readonly ExitPoint[] | undefined {
+	if(!vertex) {
+		return undefined;
 	}
-	if(v.tag === VertexType.FunctionDefinition) {
-		return v.exitPoints;
+	if(vertex.tag === VertexType.FunctionDefinition) {
+		return vertex.exitPoints;
 	}
 	// we assumed named argument
-	const n = graph.idMap?.get(nodeId);
+	const n = graph.idMap?.get(vertex.id);
 	if(!n) {
-		return def;
+		return undefined;
 	}
 	if(n.type === RType.Argument && n.value?.type === RType.FunctionDefinition) {
 		const fdefV = graph.getVertex(n.value.info.id);
 		if(fdefV?.tag === VertexType.FunctionDefinition) {
 			return fdefV.exitPoints;
 		}
 	}
-	return def;
+	return undefined;
+}
+
+function handleFdefAsCalled(nodeId: NodeId, graph: DataflowGraph, def: readonly ExitPoint[], constrain: Set<NodeId> | undefined): readonly ExitPoint[] {
+	const v = graph.getVertex(nodeId);
+	const e = getExitPoints(v, graph);
+	return e ? constrainExitPoints(e, constrain) : def;
 }
+
+function constrainExitPoints(exitPoints: readonly ExitPoint[], constrain: Set<NodeId> | undefined): readonly ExitPoint[] {
+	if(!constrain || constrain.size === 0) {
+		return exitPoints;
+	}
+	// append constrains with true
+	const cds = Array.from(constrain, id => ({ id, when: true }));
+	return exitPoints.map(e => {
+		if(e.controlDependencies) {
+			e.controlDependencies.push(...cds);
+			return e;
+		} else {
+			return { ...e, controlDependencies: cds };
+		}
+	});
+}

src/documentation/doc-util/doc-query.ts

@@ -96,6 +96,7 @@ export interface QueryDocumentation {
 	readonly type:             'virtual' | 'active';
 	readonly shortDescription: string;
 	readonly functionName:     string;
+	/** Path to the file implementing the query function, the wiki generation will fail if this isnt found */
 	readonly functionFile:     string;
 	readonly buildExplanation: (shell: RShell, ctx: GeneralDocContext) => Promise<string>;
 }

src/documentation/wiki-query.ts

@@ -48,6 +48,7 @@ import { executeFileQuery } from '../queries/catalog/files-query/files-query-exe
 import { executeCallGraphQuery } from '../queries/catalog/call-graph-query/call-graph-query-executor';
 import { executeRecursionQuery } from '../queries/catalog/inspect-recursion-query/inspect-recursion-query-executor';
 import { executeDoesCallQuery } from '../queries/catalog/does-call-query/does-call-query-executor';
+import { executeExceptionQuery } from '../queries/catalog/inspect-exceptions-query/inspect-exception-query-executor';
 
 
 registerQueryDocumentation('call-context', {
@@ -373,6 +374,39 @@ ${
 	}
 });
 
+registerQueryDocumentation('inspect-exception', {
+	name:             'Inspect Exceptions of Functions Query',
+	type:             'active',
+	shortDescription: 'Determine whether functions throw exceptions (known to flowR)',
+	functionName:     executeExceptionQuery.name,
+	functionFile:     '../queries/catalog/inspect-exceptions-query/inspect-exception-query-executor.ts',
+	buildExplanation: async(shell: RShell) => {
+		const exampleCode = `mayFail <- function(x) {
+  if(x < 0) stop("Negative value!")
+  else sqrt(x)
+}
+safeFail <- function(x) {
+  tryCatch(
+    mayFail(x),
+    error = function(e) { NA }
+  )
+}`;
+		return `
+With this query you can identify which functions in the code throw exceptions (known to flowR).
+
+Using the following example code:
+${codeBlock('r', exampleCode)}
+the following query returns the information for all identified function definitions whether they throw exceptions:
+${
+	await showQuery(shell, exampleCode, [{
+		type: 'inspect-exception',
+	}], { showCode: true, collapseQuery: true })
+}
+		`;
+	}
+});
+
+
 registerQueryDocumentation('origin', {
 	name:             'Origin Query',
 	type:             'active',