Perf improvements #1096

Frotty · 2025-09-29T07:53:27Z

Main improvements:

Use uncompressed mpq storage because compression is a bottleneck for big maps with tons of imports
- Using uncompressed for runmap only, compressing on buildmap
Better/Safer caching for compiletime results
With -noExtractMapScript option the war3map.j no longer gets forcefully reparsed on every runmap command (it's already part of the model)
Use iteration instead of recursion in places where StackOverflows happened
Reduce a lot of string allocation garbage, which reduces GC churn
Use SCC for worklist algorithm
Remove TestNG suite usage, just run all tests. Fixed tests that didn't run. Added new tests
Added RealRealMixed rewrites and tests
Java 25 now required

Copilot

Pull Request Overview

This PR introduces significant performance optimizations to the Wurst compiler, focusing on reducing compilation time through various algorithmic improvements and memory optimizations. The main goals are to improve scalability for large maps with many imports while maintaining correctness.

Replaces recursive algorithms with iterative implementations to prevent stack overflows
Introduces a two-phase validation approach to optimize control flow analysis
Implements comprehensive caching and memoization strategies
Switches to high-performance data structures (fastutil library)

Reviewed Changes

Copilot reviewed 28 out of 28 changed files in this pull request and generated 5 comments.

Show a summary per file

File	Description
SccForwardExecution.java	New SCC-based dataflow analysis implementation for better performance on large functions
ForwardMethod.java	Updates execution to use new SCC-based approach
DataflowAnomalyAnalysis.java	Converts recursive local variable collection to iterative
WurstValidator.java	Implements two-phase validation with light/heavy phases and adds subtype caching
WurstTypeClass.java	Optimizes interface collection using manual loops instead of streams
ImTranslator.java	Switches to fastutil data structures and optimizes call relation calculation
Various other files	Performance improvements including fastutil adoption, iterative algorithms, and memory optimizations

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

de.peeeq.wurstscript/src/main/java/de/peeeq/wurstscript/intermediatelang/ILconstArray.java

de.peeeq.wurstscript/src/main/java/de/peeeq/wurstscript/validation/WurstValidator.java

de.peeeq.wurstscript/src/main/java/de/peeeq/wurstscript/jurst/ExtendedJurstLexer.java

de.peeeq.wurstscript/src/main/java/de/peeeq/wurstscript/jass/ExtendedJassLexer.java

de.peeeq.wurstscript/src/main/java/de/peeeq/wurstscript/intermediatelang/interpreter/State.java

coveralls · 2025-09-29T09:10:41Z

coverage: 64.646% (+2.4%) from 62.222%
when pulling fb48177 on perf-improvements
into e96e226 on master.

Copilot

Pull Request Overview

Copilot reviewed 46 out of 46 changed files in this pull request and generated 5 comments.

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

de.peeeq.wurstscript/src/test/java/tests/wurstscript/tests/BugTests.java

de.peeeq.wurstscript/src/main/java/de/peeeq/wurstscript/validation/WurstValidator.java

Copilot · 2025-09-29T09:16:57Z

...cript/src/main/java/de/peeeq/wurstscript/validation/controlflow/DataflowAnomalyAnalysis.java

+            // visit children left→right: push in reverse
+            for (int i = e.size() - 1; i >= 0; i--) {
+                Element c = e.get(i);
+                if (!(c instanceof ExprClosure) && !(c instanceof ExprStatementsBlock)) {
+                    stack.push(c);
+                }
            }


[nitpick] The iteration is in reverse order (i--) which suggests intentional left-to-right processing, but this could be clarified with a comment explaining why the reverse iteration is necessary for correct traversal order.

Copilot · 2025-09-29T09:16:57Z

...q.wurstscript/src/main/java/de/peeeq/wurstscript/translation/imtranslation/ImTranslator.java

-        while (!functionStack.isEmpty()) {
-            ImFunction f = functionStack.pop();
+        while (!work.isEmpty()) {
+            final ImFunction f = work.removeLast(); // LIFO (DFS); change to removeFirst() for BFS


[nitpick] The comment suggests that changing to removeFirst() would make this BFS, but the data structure choice and traversal algorithm are more complex than just the removal method. Consider clarifying that DFS vs BFS also depends on how children are added to the queue.

Suggested change

final ImFunction f = work.removeLast(); // LIFO (DFS); change to removeFirst() for BFS

final ImFunction f = work.removeLast(); // LIFO (DFS); for BFS, use removeFirst() *and* ensure children are added to the end

...rstscript/src/main/java/de/peeeq/wurstscript/intermediatelang/interpreter/ILInterpreter.java

…tion/WurstValidator.java Co-authored-by: Copilot <[email protected]>

Also "an other" typo fix and use Arrays.hashCode

…rstScript into perf-improvements

Copilot

Pull Request Overview

Copilot reviewed 54 out of 70 changed files in this pull request and generated 3 comments.

Comments suppressed due to low confidence (1)

de.peeeq.wurstscript/src/test/java/tests/wurstscript/tests/OptimizerTests.java:1

The condition v <= 0 should be v < 0 for unary minus operation. Currently, -0 is being converted to 0, but mathematically -0 should remain 0, and the condition should only handle negative values.

package tests.wurstscript.tests;

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

Copilot · 2025-09-29T12:08:55Z

de.peeeq.wurstscript/src/test/java/tests/wurstscript/tests/OptimizerTests.java



+    }
+


[nitpick] Missing newline before the new test method. Add a blank line before @Test to maintain consistent formatting with other test methods in the file.

Suggested change

...urstscript/src/main/java/de/peeeq/wurstscript/intermediatelang/optimizer/SimpleRewrites.java

...eeeq.wurstscript/src/main/java/de/peeeq/wurstscript/translation/imtranslation/ImPrinter.java

…ation/imtranslation/ImPrinter.java Co-authored-by: Copilot <[email protected]>

…rstScript into perf-improvements

Copilot

Pull Request Overview

Copilot reviewed 54 out of 70 changed files in this pull request and generated 5 comments.

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

Copilot · 2025-09-29T12:22:00Z

de.peeeq.wurstscript/src/test/java/tests/wurstscript/tests/OptimizerTests.java

+            "endpackage");
        String compiledAndOptimized = Files.toString(new File("test-output/OptimizerTests_test_unused_func_remover_opt.j"), Charsets.UTF_8);
-        assertFalse("I2S should be removed", compiledAndOptimized.contains("I2S"));
+        assertFalse(compiledAndOptimized.contains("I2S"), "I2S should be removed");


[nitpick] The updated TestNG assertion syntax reverses the parameter order from the old AssertJUnit style. Consider updating all assertions consistently throughout the codebase to use the new syntax pattern assertX(actual, expected, message) for better consistency.

Copilot · 2025-09-29T12:22:01Z

...rstscript/src/main/java/de/peeeq/wurstscript/validation/controlflow/SccForwardExecution.java

+        Queue<WStatement> worklist = new ArrayDeque<>(scc);
+
+        int iterations = 0;
+        int maxIterations = scc.size() * scc.size() + 100; // Heuristic limit to prevent infinite loops


[nitpick] The magic number 100 in the iteration limit calculation should be extracted as a named constant to improve code readability and maintainability. Consider defining private static final int ITERATION_BUFFER = 100;

Copilot · 2025-09-29T12:22:01Z

de.peeeq.wurstscript/src/main/java/de/peeeq/wurstscript/validation/WurstValidator.java

+    // crude cap to avoid unbounded growth; tune as needed
+    private static final int SUBTYPE_MEMO_CAP = 16_384;


[nitpick] The cache size limit of 16,384 should be made configurable or at least documented with reasoning for this specific value. Consider adding a comment explaining the memory/performance trade-off that led to this choice.

Suggested change

// crude cap to avoid unbounded growth; tune as needed

private static final int SUBTYPE_MEMO_CAP = 16_384;

/**

* Maximum number of cached subtype checks before eviction.

* Default is 16,384, chosen to balance memory usage and performance for typical WurstScript projects.

* If the cache grows beyond this, it is fully cleared (simple eviction policy).

* You can override the default by setting the system property "wurst.subtypeMemoCap".

*/

private static final int SUBTYPE_MEMO_CAP = Integer.getInteger("wurst.subtypeMemoCap", 16_384);

Copilot · 2025-09-29T12:22:01Z

...q.wurstscript/src/main/java/de/peeeq/wurstscript/translation/imtranslation/ImTranslator.java

        return callRelations;
    }

+    public ImFunction getMainFunc() { return mainFunc; }


[nitpick] These simplified getter methods should include null-safety documentation or add null annotations to clarify the contract, since the original methods were more explicit about potential null returns.

Suggested change

public ImFunction getMainFunc() { return mainFunc; }

public ImFunction getMainFunc() { return mainFunc; }

@Nullable

Copilot · 2025-09-29T12:22:02Z

...eeeq.wurstscript/src/main/java/de/peeeq/wurstscript/translation/imtranslation/ImPrinter.java

+        // take only the last 3 digits
+        int v = h % 1000;


[nitpick] The hash collision potential increases significantly when reducing to only 3 digits (1000 possible values). Consider using a larger range or document the trade-off between readability and collision probability.

Suggested change

// take only the last 3 digits

int v = h % 1000;

// take only the last 5 digits

int v = h % 100000;

Copilot

Pull Request Overview

Copilot reviewed 86 out of 102 changed files in this pull request and generated 3 comments.

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

...rc/main/java/de/peeeq/wurstscript/intermediatelang/optimizer/ConstantAndCopyPropagation.java

de.peeeq.wurstscript/src/main/java/de/peeeq/wurstio/languageserver/requests/MapRequest.java

de.peeeq.wurstscript/src/main/java/de/peeeq/wurstscript/WLogger.java

Copilot

Pull Request Overview

Copilot reviewed 85 out of 101 changed files in this pull request and generated 4 comments.

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

...eeeq.wurstscript/src/main/java/de/peeeq/wurstscript/translation/imtranslation/ImPrinter.java

Copilot · 2025-10-01T10:45:52Z

...eeq.wurstscript/src/main/java/de/peeeq/wurstscript/attributes/prettyPrint/PrettyPrinter.java

        if (e.getAnnotationMessage() != null && e.getAnnotationMessage().length() >= 1) {
            sb.append("(");
-            sb.append(e.getAnnotationMessage());
+            sb.append(Utils.escapeString(e.getAnnotationMessage()));


[nitpick] Using Utils.escapeString() on annotation messages could introduce security vulnerabilities if the escaping is not comprehensive. Ensure that all potentially dangerous characters are properly escaped to prevent injection attacks.

Copilot · 2025-10-01T10:45:52Z

...q.wurstscript/src/main/java/de/peeeq/wurstscript/intermediatelang/optimizer/LocalMerger.java

+        @SuppressWarnings("unchecked") final ObjectOpenHashSet<ImVar>[] in  = new ObjectOpenHashSet[N];
+        @SuppressWarnings("unchecked") final ObjectOpenHashSet<ImVar>[] out = new ObjectOpenHashSet[N];


Using generic arrays with @SuppressWarnings is an anti-pattern. Consider using List<ObjectOpenHashSet<ImVar>> instead of arrays to avoid unchecked warnings and improve type safety.

de.peeeq.wurstscript/src/main/java/de/peeeq/wurstscript/validation/WurstValidator.java

…ation/imtranslation/ImPrinter.java Co-authored-by: Copilot <[email protected]>

Copilot

Pull Request Overview

Copilot reviewed 134 out of 154 changed files in this pull request and generated 5 comments.

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

de.peeeq.wurstscript/src/test/java/tests/wurstscript/tests/OptimizerTests.java

...script/src/main/java/de/peeeq/wurstscript/translation/imtranslation/StackTraceInjector2.java

...stscript/src/main/java/de/peeeq/wurstscript/intermediatelang/optimizer/FunctionSplitter.java

...eeeq.wurstscript/src/main/java/de/peeeq/wurstscript/translation/imoptimizer/ImOptimizer.java

…ediatelang/optimizer/FunctionSplitter.java Co-authored-by: Copilot <[email protected]>

…ation/imtranslation/StackTraceInjector2.java Co-authored-by: Copilot <[email protected]>

Copilot

Pull Request Overview

Copilot reviewed 134 out of 154 changed files in this pull request and generated 6 comments.

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

Copilot · 2025-10-02T13:43:14Z

de.peeeq.wurstscript/src/test/java/tests/wurstscript/tests/OptimizerTests.java

+            "		println(I2S(blablub))",
+            "endpackage");
        String output = Files.toString(new File("./test-output/OptimizerTests_test_tempVarRemover2_inlopt.j"), Charsets.UTF_8);
+        // Better not inline GetRandomInt call - it might have side effects!


[nitpick] The comment has a grammatical issue. It should be 'Better not to inline GetRandomInt call' or 'It's better not to inline GetRandomInt call'.

Suggested change

// Better not inline GetRandomInt call - it might have side effects!

// It's better not to inline GetRandomInt call - it might have side effects!

Copilot · 2025-10-02T13:43:15Z

...rstscript/src/main/java/de/peeeq/wurstscript/validation/controlflow/SccForwardExecution.java

+        while (!worklist.isEmpty()) {
+            if (iterations++ > maxIterations) {
+                // This should ideally not happen in a correct CFG with a monotonic transfer function
+                throw new RuntimeException("Dataflow analysis did not converge. Possible infinite loop in component.");


The error message could be more helpful by including details about which component failed to converge and the current iteration count.

Suggested change

throw new RuntimeException("Dataflow analysis did not converge. Possible infinite loop in component.");

throw new RuntimeException(

"Dataflow analysis did not converge after " + iterations + " iterations (max allowed: " + maxIterations + "). " +

"Possible infinite loop in component: " + scc

);

Copilot · 2025-10-02T13:43:15Z

...cript/src/main/java/de/peeeq/wurstscript/validation/controlflow/DataflowAnomalyAnalysis.java

-            r.add((LocalVarDef) e);
-        }
+    private void collectLocalVars(Set<LocalVarDef> out, Element root) {
+        ArrayDeque<Element> stack = new ArrayDeque<>();


[nitpick] The iterative implementation is good for avoiding stack overflow, but consider using a more specific initial capacity for the ArrayDeque based on expected tree depth to reduce allocations.

Suggested change

ArrayDeque<Element> stack = new ArrayDeque<>();

ArrayDeque<Element> stack = new ArrayDeque<>(32);

Copilot · 2025-10-02T13:43:15Z

de.peeeq.wurstscript/src/main/java/de/peeeq/wurstscript/validation/WurstValidator.java

+    // crude cap to avoid unbounded growth; tune as needed
+
+    private static boolean isSubtypeCached(WurstType actual, WurstType expected, Annotation site) {
+        if (actual == expected) return true;


[nitpick] The reference equality check actual == expected before equalsType is a good optimization, but ensure this doesn't bypass necessary type checking logic that might be required even for the same reference.

Suggested change

if (actual == expected) return true;

Copilot · 2025-10-02T13:43:16Z

...rc/main/java/de/peeeq/wurstscript/intermediatelang/optimizer/ConstantAndCopyPropagation.java

+                            // If no folding happened, try regular value propagation
+                            if (newValue == null) {


[nitpick] Consider extracting the constant folding logic into a separate method to improve readability of the transfer function.

Copilot · 2025-10-02T13:43:16Z

...q.wurstscript/src/main/java/de/peeeq/wurstscript/intermediatelang/optimizer/LocalMerger.java

+                return t.getPredecessors();
+            }
+        };
+        // Use the path-based strong component algorithm [1] on the reversed CFG.


The reference '[1]' is incomplete. Consider providing the full citation for the path-based strong component algorithm being referenced.

Suggested change

// Use the path-based strong component algorithm [1] on the reversed CFG.

// Use the path-based strong component algorithm (see: R. Tarjan, "Depth-First Search and Linear Graph Algorithms", SIAM J. Comput., 1972) on the reversed CFG.

Copilot

Pull Request Overview

Copilot reviewed 134 out of 154 changed files in this pull request and generated 5 comments.

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

Copilot · 2025-10-02T23:32:48Z

de.peeeq.wurstscript/src/test/java/tests/wurstscript/tests/OptimizerTests.java

+            "	function GetRandomIntt(int a, int b) returns int",
+            "     return a + b",
+            "	init",
+            "		let blablub = GetRandomIntt(0,100)",


Function name has extra 't' - should be 'GetRandomInt' not 'GetRandomIntt'.

Suggested change

" function GetRandomIntt(int a, int b) returns int",

" return a + b",

" init",

" let blablub = GetRandomIntt(0,100)",

" function GetRandomInt(int a, int b) returns int",

" return a + b",

" init",

" let blablub = GetRandomInt(0,100)",

Copilot · 2025-10-02T23:32:49Z

...q.wurstscript/src/main/java/de/peeeq/wurstscript/intermediatelang/optimizer/LocalMerger.java

 package de.peeeq.wurstscript.intermediatelang.optimizer;

-import de.peeeq.datastructures.Worklist;
+import de.peeeq.datastructures.GraphInterpreter;


Excellent use of SCC-based analysis for liveness calculation. This provides significant performance improvements over naive worklist algorithms by containing iterative analysis within strongly connected components.

Suggested change

import de.peeeq.datastructures.GraphInterpreter;

Copilot · 2025-10-02T23:32:49Z

de.peeeq.wurstscript/src/main/java/de/peeeq/wurstscript/validation/WurstValidator.java

+    private enum Phase { LIGHT, HEAVY }
+    private Phase phase = Phase.LIGHT;


Two-phase validation approach is a smart optimization - separating lightweight checks from expensive control-flow and dataflow analysis reduces overall compilation time.

Copilot · 2025-10-02T23:32:49Z

...urstscript/src/main/java/de/peeeq/wurstscript/intermediatelang/optimizer/SimpleRewrites.java

-                    ImIntVal newVal = JassIm.ImIntVal(inverseVal);
-                    opc.replaceBy(newVal);
+                int v = ((ImIntVal) expr).getValI();
+                if (v != Integer.MIN_VALUE && v <= 0) {


The condition v <= 0 for unary minus folding is incorrect. Unary minus should fold for any representable value except Integer.MIN_VALUE. The condition should be v != Integer.MIN_VALUE.

Suggested change

if (v != Integer.MIN_VALUE && v <= 0) {

if (v != Integer.MIN_VALUE) {

Copilot · 2025-10-02T23:32:50Z

...rstscript/src/main/java/de/peeeq/wurstscript/intermediatelang/interpreter/ILInterpreter.java

-    private static boolean cache = false;
    private final ProgramState globalState;
    private final TimerMockHandler timerMockHandler = new TimerMockHandler();



Removed cache parameter from constructor - this is a breaking change that could affect external code using this API.

Suggested change

/**

* @deprecated The cache parameter is no longer used. Use {@link #ILInterpreter(ImProg, WurstGui, Optional, ProgramState)} instead.

*/

@Deprecated

public ILInterpreter(ImProg prog, WurstGui gui, Optional<File> mapFile, Object cache, ProgramState globalState) {

this(prog, gui, mapFile, globalState);

}

Frotty added 2 commits September 24, 2025 19:36

decent improvements

cdced5d

some more good ones

1881f29

Frotty requested a review from Copilot September 29, 2025 07:53

Copilot AI reviewed Sep 29, 2025

View reviewed changes

Frotty added 2 commits September 29, 2025 11:00

restore determinism, fix tests

84fb523

Optimize imports

bed57ac

Frotty requested a review from Copilot September 29, 2025 09:14

Copilot AI reviewed Sep 29, 2025

View reviewed changes

Frotty and others added 4 commits September 29, 2025 11:17

Update de.peeeq.wurstscript/src/main/java/de/peeeq/wurstscript/valida…

343c032

…tion/WurstValidator.java Co-authored-by: Copilot <[email protected]>

Improve SCC processing, run all tests and fix them

a513c86

Also "an other" typo fix and use Arrays.hashCode

Merge branch 'perf-improvements' of https://github.com/wurstscript/Wu…

b96882c

…rstScript into perf-improvements

add RealRealMixed rewrites and tests

94c4f32

Frotty requested a review from Copilot September 29, 2025 12:07

Copilot AI reviewed Sep 29, 2025

View reviewed changes

Frotty and others added 3 commits September 29, 2025 14:11

Update de.peeeq.wurstscript/src/main/java/de/peeeq/wurstscript/transl…

8c88fa7

…ation/imtranslation/ImPrinter.java Co-authored-by: Copilot <[email protected]>

Update SimpleRewrites.java

f1e094a

Merge branch 'perf-improvements' of https://github.com/wurstscript/Wu…

fb48177

…rstScript into perf-improvements

Frotty requested a review from Copilot September 29, 2025 12:20

Copilot AI reviewed Sep 29, 2025

View reviewed changes

Frotty added 6 commits September 29, 2025 14:54

enable caching per default

cc35acf

enable compression for BuildMap command

16ac13c

remove graph test logging and clear subtype cache

5bf62d7

A bunch more fixes. All tests passing

2d73659

Update config.yml

7be3107

Update config.yml

10aebe0

Frotty requested a review from Copilot October 1, 2025 10:33

Copilot AI reviewed Oct 1, 2025

View reviewed changes

remove debug prints

672c7fd

Frotty requested a review from Copilot October 1, 2025 10:44

Copilot AI reviewed Oct 1, 2025

View reviewed changes

Frotty and others added 9 commits October 1, 2025 12:47

Update de.peeeq.wurstscript/src/main/java/de/peeeq/wurstscript/transl…

228f778

…ation/imtranslation/ImPrinter.java Co-authored-by: Copilot <[email protected]>

decent state

73925cc

some more good progress

a8bfb86

some caching

52d96b6

Update NameLinks.java

af896df

dont use direct buffers, make test heap settings context aware

0f2730b

fix pipeline, remove parallel streams

951d455

better vscode logging, fix caching

1212d99

Update WLoggerDefault.java

571d222

Frotty requested a review from Copilot October 2, 2025 13:35

Copilot AI reviewed Oct 2, 2025

View reviewed changes

Frotty and others added 2 commits October 2, 2025 15:38

Update de.peeeq.wurstscript/src/main/java/de/peeeq/wurstscript/interm…

7e0bedc

…ediatelang/optimizer/FunctionSplitter.java Co-authored-by: Copilot <[email protected]>

Update de.peeeq.wurstscript/src/main/java/de/peeeq/wurstscript/transl…

b40dde4

…ation/imtranslation/StackTraceInjector2.java Co-authored-by: Copilot <[email protected]>

Frotty requested a review from Copilot October 2, 2025 13:41

Copilot AI reviewed Oct 2, 2025

View reviewed changes

Frotty requested a review from Copilot October 2, 2025 23:31

Copilot AI reviewed Oct 2, 2025

View reviewed changes

Update GetCompletions.java

7bbf0cc

	final ImFunction f = work.removeLast(); // LIFO (DFS); change to removeFirst() for BFS
	final ImFunction f = work.removeLast(); // LIFO (DFS); for BFS, use removeFirst() and ensure children are added to the end

		// crude cap to avoid unbounded growth; tune as needed
		private static final int SUBTYPE_MEMO_CAP = 16_384;

-    // crude cap to avoid unbounded growth; tune as needed
-    private static final int SUBTYPE_MEMO_CAP = 16_384;
+    /**
+     * Maximum number of cached subtype checks before eviction.
+     * Default is 16,384, chosen to balance memory usage and performance for typical WurstScript projects.
+     * If the cache grows beyond this, it is fully cleared (simple eviction policy).
+     * You can override the default by setting the system property "wurst.subtypeMemoCap".
+     */
+    private static final int SUBTYPE_MEMO_CAP = Integer.getInteger("wurst.subtypeMemoCap", 16_384);

	public ImFunction getMainFunc() { return mainFunc; }
	public ImFunction getMainFunc() { return mainFunc; }
	@Nullable

		@SuppressWarnings("unchecked") final ObjectOpenHashSet<ImVar>[] in = new ObjectOpenHashSet[N];
		@SuppressWarnings("unchecked") final ObjectOpenHashSet<ImVar>[] out = new ObjectOpenHashSet[N];

	// Better not inline GetRandomInt call - it might have side effects!
	// It's better not to inline GetRandomInt call - it might have side effects!

	ArrayDeque<Element> stack = new ArrayDeque<>();
	ArrayDeque<Element> stack = new ArrayDeque<>(32);

		// If no folding happened, try regular value propagation
		if (newValue == null) {

	// Use the path-based strong component algorithm [1] on the reversed CFG.
	// Use the path-based strong component algorithm (see: R. Tarjan, "Depth-First Search and Linear Graph Algorithms", SIAM J. Comput., 1972) on the reversed CFG.

		private enum Phase { LIGHT, HEAVY }
		private Phase phase = Phase.LIGHT;

	if (v != Integer.MIN_VALUE && v <= 0) {
	if (v != Integer.MIN_VALUE) {

+    /**
+     * @deprecated The cache parameter is no longer used. Use {@link #ILInterpreter(ImProg, WurstGui, Optional, ProgramState)} instead.
+     */
+    @Deprecated
+    public ILInterpreter(ImProg prog, WurstGui gui, Optional<File> mapFile, Object cache, ProgramState globalState) {
+        this(prog, gui, mapFile, globalState);
+    }

Perf improvements #1096

Are you sure you want to change the base?

Perf improvements #1096

Uh oh!

Conversation

Frotty commented Sep 29, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Reviewed Changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

coveralls commented Sep 29, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Uh oh!

Uh oh!

Uh oh!

Copilot AI Sep 29, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Sep 29, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Uh oh!

Copilot AI Sep 29, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Uh oh!

Copilot AI Sep 29, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Sep 29, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Sep 29, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Sep 29, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Sep 29, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Uh oh!

Uh oh!

Copilot AI Oct 1, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Oct 1, 2025

Choose a reason for hiding this comment

Uh oh!

Frotty commented Sep 29, 2025 •

edited

Loading

coveralls commented Sep 29, 2025 •

edited

Loading