[Issue Backlog] Backlog Analysis - February 27, 2026

[github-actions[bot]]

Issue Backlog Analysis - 2026-02-27

Executive Summary

• Total open issues analyzed: 145
• Critical soundness bugs: 20
• Performance regressions: 25
• API/binding issues: 15
• String solver issues: 7
• Days since last backlog run: 2

---

Issues Recommended for Closure

After careful review, one issue can be safely closed:

Issue	Title	Reason for Closure
#7958	Z3 4.15.3 is not on opam	Comments confirm OCaml opam package availability is managed externally by package maintainers; Z3 releases are published to PyPI, not opam. Issue is a question that has been answered.

---

Critical Soundness Bugs (High Priority)

These issues represent potential correctness problems and should be prioritized:

#8185 - Incorrect model in mixed FP/Real logic triggered by extra string constraint

• Component: String solver, floating-point arithmetic, proof generation
• Relevant source: src/smt/theory_str.cpp, src/smt/theory_fpa.cpp, proof generation
• Issue: Model validation fails when :produce-proofs true is set with mixed FP/Real logic plus string constraints
• Status: Assigned to maintainers, 0 comments
• Suggested action: Investigate interaction between proof generation and mixed theory solving

#8194 - Incorrect model with :produce-proofs true involving String Theory and UF

• Component: Proof generation, string solver, uninterpreted functions
• Relevant source: src/smt/theory_str.cpp, proof infrastructure
• Issue: Valid model generated without proof flag, invalid model with proof flag enabled
• Status: Assigned, 1 comment
• Suggested action: Related to Incorrect model in mixed FP/Real logic triggered by extra string constraint #8185 - both triggered by proof generation with strings

#8023 - Soundness issue in string logic

• Component: String solver (src/smt/theory_str.cpp)
• Relevant source: String theory implementation, str.replace, str.indexof handling
• Issue: Z3 returns unsat, cvc5 returns sat with valid model
• Status: 8 comments with additional test cases, active investigation
• Suggested action: Deep dive into string solver axioms for str.replace operations

#8740 - Z3 Regression: solver.check() returns 'unknown' instead of 'sat'

• Component: Arithmetic solver, core SMT engine
• Relevant source: Core solving logic, bisected to commit 7395152
• Issue: Versions 4.15.8+ return unknown on satisfiable constraints that 4.15.4 solves correctly
• Status: Contributor identified the regression commit in comments
• Suggested action: Revert or fix commit 7395152, confirmed by cvc5 returning sat

#2650 - Incorrect result in NRA formula 2

• Component: Nonlinear real arithmetic (NRA)
• Age: >3 years
• Status: Long-standing soundness issue in NRA solver

#7990 - Wrong verdict with (set-option :sat.smt true)

• Component: SAT-SMT integration
• Issue: Configuration option changes solver verdict incorrectly

#7841, #7842 - Incorrect models for QF_SLIA

• Component: String solver with linear integer arithmetic
• Related to: Soundness issue in string logic #8023 string solver issues

#7730 - Spacer returns sat instead of expected unsat

• Component: Spacer CHC solver
• Issue: Model validation fails

---

Performance Regressions Needing Attention

#7502 - Regression with arith.solver=6

• Component: Arithmetic solver (src/smt/theory_arith*)
• Bisected to: Commit 28c827f
• Impact: HACL* verification went from minutes to >1 hour
• Status: Assigned, 9 comments with test cases
• Suggested fix: Review commit that enforces arith.solver=6, investigate rlimit accounting differences

#8019 - Massive performance regression in 4.15.4.0 due to lock contention

• Component: Thread synchronization, core infrastructure
• Impact: Lock contention causing significant slowdown
• Suggested investigation: Profile lock usage, consider lock-free data structures

#7991 - Infinite loop/performance regression in Z3 4.15.3

• Versions affected: 4.15.3+ (worked in 4.12.6)
• Impact: 15 seconds → non-termination, 30GB RAM usage
• Status: 14 comments, test case from CBMC
• Related to: Other 4.15.x regressions

#7951 - Another seemingly infinite loop

#7800 - Memory and compute regression when upgrade to 4.15.3

#7735 - Memory and time blow-up with z3 4.15.2

Pattern: Multiple performance regressions introduced in 4.15.x series, suggesting a systemic change that needs investigation.

---

API and Binding Issues

#6512 - NodeJS process won't stop after using 'solve' or 'solver.check()'

• Component: JavaScript/WASM bindings
• Status: Previously commented in last run (2 days ago)
• Impact: Thread non-termination blocking Node.js exit
• Duplicates: Javascript & Typescript API Runs Forever #6701, [NodeJS] Z3 doesn't seem to clean up threads/workers after finishing its job #7070 (per previous analysis)

#7684 - OCaml bindings don't install with cmake build system

• Component: CMake build system, OCaml bindings
• Status: 18 comments, affecting nixpkgs
• Issue: Python build works, CMake doesn't install OCaml outputs
• Suggested fix: Add CMake install rules for OCaml bindings in src/api/ml/CMakeLists.txt

#7687 - Inconsistent results in CLI and Z3Py API

• Component: Python bindings vs. CLI
• Impact: Same formula gives different results in different interfaces

---

String Solver Issues (Component Focus)

Multiple string solver issues suggest this component needs focused attention:

• Soundness issue in string logic #8023: Soundness with str.replace/str.indexof
• Incorrect model in mixed FP/Real logic triggered by extra string constraint #8185: Incorrect model with proof generation (FP + String)
• Incorrect model with :produce-proofs true involving String Theory and UF #8194: Incorrect model with proof generation (UF + String)
• Incorrect model for QF_SLIA #7841: Incorrect model for QF_SLIA
• Invalid model issue on String theory  #7223: Invalid model issue on String theory
• [consolidated] hang on small instances (string, bit-vector) #7164: Hang on small string instances
• invalid model issue on an incremental string formula #6346: Invalid model on incremental string formula

Recommendation: Systematic audit of string solver, particularly:

• Interaction with proof generation
• str.replace, str.indexof, str.substr axioms
• Mixed theory solving with strings

---

Notable Long-Standing Issues

Issue	Title	Age	Comments
#2650	Incorrect result in NRA formula 2	>3 years	Open
#2641	QE soundness bug	>3 years	Open
#4175	(qe) unsoundness	>2 years	Open
#4008	Soundness issue	>2 years	Open
#4934	arith.solver=6 performance	>2 years	Open
#6788	Finite sets	>1 year	Open

---

Duplicate Clusters (From Previous Analysis)

Based on previous runs:

Canonical Issue	Duplicates	Topic
#6512	#6701, #7070	JS WASM thread non-termination
#4934	#5614, #7502	arith.solver=6 performance

---

Issues Needing More Information

Issue	Title	Missing Information
#8096	Performance degradation in Z3 on QF_SLIA	Need bisection to identify regression commit
#8052	Unsat only with unnecessary asserts	Need minimal reproducer
#7677	Potential issue in arithmetic optimization	Need verification if issue still present

---

Recommendations for Maintainers

1. Immediate Priority: Address soundness bugs (Incorrect model in mixed FP/Real logic triggered by extra string constraint #8185, Incorrect model with :produce-proofs true involving String Theory and UF #8194, Soundness issue in string logic #8023, Z3 Regression: solver.check() returns 'unknown' instead of 'sat' #8740)

   • Z3 Regression: solver.check() returns 'unknown' instead of 'sat' #8740 has regression commit identified - can be fixed quickly
   • Incorrect model in mixed FP/Real logic triggered by extra string constraint #8185 and Incorrect model with :produce-proofs true involving String Theory and UF #8194 both involve proof generation + strings - likely related

2. 4.15.x Regression Investigation: Multiple performance regressions in 4.15.x series

   • Consider 4.15.x hotfix release after identifying root causes
   • Issues: Infinite loop/performance regression in Z3 4.15.3 #7991, Another seemingly infinite loop #7951, Observe both compute and memory regression when upgrade to 4.15.3 from 4.14.1 #7800, Memory and time blow-up with z3 4.15.2 #7735, Massive performance regression in 4.15.4.0 apparently due to lock contention #8019

3. String Solver Focus: 7+ issues point to string solver needing attention

   • Particularly proof generation interaction
   • Mixed theory scenarios

4. Build System Issues: CMake OCaml bindings (OCaml bindings don't install with cmake build system #7684)

5. Close #7958: Package distribution question already answered

---

Automated by Issue Backlog Processor - Last run: 2026-02-27 - Runs every 2 days

AI generated by Issue Backlog Processor

• expires on Mar 6, 2026, 12:20 AM UTC

[github-actions[bot]]

This discussion was automatically closed because it expired on 2026-03-06T00:20:39.565Z.

Closed by Workflow