[Issue Backlog] Backlog Analysis - March 15, 2026

[github-actions[bot]]

Issue Backlog Analysis — March 15, 2026

Executive Summary

• Total open issues analyzed: 138
• Issues recommended for closure: 1
• Potential duplicates / merge candidates: 7 clusters (carried from previous analyses)
• Issues with suggested fixes: 8
• Issues commented on this run: 2 (newly surfaced, never commented)

Previous run: 2026-03-13 · Issues already commented in last 4 days: #8893, #8740, #8194 (skipped to avoid duplicate noise)

---

Issues Recommended for Closure

Issue	Title	Reason for Closure
#7958	Z3 4.15.3 is not on opam	Fully answered: opam is maintained externally by community. Commenting was done on 2026-02-27. No Z3 team action items remain.

---

Potential Duplicates / Merge Candidates

Issue	Title	Duplicate of	Notes
#6701	JavaScript & TypeScript API Runs Forever	#6512	Same JS/WASM thread non-termination root cause
#7070	NodeJS doesn't clean up threads/workers	#6512	Same JS/WASM thread cleanup root cause
#5614	Perf regression arith.solver=2 → arith.solver=6	#4934	Same arith.solver=6 regression cluster
#7321	Invalid model issue on floats	#7162	FP invalid model cluster (also: #7431, #7842, #8185)
#7431	Invalid model issue on float formula	#7162	FP invalid model cluster
#7841	Incorrect model for QF_SLIA	#8023	String/Seq invalid model cluster (also: #7664, #7593, #6346, #6982)
#8940	Is there a way to combine QF_BV and QF_IDL?	#4934	arith.solver=6 underperforms for BV+IDL; workaround: smt.arith.solver=3; answered by maintainer

---

Issues with Suggested Fixes

#7135 — Refutational Soundness Issue (QF_BVFP)

• Component: Float+Bitvector theory (src/ast/fpa/, src/smt/theory_fpa*)
• Relevant source files: src/ast/fpa/fpa_rewriter.cpp, src/smt/theory_fpa.cpp, src/util/mpf.cpp
• Suggested fix direction: Z3 returns unsat where cvc5 produces a validated sat model. Likely a reasoning error in to_fp_unsigned or bvsdiv interaction with floating-point conversion. Related to [consolidated] new core, floats #7026 (consolidated floats) cluster. Needs test with current version.

#7132 — Unsound Model with ABV Logic

• Component: Array theory model generation (src/smt/theory_array*, src/smt/smt_model_generator.cpp)
• Relevant source files: src/smt/theory_array.cpp, src/smt/theory_array_base.cpp
• Suggested fix direction: The generated model for array select operations violates constraints asserted in the problem. Model builder constructs a composed function (k!149 ∘ k!143!151) that yields incorrect values. Bug in as-array model construction for BV-indexed arrays.

#8740 — solver.check() returns 'unknown' instead of expected result

• Component: nlsat solver (src/nlsat/)
• Relevant source files: src/nlsat/nlsat_solver.cpp, src/math/polynomial/
• Suggested fix direction: Bisected to commit 7395152 (PR factor out coi, use polynomial elaboration for nlsat solver #8039, "polynomial elaboration for nlsat"). Polynomial elaboration change appears to introduce a completeness regression in QF_NIA.

#7991 — Infinite loop/performance regression in Z3 4.15.3 (BV+Quantifiers)

• Component: BV rewriter, quantifier instantiation (src/ast/rewriter/, src/smt/)
• Relevant source files: src/ast/rewriter/bv_rewriter.cpp
• Suggested fix direction: Bisected to commit d66609e (Sept 22, 2024). Workaround: rewriter.bv_le2extract=false rewriter.bv_sort_ac=true. The bv_le2extract rewrite destroys E-matching patterns for quantifiers over BV expressions.

#8023 — Soundness issue in string logic

• Component: String/Seq solver (src/smt/seq_*.cpp, src/smt/theory_seq*)
• Relevant source files: src/smt/seq_unicode.cpp, src/smt/seq_solver.cpp
• Suggested fix direction: Invalid model produced for QF_SLIA with sequence operations. Related cluster: Incorrect model for QF_SLIA #7841, invalid model #7664, Invalid model issue with sequences: foldl, nth and ++ #7593, invalid model issue on an incremental string formula #6346, invalid model issue on an incremental QF_SLIA formula #6982.

#7730 — Spacer returns sat instead of unsat

• Component: Spacer Horn solver (src/muz/spacer/)
• Relevant source files: src/muz/spacer/spacer_mbu.cpp (native MBP)
• Suggested fix direction: Bisection shows fp.spacer.native_mbp=true triggers bug; setting to false restores unsat. Native MBP implementation has unsoundness for this class of Horn problems. Assigned to @agurfinkel.

#8893 — String solver performance regression in 4.16 (QF_SLIA/regex)

• Component: String solver, regex automaton (src/smt/seq_regex.cpp)
• Relevant source files: src/smt/seq_regex.cpp, src/smt/seq_solver.cpp
• Suggested fix direction: 25x slowdown on regex-heavy QF_SLIA. Compare regex automaton construction commits between 4.15.2 and 4.16.0. Related to Performance degradation in Z3 on QF_SLIA instances with fixed-length strings and prefix/suffix constraints #8096 (fixed-length string perf). Assigned to @NikolajBjorner, @Copilot.

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

• Component: OCaml bindings build system (src/api/ml/CMakeLists.txt)
• Relevant source files: src/api/ml/CMakeLists.txt, cmake/FindOCaml.cmake
• Suggested fix direction: WIP fix exists (commit f7aec02). API changes since last working version need to be incorporated. Contributor @arbipher is working on it.

---

Issues Needing More Information

Issue	Title	Missing Information
#8052	Unsat only with unnecessary asserts	Needs minimal reproducible SMT-LIB2 formula (file attached but not accessible to tooling); clarify Z3 version and whether it reproduces in 4.16
#7592	PartialOrder hangs on bottom element	No SMT-LIB2 example provided inline; needs reproducible test case
#7185	Error("invalid argument") when no expr update	API usage example needed; Python or C API?
#6303	Z3 switches unknown/unsat by changing range sort	Would benefit from bisection to find regressing commit

---

Notable Issues Deserving Attention

Issue	Title	Age	Notes
#438	QF_ABV query causes Z3 to hang	~10.1 years	Oldest open issue; classic test case for non-termination
#509	shell/libz3/test-z3 should use same generated files	~10.0 years	Long-standing build consistency issue
#578	Infinite loop for recursive datatype + quantifiers	~9.9 years	Model generation with recursive types
#1241	Spacer/PDR report incomplete/spurious models	~8.5 years	23 comments; ongoing Horn solver correctness
#7026	[consolidated] new core, floats	~2.3 years	Collector issue for float bugs in EUF core; several assertion violations
#7164	[consolidated] hang on small instances (string/BV)	~2.0 years	10 comments; collector for completeness issues
#8023	Soundness issue in string logic	~4 months	Critical; string model is unsound
#7991	BV+Quantifier infinite loop regression	~5 months	14 comments; has bisection + workaround, needs fix
#7730	Spacer returns sat instead of unsat	~8 months	Critical soundness; assigned to @agurfinkel
#8893	String solver 25x perf regression in 4.16	6 days	Newly filed; actively assigned

---

Cluster Summary

String/Seq soundness cluster (canonical: #8023): #7841, #7664, #7593, #6346, #6982 — 6 issues total. No fix yet.

Float invalid model cluster (canonical: #7162): #7321, #7431, #7842, #8185 — 5 issues total. Also related: #7026, #6633, #6078, #7135.

arith.solver=6 performance cluster (canonical: #4934): #5614, #7502, #8940 — Long-standing; workaround is smt.arith.solver=2 or smt.arith.solver=3.

JS/WASM thread cleanup (canonical: #6512): #6701, #7070 — JS bindings don't terminate workers.

Horn/Spacer soundness: #7730, #7611, #6471, #4008, #2641 — Spacer has multiple open soundness reports.

---

Automated by Issue Backlog Processor — runs every 2 days
This run analyzed 138 open issues (100 + 38, 2 pages). Previous run: 2026-03-13.

AI generated by Issue Backlog Processor · history

• expires on Mar 22, 2026, 12:22 AM UTC

[github-actions[bot]]

This discussion has been marked as outdated by Issue Backlog Processor.

A newer discussion is available at Discussion #9019.