Renderingen-US

Radix 0.62.0

FieldValue
ProductRadix
Version0.62.0
Tagradix-v0.62.0
GitHubradix-v0.62.0
Published2026-07-16
SourceClosed for now — see Open source

Install this version#

Pinned download for Radix 0.62.0. For the current release, use Install instead.

PlatformArchiveSizeChecksum
macOS arm64radix-v0.62.0-aarch64-apple-darwin.tar.gz2.0 MBsha256
curl -fsSL -o radix.tgz \
  https://github.com/faberlang/releases/releases/download/radix-v0.62.0/radix-v0.62.0-aarch64-apple-darwin.tar.gz
tar -xzf radix.tgz
sudo mv radix-v0.62.0-aarch64-apple-darwin/radix /usr/local/bin/
radix --version

Release notes#

Synthetic retrospective minor tag covering the HIR/MIR semantic sync and stdlib merge milestone. This is the first release where fail-closed lowering invariants gate both compiler lanes, and the first where norma contracts unify across the stdlib effect interface.

Scale#

SignalCount
Commits (no merges)89
Date span2026-06-30 → 2026-06-30

Reconstruct the full log:

git log v0.61.0..v0.62.0 --oneline --no-merges

Major tracks#

HIR/MIR semantic sync — lista/tensor conversio#

The dominant theme of this release. The compiler enforces fail-closed lowering invariants — any HIR construct that lacks a MIR lowering partner now produces a hard error rather than silently dropping predicates.

  • Shaped, rectangular lista literal → tensor conversio through MIR and Rust emit (b65fb81e8, c74507750)
  • Fail-closed lista tensor MIR convert and lowering invariants (a6ab2a7c0, 039338f7e)
  • Close HIR/MIR cleanse validation across both lanes (ce0fb6024)
  • Split conversio into focused submodules (9cb660e1d)
  • Route intrinsic emit through a shared registry (9b8d8b73e)
  • Redundant conversio emit guard removed (96dda1295)
  • Centralize scalar display helpers and failable conversio channel checks (f5a05945c, 4a95ca091)

Width type sugar for listas#

Introduced per-width type aliases for lista literals (lu32, li64, lf32, lf64, …) parallel to the existing tensor width sugar, closing a long-standing parity gap.

  • Add lista width type sugar (57b905e5c)
  • Route width sugar keyword checks through a shared helper (d998c6e1e)
  • Extract and unify lista/tensor width sugar type parsers and marker helpers (3798520e0, 9168c4e3e)
  • Fix tensor sugar conversio targets (ce166c411)

Tensor bracket access policy#

Unified the bracket-indexing path through a shared tensor_bracket policy module, replacing ad-hoc typecheck/MIR/Rust emit branches.

  • Add shared tensor bracket access policy module (8d19953bd)
  • Route tensor bracket typecheck through tensor_bracket policy (be07eb5cd)
  • Use tensor_bracket list-index probe in Rust bracket emit (47f6d7f35)
  • Unify MIR tensor bracket lowering on tensor_bracket policy (9fa8ecb4c)

Stdlib merge (factory/stdlib)#

Eight-phase migration that wraps effect bodies in ad (automatic dispatch) contracts and unifies norma kernel interfaces with faber runtime types.

PhaseContent
Phase 0Norma contract unification (7efa08447)
Phase 1aSolum ad-wrapped effect bodies (db2ee8eeb)
Phase 1bProcessus sync ad bodies (f57712898)
Phase 1cAleator sync ad bodies (5dd0c867e)
Phase 1dConsolum sync ad bodies + R2a TTY (2c52cf7d4)
Phase 1eTempus clock ad bodies (e80f16684)
Phase 2Faber/norma kernel parity + Faber-vs-ad note (18089ddbe)
Phase 3Native chorda.diducta (aa830792c)

Phase 4 (native valor navigation in norma:toml, 914cad0b2) and Phase 5 closeout (toml.exige_claves and explora contract exempla, b893bef4f) round out the pipeline.

HIR/MIR stage parity fixtures#

Systematic test coverage closing the parity gap between HIR and MIR stages.

  • Close HIR MIR stage1 parity fixtures (79d002aa3)
  • Close scalar conversio parity fixtures (25f837211)
  • Align regex conversio parity (d8448dc2d)
  • Align valor scalar, boxing, recursive aggregate, and aggregate-defaults parity (ffc1e4a7c, 087adfdbc, c8168ec21, 80b5df2f0)
  • Align collection method policy (9dac01b61)
  • Decompose scalar valor conversio sync rows (3204b4707)

Lista-to-tensor polish (dedup & extraction)#

Systematic deduplication and extraction across the lista-to-tensor conversion pipeline — MIR lowering, Rust emit, typecheck, and stepper.

  • Unify lista-to-tensor MIR lowering through TensorFromFlat helper (a9d599b49)
  • Extract rank-2 literal helpers (a125e5e26)
  • Share Tensor structa emission for lista tensor conversio (fbf035aa7)
  • Align lista tensor conversio structa emit with outcome ladder (dfe343c64)
  • Tighten lista-to-tensor conversio typecheck arms (6f0f00d2f)
  • Dedupe lista-to-tensor MIR lowering and Rust emit regression assertions (64db01385, 8f5a97904)
  • Polish stepper Convert fallback path and shape helper (8fcd6cd5b, ee0cf2c90)
  • Polish lista-to-tensor structa emit helper return type (f40ad1b6d)
  • Polish tensor bracket index typecheck signature (8669860b4)

Ordering, overflow, and lexer consolidation#

  • Centralize ordering comparison emission helper and dedupe Rust emit assertions (9d71e8dfb, 17e4019ae)
  • Cross-reference scalar longitudo ordering emit coverage (f58947eda)
  • Align numerus overflow policy and polish overflow message constants (014637b67, b79e21fc6)
  • Centralize line-start comment lexer policy and dedupe regression helpers (4ed3273a8, 5f13b9652, a7ebc4f8d)

Generic param call-site unification#

  • Extend generic param call-site ratchet tests (460019de4)
  • Extract generic call-site instantiation helpers (e6b945f20)
  • Document Type::Param unification policy in infer (784931f66)

Factory planning & product framing docs#

Extensive factory documentation shaping the next campaign waves.

  • HIR/MIR semantic sync discovery ledgers (families 1–7, 8–12) and campaign staging (70694bad9, 8de340f14, a148341be)
  • Rectangular lista tensor conversio goal, Phase 1 closeout, and Phase 2 library-method expansion (0b091d0f5, 0c438ac77, 6e6521de3, 63bba16d6)
  • App-stdlib naming contract, method naming table, and renaming tracks (f65094786, dde83227c, a047619d4)
  • Architectural cleanse campaign: scope, mechanical posture, follow-up (d068e01d3, 137e9a4a1, 9c563d1cd)
  • Two-lane product framing in AGENTS.md and README.md (bdcd75220, a54fd6881)
  • Numeric type sugar spelling preference notes (31699d3d9, 9bb78fabf)
  • DEFER-066 (textus escape decoding) and DEFER-067 (si-block parse bug) documentation (88056723a, 838ecb303)

Other changes#

  • Share tensor policy helpers across faber (e01653d08)
  • Polish exempla_e2e parity harness (249ce2f71)
  • Link intentional split capability matrix (3ea45438d)

---

All releases · Install the current release