asciichem 0.29.2 → 0.29.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: f372a8539906c658a399b015255e574de5feb73d1b5e7183958901681bfbce34
-  data.tar.gz: 7269dc4ad6e39e7c5ea6726b54a8ffe1508c039bdf9ced5c0a74e23efdc9b00f
+  metadata.gz: 6aed1bf1e1bf05bd4f3459f3bf147d39f08f283dc49b8452040643171e836a30
+  data.tar.gz: 557b7506d58e8505c6b9e3ffa36da6ef7025e054850eba7a9fdee7e03503c337
 SHA512:
-  metadata.gz: 6f936bc0e95907d589c0095d34e4e65c554dcbdccd276aab7186db8f1218cbb179652b082f088f1b70a76391e830db53de2259ea8d333b64dd85ba367b16c610
-  data.tar.gz: 63d979830e4a8ab9aca450006a9740e83428e26073323af0ed4f8ddeed4be6f7aecb8bcfe1e5173f7a6049f92ff8b0a8e0a2b374fbd13630ae00b0e9e7a0db60
+  metadata.gz: b3ff4937b6f92f253c16f03d414ef4fa177ab0df85fea38c61be07e83a04e7889c2485e4b4d8b7dd3c2a7e0042711e533cdf5968e11abc130d68d21620e6ed07
+  data.tar.gz: dfeada783a748286cf51fbbb9cbaa0a8f7a485af352b511dd54a67d53fce0ec77c9b58d339d980853664ff58438f0e5be56ae6632802d70a44e12967d5418285

data/CHANGELOG.md

@@ -3,6 +3,16 @@
 All notable changes to AsciiChem are documented here.
 This project follows [Semantic Versioning](https://semver.org/).
 
+## [0.29.3] - 2026-09-22
+
+### Changed
+- Dependency floors raised to the corpus-validated versions,
+  expressed pessimistically (`~>`): `lutaml-model ~> 0.8` (from
+  `>= 0.8, < 2`), `relaton-bib ~> 2.1` (from `>= 0.1, < 3`),
+  `plurimath ~> 0.11`, `nokogiri ~> 1.18`; `chemicalml`, `elkrb`,
+  `mml`, `parslet`, `thor` unchanged. Full suite green under the
+  raised floors (1985 examples).
+
 ## [0.29.2] - 2026-09-17
 
 ### Changed

data/Gemfile

@@ -5,6 +5,7 @@ source "https://rubygems.org"
 gemspec
 
 group :development do
+  gem "parsanol"
   gem "json_schemer", "~> 2.4"
   gem "benchmark", "~> 0.4"
   gem "benchmark-ips", "~> 2.14", require: false

data/asciichem.gemspec

@@ -33,14 +33,16 @@ Gem::Specification.new do |spec|
   spec.executables   = spec.files.grep(%r{^exe/}) { |f| File.basename(f) }
   spec.require_paths = ["lib"]
 
+  # Floors are the versions the corpus suite validates against
+  # (pessimistic ~>); raise them only with a full-suite run.
   spec.add_dependency "chemicalml", "~> 0.3.0"
   spec.add_dependency "elkrb", "~> 1.0"
-  spec.add_dependency "lutaml-model", ">= 0.8", "< 2"
+  spec.add_dependency "lutaml-model", "~> 0.8"
   spec.add_dependency "mml", "~> 2.3"
-  spec.add_dependency "nokogiri", "~> 1.16"
+  spec.add_dependency "nokogiri", "~> 1.18"
   spec.add_dependency "parslet", "~> 2.0"
-  spec.add_dependency "relaton-bib", ">= 0.1", "< 3"
-  spec.add_dependency "plurimath", "~> 0.8"
+  spec.add_dependency "relaton-bib", "~> 2.1"
+  spec.add_dependency "plurimath", "~> 0.11"
   spec.add_dependency "thor", "~> 1.3"
 
   spec.add_development_dependency "json_schemer", "~> 2.4"

data/benchmarks/README.md

@@ -168,3 +168,65 @@ non-blocking (#36 bare repeated sibling captures remains open but
 is worked around in `ParsanolEngine` via single `.as(...)` capture
 wrapping).
 
+
+### Re-check 8 (2026-09-17, parsanol 1.3.27)
+
+All four upstream issues we filed are now closed (#36-#39; seven
+releases since 1.3.20). Validation:
+
+- Gate **221/221** through the shipped `ParsanolEngine`.
+- **#36 verified fixed at the source**: the bare-repeated-sibling
+  repro (`A->B->C`) now returns parslet's array-of-segment-hashes —
+  every match preserved. Our `split_merged_formula` seam in
+  `ParsanolEngine` is therefore a compatibility no-op on current
+  parsanol (it still normalizes the merged-hash shape for older
+  parsanol lines, which the opt-in floor allows).
+- **Perf: ratio-only this time.** The machine ran at load ~45
+  during measurement (parallel spec suites in other sessions), so
+  absolute numbers are meaningless — the parslet control itself
+  measured 15-20x slower than its quiet-machine baseline.
+  Same-process ratio: parsanol **~2.1x parslet** (8.3 vs 3.8 i/s,
+  and 8.8 vs 4.5 on the repeat), consistent with the 2.6x
+  quiet-machine figure from re-check 7.
+
+### Re-check 9 (2026-09-22, parsanol 1.3.49)
+
+Twenty-one releases since re-check 8, all perf-focused upstream
+(#59 roadmap: first-set BYTE_DISPATCH 1.3.29, VM memoisation
+1.3.35, VM phase-2 wiring 1.3.33, native dynamic bridge fixes
+1.3.40-1.3.41, `Parsanol::IncrementalSession` 1.3.42). Validation:
+
+- Gate **221/221** through the shipped `ParsanolEngine` (run on
+  1.3.48/1.3.49 within the same day — the line is moving fast).
+- **Perf: ratio-only again.** Load was 34-77 during measurement
+  (parslet control itself ran 8-12 i/s vs its quiet ~75), so
+  absolute numbers are excluded. Same-process ratio across three
+  runs: parsanol **1.7-3.2x parslet** (37.7/11.7, 19.0/11.1,
+  25.5/8.4), centring ~2.5-3x — consistent with the quiet-machine
+  2.6x from re-check 7; under contention the native parse path
+  degrades less than pure-Ruby parslet.
+- Upstream's incremental (`Parsanol::IncrementalSession`) and VM
+  memoisation work benefits the compat layer automatically; no
+  asciichem-side change needed or made.
+
+## Leptris note (2026-09-22, 1.9.221)
+
+Leptris is moxml's PREFERRED_ADAPTER: when installed above its
+binding floor, lutaml-model's XML layer (our CML wire path) runs on
+it. The version is fully transitive — lutaml-model constrains
+`~> 1.9.178`; asciichem pins nothing. The line moves fast
+(1.9.178 floor -> 1.9.222 within days).
+
+- **Compatibility:** full suite **1985/0** at 1.9.221.1, including
+  every CML round-trip and three-way wire spec. 1.9.222's namespace
+  fix (`xml:space` in the interleaved lane, reported upstream by
+  Canon) does not affect our documents; no action.
+- **Perf on our CML workload:** leptris is ~15-25% behind nokogiri
+  (round-trip 12.7 vs 15.0 i/s; emit 30.6 vs 39.8 i/s; load-noisy
+  ±20%). The workload is dominated by lutaml-model's Ruby-side
+  model building, not the adapter — leptris's speed gains target
+  its native parse lanes (HTML/XQuery per its release notes).
+  Measurement caveat: forcing an adapter for A/B runs requires
+  stubbing `leptris_preferred_available?` — lutaml's
+  `detect_xml_adapter` calls `runtime_default_adapter`, which
+  ignores `default_adapter=`.

data/lib/asciichem/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module AsciiChem
-  VERSION = "0.29.2"
+  VERSION = "0.29.3"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: asciichem
 version: !ruby/object:Gem::Version
-  version: 0.29.2
+  version: 0.29.3
 platform: ruby
 authors:
 - Ribose Inc.
@@ -41,22 +41,16 @@ dependencies:
   name: lutaml-model
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ">="
+    - - "~>"
       - !ruby/object:Gem::Version
         version: '0.8'
-    - - "<"
-      - !ruby/object:Gem::Version
-        version: '2'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ">="
+    - - "~>"
       - !ruby/object:Gem::Version
         version: '0.8'
-    - - "<"
-      - !ruby/object:Gem::Version
-        version: '2'
 - !ruby/object:Gem::Dependency
   name: mml
   requirement: !ruby/object:Gem::Requirement
@@ -77,14 +71,14 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.16'
+        version: '1.18'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.16'
+        version: '1.18'
 - !ruby/object:Gem::Dependency
   name: parslet
   requirement: !ruby/object:Gem::Requirement
@@ -103,36 +97,30 @@ dependencies:
   name: relaton-bib
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: '0.1'
-    - - "<"
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: '3'
+        version: '2.1'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: '0.1'
-    - - "<"
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: '3'
+        version: '2.1'
 - !ruby/object:Gem::Dependency
   name: plurimath
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.8'
+        version: '0.11'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.8'
+        version: '0.11'
 - !ruby/object:Gem::Dependency
   name: thor
   requirement: !ruby/object:Gem::Requirement