# Changelog

All notable changes to this package are documented here.

This package is **data**, so "breaking" mostly means *the text of a string changed*.
Releases that alter recited text say so explicitly, and list what moved.

The format follows [Keep a Changelog](https://keepachangelog.com/en/1.1.0/);
this package follows semantic versioning from 1.0.0 onward. While at `0.x`,
minor versions may change shipped text.

## [0.8.0] — 2026-09-20

A content release. **No API change**: the exported surface, every type
declaration, and every collection count are byte-identical to 0.7.0. Nothing was
added, removed, or renamed — but a large amount of text was corrected, and
**65 recited archana lines changed**. If you snapshot-test or cache strings from
this package, expect diffs.

### Added

- **Rāmcharitmanas meaning layer completed.** 938 new verse meanings (English
  prose + Hindi bhāvārtha), bringing the scripture to **1,077 / 1,077 verses**
  covered in both languages, across all seven kāṇḍs. Previously only part of the
  text carried a meaning layer.
- **`about` prose for 130 entries** that previously shipped an empty string —
  chiefly aartis and chalisas. Consumers rendering `about` no longer need an
  empty-state branch for these.

### Changed

- **2,118 existing strings revised.** The main threads:
  - **English prose is now plain ASCII.** Diacritics were removed from `en`,
    `english` and `titleEnglish` — `Kṛṣṇa` → `Krishna`, `Pūrva Pīṭhikā` →
    `Purva Pithika`, `Śarva` → `Sharva`. The `iast` fields are unaffected and
    keep full diacritics; this change is only about readable prose.
  - **British English throughout**, with Oxford `-ize` (colour, honour, centre,
    fulfil — but realize, recognize).
  - **Hindi nasals normalised** per consonant class: dental and labial take the
    conjunct (`संप्रदाय` → `सम्प्रदाय`, `परंतु` → `परन्तु`), velar and palatal
    keep anusvāra. Sanskrit-register fields are untouched.
  - **Typography and quoting** normalised; straight quotes throughout.
  - Implementation vocabulary removed from user-facing Hindi
    (`मन्त्र-मॉड्यूल` → `मन्त्र-संग्रह`).

- **Namāvalī editorial policy changed: the `ashtottara/` and `sahasranama/`
  modules now ship a corrected recension, not a verbatim transcription.**
  Both modules previously forbade word-level emendation, which meant forms no
  Sanskrit derivation can produce stayed in the litany because a source printed
  them. **65 archana lines are corrected**, each with its paired `iast`
  regenerated so the round-trip invariant still holds. Every correction is
  recorded against its shipped-before form in the relevant module's
  `SOURCES.md` in the repository.

  Representative corrections:

  | | was | now | |
  |---|---|---|---|
  | `durga` 41 | `मनोम्नन्यै` | `मनोन्मन्यै` | transposed conjunct; Manonmanī |
  | `hanuman` 23 | `सर्वयन्त्रात्मिकाय` | `सर्वयन्त्रात्मकाय` | feminine stem on a masculine ending |
  | `krishna` 52 | `…श्रीवक्षलाञ्छनाय` | `…श्रीवत्सलाञ्छनाय` | the śrīvatsa mark |
  | `lakshmi` 88 | `शुभप्रदाये` | `शुभप्रदायै` | invalid ā-stem dative |
  | `lalita` 258 | `स्वपत्न्यै` | `स्वपन्त्यै` | restored the dream-state name |
  | `vishnu` 918 | `दक्षिणस्यै` | `दक्षिणाय` | feminine dative in a masculine list |
  | `vishnu` 63 | `मङ्गल पराय` | `मङ्गलाय परस्मै` | uninflected stem beside a dative |

- **Internal provenance notes are no longer published.** `SOURCES.md` and
  `AUDIT.md` files are excluded from the npm tarball. They remain in the
  repository; they were always internal and were never meant to be displayed.

### Fixed

- Two Bhagavad Gītā verses shipped the literal string `"No changes needed."` as
  their English translation. Repaired, along with other transcription defects in
  the Gītā.
- Unicode NFC normalisation: 11 strings encoded `ड़` as U+095C while 1,347 used
  U+0921+U+093C. The same glyph in two encodings made `includes()` and search
  indexes silently miss.
- Malformed Devanāgarī: orphaned combining marks, a virāma followed by a vowel
  sign, and runs of three or more daṇḍas.
- Words wrongly fused across a line break in aartis and chalisas
  (`जग मेंबड़ा` → `जग में बड़ा`).
- Anusvāra in `iast` is now consistently `ṃ`; the mantras module had been
  using the overdot `ṁ`.

### Known issues

Carried knowingly into this release and documented in the repository:

- Four archana lines are **known to be defective and are retained unchanged**,
  because the correct reading is underdetermined rather than obvious:
  `ganesha` 78 `बलोद्धाय`, `durga` 80 `त्रिवद्यै`, `santoshi-mata` 63 `यारायै`,
  and `santoshi-mata` 85/86. Resolving them needs a printed edition.
- `santoshi-mata` 7/8 repeat `ॐ अनन्तायै नमः`. This is **correct** — the parent
  stotram carries the name in two separate verses and the list is alphabetised,
  which brings them adjacent.
- The Viṣṇu Sahasranāma repeats 109 names by design, two of them adjacently
  (663/664, 693/694). Also correct, and inherited from the recited verse.
- Most modules remain `draft` and a number of items carry `reviewPending: true`,
  pending verification against printed canonical editions. Surface a
  "provisional" caption where `reviewPending` is set.

## [0.7.0] and earlier

Not documented here; see the git history.