polytypo 1.11.0 → 1.13.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 2005b3a52f6a24868969c8c42da7d245336a2163e629e9ab275ba00ff3bf2337
-  data.tar.gz: 63fe30887b06905f4d0de3d780559d3fe6f68e4d1718a499b49d68222281690e
+  metadata.gz: 5085a8645d4f2ef3be4dfc709d3be2ddeadee79c69b4a12c3fe3f2fa771f9099
+  data.tar.gz: e6994423f5fc58dc7142b27deae4dfb1504dbda181370c9d349feb9682a3672f
 SHA512:
-  metadata.gz: 9eb140ef0546ab98091237959f7318d10a70eee753419f173aaaa7d36532ed7688bf6de432ade2c14d8d6e872b03887cd9e5128db6476aba21ff75278b3f8edd
-  data.tar.gz: ebf0496b94a294d4a9bd29e4a20ddb13152cd6e8ce333c0c25f8d8a746000b2bc1372a745cc522df7f269e20f7aeffde12d1a61c37a4ea07dbf4c903f5c38ba0
+  metadata.gz: 90dcb32715c71d9305f6d71730409c232cdf13d920a0e2173444832c534c8a1195e7149d322484de45afad87a51b8c0fb22d0a30df8cd1408fbdf43331a14ebc
+  data.tar.gz: d365d8dec671abfb9230edb85509cc8a2ef0aaa87a6df4a6dd3bedbac2bacccea7752b9cd2b15cadaf9f7f023dbe49fa3ce3244ff16457534b7e8de97cb44394

data/README.md

@@ -110,6 +110,26 @@ the `''` is left alone and the text on either side is processed. The block is al
 the body, which shows up in exactly one place: an unbalanced quotation mark in `title` cannot pair
 with a mark in your first paragraph.
 
+### Protecting regions of a text
+
+In `text` mode, `protect:` names ranges of the input that are never edited and that the rules read
+as opaque content, which is what a template hole is: `${…}` in a JavaScript template literal, an
+ICU argument, a `{{ … }}` tag. Ranges are `[start, end]` pairs of code-point offsets (not bytes and
+not UTF-16 units), half-open, ascending and non-overlapping; touching ranges are one run.
+
+```ruby
+Polytypo.transform("${x} - Rules", locale: "en-US", protect: [[0, 4]])
+# => "${x}—Rules"
+Polytypo.transform("${x} - Rules", locale: "en-GB", protect: [[0, 4]])
+# => "${x} – Rules"
+```
+
+The protected code points come back exactly as written. Edits before a range shift it, so a second
+call over the output needs the ranges at their new offsets; `transform` does not return them.
+`nil` and `[]` mean no protection. A malformed range, or a non-empty list in `html`, `markdown` or
+`yaml` mode, raises `Polytypo::Error` with `POLYTYPO_INVALID_OPTION`. `Polytypo.analyze` takes the
+same option and reports offsets into the input as passed.
+
 `require "polytypo"` never loads the native `commonmarker` extension unless you actually call
 `Polytypo.transform` with `mode: "markdown"` — the require happens lazily, inside the markdown
 pipeline only. A single `Polytypo.transform` module method (not separate

data/lib/polytypo/data/VERSION

@@ -1 +1 @@
-1.11.0
+1.13.0

data/lib/polytypo/data/fixtures/bg.json

@@ -0,0 +1,253 @@
+{
+  "spec": "1.13.0",
+  "locale": "bg",
+  "cases": [
+    {
+      "id": "bg-quotes-primary",
+      "rule": "quotes",
+      "mode": "text",
+      "in": "Той каза: \"не\" и си тръгна.",
+      "out": "Той каза: „не“ и си тръгна.",
+      "note": "ИБЕ, s.v. «кавички»: the base form is „…“ (U+201E, U+201C); the close is U+201C, not U+201D."
+    },
+    {
+      "id": "bg-quotes-nested-same-pair",
+      "rule": "quotes",
+      "mode": "text",
+      "in": "Той каза: \"стихотворението \"Кафене Славия\" е хубаво\" и спря.",
+      "out": "Той каза: „стихотворението „Кафене Славия“ е хубаво“ и спря.",
+      "note": "ОПР т. 135.1.5, via ИБЕ: a quote inside a quote takes the same „ “ pair, so the inner level is not a different glyph."
+    },
+    {
+      "id": "bg-quotes-nested-single-straight",
+      "rule": "quotes",
+      "mode": "text",
+      "in": "Той каза: \"стихотворението 'Кафене' е хубаво\".",
+      "out": "Той каза: „стихотворението „Кафене“ е хубаво“.",
+      "note": "A straight single mark inside a double quote becomes the secondary pair, which in bg equals the primary pair."
+    },
+    {
+      "id": "bg-quotes-already-typeset-kept",
+      "rule": "quotes",
+      "mode": "text",
+      "in": "Той каза: „стихотворението „Кафене Славия“ е хубаво“ и спря.",
+      "out": "Той каза: „стихотворението „Кафене Славия“ е хубаво“ и спря.",
+      "note": "Already typeset at both levels: left as it is."
+    },
+    {
+      "id": "bg-quotes-nest-at-end-declined",
+      "rule": "quotes",
+      "mode": "text",
+      "in": "Той каза: \"това е 'Славия'\".",
+      "out": "Той каза: \"това е ’Славия'\".",
+      "note": "Accepted miss: secondary equals primary, so a nest whose inner mark touches the outer one is declined by the quotes rule (quotes.md §3.2 V1, §3.5), and apostrophe then turns the inner opening mark into U+2019. Pinned as measured; the cost of the identical pair."
+    },
+    {
+      "id": "bg-quotes-nest-at-end-double-declined",
+      "rule": "quotes",
+      "mode": "text",
+      "in": "\"Кафене \"Славия\"\"",
+      "out": "\"Кафене \"Славия\"\"",
+      "note": "Accepted miss: the ОПР т. 135.1.5 shape typed with straight marks stays entirely straight; the identical pair would put two like glyphs side by side."
+    },
+    {
+      "id": "bg-quotes-nest-at-start-declined",
+      "rule": "quotes",
+      "mode": "text",
+      "in": "\"'b' a\"",
+      "out": "\"'b’ a\"",
+      "note": "Accepted miss: the inner mark at the start of the outer quote; the decline leaves a half-typeset result."
+    },
+    {
+      "id": "bg-quotes-decline-spreads",
+      "rule": "quotes",
+      "mode": "text",
+      "in": "\"x\" и \"a 'b'\"",
+      "out": "\"x\" и \"a ’b'\"",
+      "note": "Accepted miss: the decline is not local, the plain pair before the failing nest also stays straight."
+    },
+    {
+      "id": "bg-quotes-no-inner-space",
+      "rule": "quotes",
+      "mode": "text",
+      "in": "„ не “",
+      "out": "„не“",
+      "note": "ИБЕ spacing article: no space inside the marks."
+    },
+    {
+      "id": "bg-ellipsis-three-dots",
+      "rule": "ellipsis",
+      "mode": "text",
+      "in": "Чакай... какво?",
+      "out": "Чакай… какво?",
+      "note": "ИБЕ №2158: three dots, not four; the mark attaches to the preceding word."
+    },
+    {
+      "id": "bg-ellipsis-not-abbreviated",
+      "rule": "ellipsis",
+      "mode": "text",
+      "in": "Наистина?.. да",
+      "out": "Наистина?… да",
+      "note": "abbreviatedAfterTerminal is false: `?..` is a typing slip, not an abbreviated ellipsis (decision, no source)."
+    },
+    {
+      "id": "bg-dashes-parenthetical",
+      "rule": "dashes",
+      "mode": "text",
+      "in": "Утре - ако вали - няма да дойда.",
+      "out": "Утре – ако вали – няма да дойда.",
+      "note": "ИБЕ №572: a space before and after the long dash; U+2013 by decision."
+    },
+    {
+      "id": "bg-dashes-existing-em-dash",
+      "rule": "dashes",
+      "mode": "text",
+      "in": "Утре — ако вали — няма да дойда.",
+      "out": "Утре – ако вали – няма да дойда.",
+      "note": "An authored em dash becomes the locale's en dash (the DGT handbook's U+2014 is house style)."
+    },
+    {
+      "id": "bg-ranges-years-hyphen",
+      "rule": "ranges",
+      "mode": "text",
+      "in": "през 2000-2010 г. и I-II в.",
+      "out": "през 2000 – 2010 г. и I-II в.",
+      "note": "ranges.md §3: ОПР т. 103.4 prints «2000 – 2010 г.» with spaces; the Roman-numeral member is declined, so `I-II` stays.",
+      "rules": {
+        "ranges": true
+      }
+    },
+    {
+      "id": "bg-ranges-en-dash-gets-spaces",
+      "rule": "ranges",
+      "mode": "text",
+      "in": "2000–2010 г.",
+      "out": "2000 – 2010 г.",
+      "note": "ranges.md §3: a tight en dash is the pre-2012 form and takes the spaced form.",
+      "rules": {
+        "ranges": true
+      }
+    },
+    {
+      "id": "bg-ranges-spaced-kept",
+      "rule": "ranges",
+      "mode": "text",
+      "in": "2000 – 2010 г.",
+      "out": "2000 – 2010 г.",
+      "note": "ranges.md §3: the spaced en dash is already the correct form.",
+      "rules": {
+        "ranges": true
+      }
+    },
+    {
+      "id": "bg-ranges-dates",
+      "rule": "ranges",
+      "mode": "text",
+      "in": "13.10. - 20.11.2018 г.",
+      "out": "13.10. – 20.11.2018 г.",
+      "note": "ranges.md §3: ОПР т. 109.4 prints «13.10. – 20.11.2018 г.».",
+      "rules": {
+        "ranges": true
+      }
+    },
+    {
+      "id": "bg-ranges-word-hyphen-untouched",
+      "rule": "ranges",
+      "mode": "text",
+      "in": "Българо-руски речник",
+      "out": "Българо-руски речник",
+      "note": "ranges.md §3: a hyphenated word is not a digit range and is left alone.",
+      "rules": {
+        "ranges": true
+      }
+    },
+    {
+      "id": "bg-hyphen-noop",
+      "rule": "hyphen",
+      "mode": "text",
+      "in": "северо-западен и по-добър",
+      "out": "северо-западен и по-добър",
+      "note": "ОПР т. 65–66 govern syllable breaks only; no list ships, nothing changes."
+    },
+    {
+      "id": "bg-nbsp-units",
+      "rule": "nbsp",
+      "mode": "text",
+      "in": "100 км, 5 кг и 20 лв.",
+      "out": "100 км, 5 кг и 20 лв.",
+      "note": "ИБЕ №1601: a space separates the unit or abbreviation from its number."
+    },
+    {
+      "id": "bg-nbsp-year-abbreviation",
+      "rule": "nbsp",
+      "mode": "text",
+      "in": "през 2020 г. беше",
+      "out": "през 2020 г. беше",
+      "note": "ИБЕ №1601: «2020 г.» with a space."
+    },
+    {
+      "id": "bg-nbsp-percent-inert",
+      "rule": "nbsp",
+      "mode": "text",
+      "in": "5% и 5 %",
+      "out": "5% и 5 %",
+      "note": "ИБЕ №626 writes 5% without a space, so % is not a unit here and nothing is bound."
+    },
+    {
+      "id": "bg-nbsp-single-letter-unit-inert",
+      "rule": "nbsp",
+      "mode": "text",
+      "in": "10 м и 5 с",
+      "out": "10 м и 5 с",
+      "note": "Single-letter symbols are excluded by decision, not by citation (ИБЕ №1601 writes «10 м» with a space): `с` is a Bulgarian preposition."
+    },
+    {
+      "id": "bg-nbsp-short-words-inert",
+      "rule": "nbsp",
+      "mode": "text",
+      "in": "в къщата и с приятел",
+      "out": "в къщата и с приятел",
+      "note": "afterShortWords ships empty: no source gives a rule for в, с, к, и."
+    },
+    {
+      "id": "bg-nbsp-initials-unbound",
+      "rule": "nbsp",
+      "mode": "text",
+      "in": "Иван Вазов и И. Вазов",
+      "out": "Иван Вазов и И. Вазов",
+      "note": "initialBinding is none: no source states a rule on initials."
+    },
+    {
+      "id": "bg-nbsp-no-space-before-question",
+      "rule": "nbsp",
+      "mode": "text",
+      "in": "Наистина ? Да !",
+      "out": "Наистина? Да!",
+      "note": "ИБЕ spacing article: no space before ? and !, so nothing is put back."
+    },
+    {
+      "id": "bg-spaces-brackets",
+      "rule": "spaces",
+      "mode": "text",
+      "in": "Текст ( нормален ) остава  .",
+      "out": "Текст (нормален) остава.",
+      "note": "spaces.md: brackets tighten, no space before the final stop."
+    },
+    {
+      "id": "bg-symbols-copyright",
+      "rule": "symbols",
+      "mode": "text",
+      "in": "Copyright (c) 2026",
+      "out": "Copyright © 2026",
+      "note": "symbols rule, locale-neutral."
+    },
+    {
+      "id": "bg-apostrophe-foreign-name",
+      "rule": "apostrophe",
+      "mode": "text",
+      "in": "Това е McDonald's",
+      "out": "Това е McDonald’s",
+      "note": "apostrophe rule, locale-neutral."
+    }
+  ]
+}

data/lib/polytypo/data/fixtures/cs.json

@@ -1,5 +1,5 @@
 {
-  "spec": "1.11.0",
+  "spec": "1.13.0",
   "locale": "cs",
   "cases": [
     {

data/lib/polytypo/data/fixtures/de-CH.json

@@ -1,5 +1,5 @@
 {
-  "spec": "1.11.0",
+  "spec": "1.13.0",
   "locale": "de-CH",
   "cases": [
     {

data/lib/polytypo/data/fixtures/de-DE.json

@@ -1,5 +1,5 @@
 {
-  "spec": "1.11.0",
+  "spec": "1.13.0",
   "locale": "de-DE",
   "cases": [
     {
@@ -787,6 +787,82 @@
       "in": "description: 'a''--''b'\n",
       "out": "description: 'a''--''b'\n",
       "note": "modes.md §3.8.6 and §3.4: de-DE is a `-spaced` locale, so the dash would grow against a unit and the edge-growth rule discards it, as in fr. A second locale so that an implementation that does not split cannot pass on one cell alone."
+    },
+    {
+      "id": "de-de-protect-two-holes-dash",
+      "rule": "dashes",
+      "mode": "text",
+      "in": "${a} - ${b}",
+      "protect": [
+        {
+          "start": 0,
+          "end": 4
+        },
+        {
+          "start": 7,
+          "end": 11
+        }
+      ],
+      "out": "${a} – ${b}",
+      "protectOut": [
+        {
+          "start": 0,
+          "end": 4
+        },
+        {
+          "start": 7,
+          "end": 11
+        }
+      ],
+      "note": "modes.md §3.9.3: a token between two holes sees opaque content on both sides."
+    },
+    {
+      "id": "de-de-protect-dash-adjacent-growth-declined",
+      "rule": "dashes",
+      "mode": "text",
+      "in": "${a}--${b}",
+      "protect": [
+        {
+          "start": 0,
+          "end": 4
+        },
+        {
+          "start": 6,
+          "end": 10
+        }
+      ],
+      "out": "${a}--${b}",
+      "protectOut": [
+        {
+          "start": 0,
+          "end": 4
+        },
+        {
+          "start": 6,
+          "end": 10
+        }
+      ],
+      "note": "modes.md §3.9.4, accepted cost: '--' → ' – ' grows by one code point at both edges of its span, so the edge-growth rule (§3.4) discards it. The same string with no protect converts (§7.9)."
+    },
+    {
+      "id": "de-de-protect-quotes-around-hole",
+      "rule": "quotes",
+      "mode": "text",
+      "in": "Er sagte \"${name}\" zweimal.",
+      "protect": [
+        {
+          "start": 10,
+          "end": 17
+        }
+      ],
+      "out": "Er sagte „${name}“ zweimal.",
+      "protectOut": [
+        {
+          "start": 10,
+          "end": 17
+        }
+      ],
+      "note": "modes.md §3.9.3: the same pairing in a locale whose glyphs differ in shape."
     }
   ]
 }

data/lib/polytypo/data/fixtures/el.json

@@ -1,5 +1,5 @@
 {
-  "spec": "1.11.0",
+  "spec": "1.13.0",
   "locale": "el",
   "cases": [
     {

data/lib/polytypo/data/fixtures/en-GB.json

@@ -1,5 +1,5 @@
 {
-  "spec": "1.11.0",
+  "spec": "1.13.0",
   "locale": "en-GB",
   "cases": [
     {
@@ -1388,6 +1388,94 @@
       "in": "(aside)'quoted' here",
       "out": "(aside)’quoted’ here",
       "note": "apostrophe.md §6 row 25 and §7 item 9 — case 2a's accepted cost, pinned so no port can narrow or widen it silently. The author meant a quotation and both marks now read as closing glyphs. It was mismatched before spec 1.5.0 too, in the other direction: 1.4.0 gave (aside)'quoted’ here, case 3 having curled the trailing mark while the leading one matched no case. `quotes` declines the leading mark because canOpen's right-test rejects a CLOSEISH left neighbour, so a quotation opening flush after a closing delimiter is exactly the shape it cannot pair, and two neighbours cannot separate that from a possessive."
+    },
+    {
+      "id": "en-gb-protect-leading-hole-dash",
+      "rule": "dashes",
+      "mode": "text",
+      "in": "${x} - Rules",
+      "protect": [
+        {
+          "start": 0,
+          "end": 4
+        }
+      ],
+      "out": "${x} – Rules",
+      "protectOut": [
+        {
+          "start": 0,
+          "end": 4
+        }
+      ],
+      "note": "modes.md §3.9.3: the same hole in a spaced-dash locale; the three-code-point token ' - ' becomes ' – ' without touching the hole."
+    },
+    {
+      "id": "en-gb-protect-trailing-hole-dash",
+      "rule": "dashes",
+      "mode": "text",
+      "in": "Rules - ${x}",
+      "protect": [
+        {
+          "start": 8,
+          "end": 12
+        }
+      ],
+      "out": "Rules – ${x}",
+      "protectOut": [
+        {
+          "start": 8,
+          "end": 12
+        }
+      ],
+      "note": "modes.md §3.9.3: a protected run at the end of the input is a marker too."
+    },
+    {
+      "id": "en-gb-protect-icu-quoted-brace",
+      "rule": "quotes",
+      "mode": "text",
+      "in": "Call it '{b}' now.",
+      "protect": [
+        {
+          "start": 8,
+          "end": 13
+        }
+      ],
+      "out": "Call it '{b}' now.",
+      "protectOut": [
+        {
+          "start": 8,
+          "end": 13
+        }
+      ],
+      "note": "polytypo/polytypo#14: in ICU MessageFormat an apostrophe before a brace is an escape. The caller protects the quoted brace including both apostrophes and they stay straight; without protect they become U+2018/U+2019."
+    },
+    {
+      "id": "en-gb-protect-touching-ranges",
+      "rule": "dashes",
+      "mode": "text",
+      "in": "a ${x}${y} - b",
+      "protect": [
+        {
+          "start": 2,
+          "end": 6
+        },
+        {
+          "start": 6,
+          "end": 10
+        }
+      ],
+      "out": "a ${x}${y} – b",
+      "protectOut": [
+        {
+          "start": 2,
+          "end": 6
+        },
+        {
+          "start": 6,
+          "end": 10
+        }
+      ],
+      "note": "modes.md §3.9.2: touching ranges are one protected run and give one marker."
     }
   ]
 }