polytypo 1.11.0 → 1.12.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 2005b3a52f6a24868969c8c42da7d245336a2163e629e9ab275ba00ff3bf2337
-  data.tar.gz: 63fe30887b06905f4d0de3d780559d3fe6f68e4d1718a499b49d68222281690e
+  metadata.gz: 8c8aba5b27eb6b2b852217b3c4ba1652d78fee2fd291ce2a14e8cfc36da9ea5f
+  data.tar.gz: 321cf32a34decf2b526b6ae4231df91cf8a7d07c96b55a211b25582e701c9abe
 SHA512:
-  metadata.gz: 9eb140ef0546ab98091237959f7318d10a70eee753419f173aaaa7d36532ed7688bf6de432ade2c14d8d6e872b03887cd9e5128db6476aba21ff75278b3f8edd
-  data.tar.gz: ebf0496b94a294d4a9bd29e4a20ddb13152cd6e8ce333c0c25f8d8a746000b2bc1372a745cc522df7f269e20f7aeffde12d1a61c37a4ea07dbf4c903f5c38ba0
+  metadata.gz: d4ba1cd320aeaca4076180463c47d0d802fa3c0c9672a34326bbef2cd39ce6308b1209454792eccc2a534aec8476c840413b9f4cdddbd8cf248ef0f7205c1f0f
+  data.tar.gz: f5dd45b49cec16dcfdca6994646d6083242cfc8cb1357ec687786bfb73f24390af6b940fe49570ab0183d6507c79c4bb29581c5fdd40caa208a4bc2d3666845d

data/lib/polytypo/data/VERSION

@@ -1 +1 @@
-1.11.0
+1.12.0

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

@@ -0,0 +1,253 @@
+{
+  "spec": "1.12.0",
+  "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": "1.12.0",
   "locale": "cs",
   "cases": [
     {

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

@@ -1,5 +1,5 @@
 {
-  "spec": "1.11.0",
+  "spec": "1.12.0",
   "locale": "de-CH",
   "cases": [
     {

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

@@ -1,5 +1,5 @@
 {
-  "spec": "1.11.0",
+  "spec": "1.12.0",
   "locale": "de-DE",
   "cases": [
     {

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

@@ -1,5 +1,5 @@
 {
-  "spec": "1.11.0",
+  "spec": "1.12.0",
   "locale": "el",
   "cases": [
     {

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

@@ -1,5 +1,5 @@
 {
-  "spec": "1.11.0",
+  "spec": "1.12.0",
   "locale": "en-GB",
   "cases": [
     {

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

@@ -1,5 +1,5 @@
 {
-  "spec": "1.11.0",
+  "spec": "1.12.0",
   "locale": "en-US",
   "cases": [
     {

data/lib/polytypo/data/fixtures/es.json

@@ -1,5 +1,5 @@
 {
-  "spec": "1.11.0",
+  "spec": "1.12.0",
   "locale": "es",
   "cases": [
     {

data/lib/polytypo/data/fixtures/fi.json

@@ -1,5 +1,5 @@
 {
-  "spec": "1.11.0",
+  "spec": "1.12.0",
   "locale": "fi",
   "cases": [
     {

data/lib/polytypo/data/fixtures/fr-CA.json

@@ -1,5 +1,5 @@
 {
-  "spec": "1.11.0",
+  "spec": "1.12.0",
   "locale": "fr-CA",
   "cases": [
     {

data/lib/polytypo/data/fixtures/fr.json

@@ -1,5 +1,5 @@
 {
-  "spec": "1.11.0",
+  "spec": "1.12.0",
   "locale": "fr",
   "cases": [
     {

data/lib/polytypo/data/fixtures/hr.json

@@ -0,0 +1,253 @@
+{
+  "spec": "1.12.0",
+  "locale": "hr",
+  "cases": [
+    {
+      "id": "hr-quotes-primary",
+      "rule": "quotes",
+      "mode": "text",
+      "in": "Rekao je: \"da\" i otišao.",
+      "out": "Rekao je: „da” i otišao.",
+      "note": "Hrvatski pravopis 8.16: the base form is „…” (U+201E, U+201D); the close is U+201D, not U+201C."
+    },
+    {
+      "id": "hr-quotes-nested-secondary",
+      "rule": "quotes",
+      "mode": "text",
+      "in": "Rekao je: \"pjesma 'Lijepa naša' je himna\" i otišao.",
+      "out": "Rekao je: „pjesma ‘Lijepa naša’ je himna” i otišao.",
+      "note": "Hrvatski pravopis 8.17: a quotation inside a quotation takes the single upper pair ‘…’ (U+2018, U+2019), not the low U+201A."
+    },
+    {
+      "id": "hr-quotes-nested-at-end",
+      "rule": "quotes",
+      "mode": "text",
+      "in": "Rekao je: \"pjesma 'Lijepa naša'\".",
+      "out": "Rekao je: „pjesma ‘Lijepa naša’”.",
+      "note": "Hrvatski pravopis 8.17: the inner level may end where the outer one does; the secondary pair differs from the primary, so nothing is declined."
+    },
+    {
+      "id": "hr-quotes-two-pairs",
+      "rule": "quotes",
+      "mode": "text",
+      "in": "x \"a\" y \"b 'c'\"",
+      "out": "x „a” y „b ‘c’”",
+      "note": "Two sibling pairs, the second with a nest: both are typeset."
+    },
+    {
+      "id": "hr-quotes-already-typeset-kept",
+      "rule": "quotes",
+      "mode": "text",
+      "in": "Rekao je: „pjesma ‘Lijepa naša’ je himna” i otišao.",
+      "out": "Rekao je: „pjesma ‘Lijepa naša’ je himna” i otišao.",
+      "note": "Already typeset at both levels: left as it is."
+    },
+    {
+      "id": "hr-quotes-no-inner-space",
+      "rule": "quotes",
+      "mode": "text",
+      "in": "„ da ”",
+      "out": "„da”",
+      "note": "Hrvatski pravopis 8.16: no space after the opening or before the closing mark."
+    },
+    {
+      "id": "hr-ellipsis-three-dots",
+      "rule": "ellipsis",
+      "mode": "text",
+      "in": "Čekaj... što?",
+      "out": "Čekaj… što?",
+      "note": "Hrvatski pravopis 8.10: three dots become the single ellipsis character."
+    },
+    {
+      "id": "hr-ellipsis-not-abbreviated",
+      "rule": "ellipsis",
+      "mode": "text",
+      "in": "Što?.. da",
+      "out": "Što?… da",
+      "note": "abbreviatedAfterTerminal is false: `?..` is an ordinary ellipsis after the question mark, not a two-dot abbreviated one (decision; neither 8.10 nor 8.3 states a two-dot form)."
+    },
+    {
+      "id": "hr-dashes-parenthetical",
+      "rule": "dashes",
+      "mode": "text",
+      "in": "Idemo - ako pada - kasnije.",
+      "out": "Idemo – ako pada – kasnije.",
+      "note": "Hrvatski pravopis 8.14: the dash is written with a space on each side; U+2013 read from rendered text."
+    },
+    {
+      "id": "hr-dashes-existing-em-dash",
+      "rule": "dashes",
+      "mode": "text",
+      "in": "Idemo — ako pada — kasnije.",
+      "out": "Idemo – ako pada – kasnije.",
+      "note": "An authored em dash becomes the locale's en dash: 8.14 does not mention U+2014."
+    },
+    {
+      "id": "hr-ranges-year-tight-untouched",
+      "rule": "ranges",
+      "mode": "text",
+      "in": "2010.-2012.",
+      "out": "2010.-2012.",
+      "note": "ranges.md §3.2: the left flank of the hyphen is the ordinal period U+002E, not a digit, so this attested year form is not a range candidate and stays as typed.",
+      "rules": {
+        "ranges": true
+      }
+    },
+    {
+      "id": "hr-ranges-year-spaced-gets-en-dash",
+      "rule": "ranges",
+      "mode": "text",
+      "in": "2010. - 2012.",
+      "out": "2010. – 2012.",
+      "note": "ranges.md §3: a spaced hyphen between years is taken by the dashes rule and becomes the spaced en dash that 8.14 prints («2010. – 2012.»).",
+      "rules": {
+        "ranges": true
+      }
+    },
+    {
+      "id": "hr-ranges-digits",
+      "rule": "ranges",
+      "mode": "text",
+      "in": "Posjeti: 15-17 i 30-45",
+      "out": "Posjeti: 15 – 17 i 30 – 45",
+      "note": "ranges.md §3: 8.1 prints «Posjeti: 15 – 17»; digit-to-digit hyphens take the spaced en dash.",
+      "rules": {
+        "ranges": true
+      }
+    },
+    {
+      "id": "hr-ranges-page-range-spaced",
+      "rule": "ranges",
+      "mode": "text",
+      "in": "str. 145-162",
+      "out": "str. 145 – 162",
+      "note": "ranges.md §3: ACCEPTED COST. 8.14 writes bibliographic page ranges tight («str. 145–162»), but locale data cannot select by context, so the rule spaces them (the ranges rule is off by default).",
+      "rules": {
+        "ranges": true
+      }
+    },
+    {
+      "id": "hr-ranges-word-hyphen-untouched",
+      "rule": "ranges",
+      "mode": "text",
+      "in": "Zagreb-Split i hrvatsko-engleski",
+      "out": "Zagreb-Split i hrvatsko-engleski",
+      "note": "ranges.md §3: a hyphenated word is not a digit range and is left alone.",
+      "rules": {
+        "ranges": true
+      }
+    },
+    {
+      "id": "hr-hyphen-noop",
+      "rule": "hyphen",
+      "mode": "text",
+      "in": "hrvatsko-engleski rječnik i bez-",
+      "out": "hrvatsko-engleski rječnik i bez-",
+      "note": "Hrvatski pravopis 8.13: breaking at a hyphen is provided for and no closed list ships, so nothing changes."
+    },
+    {
+      "id": "hr-nbsp-units",
+      "rule": "nbsp",
+      "mode": "text",
+      "in": "3 kg, 50 %, 0,5 ‰, 99,90 € i 5 °C",
+      "out": "3 kg, 50 %, 0,5 ‰, 99,90 € i 5 °C",
+      "note": "Hrvatski pravopis 8.1 and 8.19.11: a space separates the number from the unit, percent, per-mille, currency sign and degree Celsius."
+    },
+    {
+      "id": "hr-nbsp-single-letter-unit-inert",
+      "rule": "nbsp",
+      "mode": "text",
+      "in": "2 m, 220 V i 60 W",
+      "out": "2 m, 220 V i 60 W",
+      "note": "Single-letter symbols are excluded by decision (8.1 writes «2 m» with a space), as in sk and bg."
+    },
+    {
+      "id": "hr-nbsp-defunct-currency-inert",
+      "rule": "nbsp",
+      "mode": "text",
+      "in": "34 kn",
+      "out": "34 kn",
+      "note": "kn is excluded as defunct (decision, not citation); 8.1 still prints «34 kn»."
+    },
+    {
+      "id": "hr-nbsp-article-sign",
+      "rule": "nbsp",
+      "mode": "text",
+      "in": "§ 25. i § 3",
+      "out": "§ 25. i § 3",
+      "note": "Hrvatski pravopis 8.1 and 8.19.8: the section sign is separated from its number by a space."
+    },
+    {
+      "id": "hr-nbsp-page-abbreviation",
+      "rule": "nbsp",
+      "mode": "text",
+      "in": "str. 12",
+      "out": "str. 12",
+      "note": "beforeNumber lists str., from «str. 145–162» in 8.14 (form evidence; a decision, MEDIUM)."
+    },
+    {
+      "id": "hr-nbsp-complex-abbreviations",
+      "rule": "nbsp",
+      "mode": "text",
+      "in": "tvrtka d. o. o. i dipl. ing. Ivo",
+      "out": "tvrtka d. o. o. i dipl. ing. Ivo",
+      "note": "Hrvatski pravopis 8.13 rule 6: complex abbreviations such as dipl. ing. and d. o. o. should not be split at a line end."
+    },
+    {
+      "id": "hr-nbsp-other-abbreviations-inert",
+      "rule": "nbsp",
+      "mode": "text",
+      "in": "d. d. i dr. sc. Ana",
+      "out": "d. d. i dr. sc. Ana",
+      "note": "The other spaced abbreviations of 7.1 are not listed: extending 8.13's line-break intent to them would be a derivation, not a citation."
+    },
+    {
+      "id": "hr-nbsp-short-words-inert",
+      "rule": "nbsp",
+      "mode": "text",
+      "in": "u kući i s prijateljem",
+      "out": "u kući i s prijateljem",
+      "note": "afterShortWords ships empty: no online section of the pravopis states a rule for one-letter prepositions or conjunctions at a line end."
+    },
+    {
+      "id": "hr-nbsp-initials-unbound",
+      "rule": "nbsp",
+      "mode": "text",
+      "in": "A. G. Matoš",
+      "out": "A. G. Matoš",
+      "note": "initialBinding is none: 7.3 and 8.1 give the spaced form of initials and say nothing about line breaks."
+    },
+    {
+      "id": "hr-nbsp-no-space-before-question",
+      "rule": "nbsp",
+      "mode": "text",
+      "in": "Što ? Da !",
+      "out": "Što? Da!",
+      "note": "Hrvatski pravopis 8.1: no space before ? and !, so nothing is put back."
+    },
+    {
+      "id": "hr-spaces-brackets",
+      "rule": "spaces",
+      "mode": "text",
+      "in": "Tekst ( normalan ) ostaje  .",
+      "out": "Tekst (normalan) ostaje.",
+      "note": "spaces.md: brackets tighten, no space before the final stop."
+    },
+    {
+      "id": "hr-symbols-copyright",
+      "rule": "symbols",
+      "mode": "text",
+      "in": "Copyright (c) 2026",
+      "out": "Copyright © 2026",
+      "note": "symbols rule, locale-neutral."
+    },
+    {
+      "id": "hr-apostrophe-foreign-name",
+      "rule": "apostrophe",
+      "mode": "text",
+      "in": "To je McDonald's",
+      "out": "To je McDonald’s",
+      "note": "apostrophe rule, locale-neutral; 8.18 uses U+2019 for the elision mark as well."
+    }
+  ]
+}