interscript-maps 2.5.2 → 3.0.0

data/maps/un-bul-Cyrl-Latn-1977.isc

@@ -12,17 +12,32 @@ metadata {
   description {    The System of Transliteration and Transcription of Bulgarian Geographical Names into Roman letters,
     as shown in the annex below, be adopted as the international system for the romanization of Bulgarian geographical names.
   }
+
+  notes {
+    note "Diffed against the UN WGRS Report on the Current Status of UN Romanization\n Systems for Geographical Names, version 4.0 (February 2013)\n (eki.ee/wgrs/rom1_bg.htm; also unstats wgrr4bulgarian): the 1977 system\n (res. III/10, adopting the 1972 Council of Orthography system) is printed there\n as a comparison against the superseding 2009/2012 system; all 30 rows verified."
+
+    note "Rows ю and я now render ju/ja per the 1977 system; the map previously\n carried yu/ya, the values of the superseding current system (2009 Transliteration\n Act, UN-approved 2012 X/7). Reissue note: the current system differs from 1977\n in zh/j/ts/ch/sh/sht/a/y/yu/ya; it is NOT encoded here."
+
+    note "The full text of res. III/10 (eki.ee/wgrs/res/res_3_10.htm) is not in the\n reference library; the 1977 rows were verified against the UN WGRS 4.0\n comparison section, which lists every row in which the 1977 system differs\n from the current one. The -iya -> -ia word-final footnote of the current table\n is likewise not attested for the 1977 system and is not encoded."
+  }
 }
 
 tests {
+   "ю" ->  "ju"
+   "я" ->  "ja"
+   "Ю" ->  "Ju"
+   "Я" ->  "Ja"
+   "ь" ->  "j"
+   "ъ" ->  "ǎ"
+   "щ" ->  "št"
    "нунатак Абрит" ->  "nunatak Abrit"
-   "връх Академия" ->  "vrǎh Akademiya"
+   "връх Академия" ->  "vrǎh Akademija"
    "връх Ами Буе" ->  "vrǎh Ami Bue"
    "нос Айтос" ->  "nos Ajtos"
    "залив Баба Тонка" ->  "zaliv Baba Tonka"
    "Балабански камък" ->  "Balabanski kamǎk"
    "Бедечки поток" ->  "Bedečki potok"
-   "нос Бяга БЯГА" ->  "nos Byaga BYAGA"
+   "нос Бяга БЯГА" ->  "nos Bjaga BJAGA"
    "остров Качо" ->  "ostrov Kačo" # wiki ostrov Cacho
    "Чакъров остров" ->  "Čakǎrov ostrov"
    "връх Дъбник" ->  "vrǎh Dǎbnik"
@@ -32,10 +47,10 @@ tests {
    "Нунатак Едуард" ->  "Nunatak Eduard"
    "Елховска седловина" ->  "Elhovska sedlovina"
    "ледник Етър" ->  "lednik Etǎr"
-   "нунатак Филип Тотю" ->  "nunatak Filip Totyu"
+   "нунатак Филип Тотю" ->  "nunatak Filip Totju"
    "ледник Габаре" ->  "lednik Gabare"
    "риф Гергини" ->  "rif Gergini"
-   "Гяуров връх" ->  "Gyaurov vrǎh"
+   "Гяуров връх" ->  "Gjaurov vrǎh"
    "Гуцалски рид" ->  "Gucalski rid" # not found
    "Хараламбиев остров" ->  "Haralambiev ostrov"
    "връх Ичера" ->  "vrǎh Ičera"
@@ -45,36 +60,36 @@ tests {
    "нос Кърджали" ->  "nos Kǎrdžali"
    "седловина Кърнаре" ->  "sedlovina Kǎrnare"
    "нунатак Керсеблепт" ->  "nunatak Kerseblept"
-   "Кондофрейски възвишения" ->  "Kondofrejski vǎzvišeniya"
+   "Кондофрейски възвишения" ->  "Kondofrejski vǎzvišenija"
    "Костинбродски проход" ->  "Kostinbrodski prohod"
    "връх Кожух" ->  "vrǎh Kožuh"
    "Кукерски нунатаци" ->  "Kukerski nunataci"
-   "залив Лазурен бряг" ->  "zaliv Lazuren bryag"
+   "залив Лазурен бряг" ->  "zaliv Lazuren brjag"
    "връх Лудогорие" ->  "vrǎh Ludogorie"
-   "Лютибродски скали" ->  "Lyutibrodski skali"
+   "Лютибродски скали" ->  "Ljutibrodski skali"
    "Масларов нунатак" ->  "Maslarov nunatak"
    "Михневски връх" ->  "Mihnevski vrǎh"
    "рид Митино" ->  "rid Mitino"
-   "езеро Наяда" ->  "ezero Nayada"
-   "нос Никюп НИКЮП" ->  "nos Nikyup NIKYUP"
+   "езеро Наяда" ->  "ezero Najada"
+   "нос Никюп НИКЮП" ->  "nos Nikjup NIKJUP"
    "рид Оборище ОБОРИЩЕ" ->  "rid Oborište OBORIŠTE"
    "залив Олуша" ->  "zaliv Oluša"
-   "Оряховски възвишения" ->  "Oryahovski vǎzvišeniya"
+   "Оряховски възвишения" ->  "Orjahovski vǎzvišenija"
    "нунатак Памидово" ->  "nunatak Pamidovo"
    "връх Парангалица" ->  "vrǎh Parangalica"
    "Първомайски провлак" ->  "Pǎrvomajski provlak"
    "ледник Патлейна" ->  "lednik Patlejna"
    "полуостров Перник" ->  "poluostrov Pernik"
    "връх Петко Войвода" ->  "vrǎh Petko Vojvoda"
-   "остров Фанагория" ->  "ostrov Fanagoriya"
+   "остров Фанагория" ->  "ostrov Fanagorija"
    "нос Плас" ->  "nos Plas"
-   "Пресиянов рид" ->  "Presiyanov rid"
+   "Пресиянов рид" ->  "Presijanov rid"
    "връх Принсипе де Астуриас" ->  "vrǎh Prinsipe de Asturias" # wiki: vrǎh Príncipe de Asturias
    "нунатак Ръченица" ->  "nunatak Rǎčenica"
-   "връх Райна Княгиня" ->  "vrǎh Rajna Knyaginya"
+   "връх Райна Княгиня" ->  "vrǎh Rajna Knjaginja"
    "Рид Ръжана" ->  "Rid Rǎžana"
    "връх Ригс" ->  "vrǎh Rigs"
-   "остров Рогулят" ->  "ostrov Rogulyat"
+   "остров Рогулят" ->  "ostrov Roguljat"
    "ледник Сабазий" ->  "lednik Sabazij"
    "ледник Съединение" ->  "lednik Sǎedinenie"
    "нунатак Сенокос" ->  "nunatak Senokos"
@@ -84,37 +99,37 @@ tests {
    "ледник Сигмен" ->  "lednik Sigmen"
    "Седловина Синитово" ->  "Sedlovina Sinitovo"
    "Ледник Скаплизо" ->  "Lednik Skaplizo"
-   "залив Слънчев бряг" ->  "zaliv Slǎnčev bryag"
+   "залив Слънчев бряг" ->  "zaliv Slǎnčev brjag"
    "остров Соатрис" ->  "ostrov Soatris"
    "планина Софийски Университет" ->  "planina Sofijski Universitet"
    "ледник Сребърна" ->  "lednik Srebǎrna"
-   "Средногорски възвишения" ->  "Srednogorski vǎzvišeniya"
+   "Средногорски възвишения" ->  "Srednogorski vǎzvišenija"
    "Св. Евтимиев камък" ->  "Sv. Evtimiev kamǎk"
    "база Св. Климент Охридски" ->  "baza Sv. Kliment Ohridski"
    "връх Стъргел" ->  "vrǎh Stǎrgel"
-   "нунатак Сторгозия" ->  "nunatak Storgoziya" # nunatak Storgozia according to wiki
+   "нунатак Сторгозия" ->  "nunatak Storgozija" # nunatak Storgozia according to wiki
    "нунатак Сурвакари" ->  "nunatak Survakari"
    "ледник Световрачене" ->  "lednik Svetovračene"
    "остров Теменуга" ->  "ostrov Temenuga"
-   "Тракийски възвишения" ->  "Trakijski vǎzvišeniya"
+   "Тракийски възвишения" ->  "Trakijski vǎzvišenija"
    "хълм Цамблак" ->  "hǎlm Camblak"
    "ледник Урдовиза" ->  "lednik Urdoviza"
    "остров Вълчедръм" ->  "ostrov Vǎlčedrǎm"
    "острови Вардим" ->  "ostrovi Vardim"
    "Владигеров проток" ->  "Vladigerov protok"
-   "ледник Ябланица" ->  "lednik Yablanica"
-   "залив Ямфорина" ->  "zaliv Yamforina"
+   "ледник Ябланица" ->  "lednik Jablanica"
+   "залив Ямфорина" ->  "zaliv Jamforina"
    "Йовков нос" ->  "Jovkov nos"
    "рид Заберново" ->  "rid Zabernovo"
    "ледник Збелсурд" ->  "lednik Zbelsurd"
    "Жефарович камък" ->  "Žefarovič kamǎk"
    "връх Зиези" ->  "vrǎh Ziezi"
-   "залив Златни пясъци" ->  "zaliv Zlatni pyasǎci"
+   "залив Златни пясъци" ->  "zaliv Zlatni pjasǎci"
    "ледник Злокучене" ->  "lednik Zlokučene"
    "проток Злогош" ->  "protok Zlogoš"
 
   # UN card tests
-   "Република България" ->  "Republika Bǎlgariya"
+   "Република България" ->  "Republika Bǎlgarija"
 }
 
 dependency "posix"
@@ -152,8 +167,8 @@ stage main {
     sub "Щ" "Št"
     sub "Ъ" "Ǎ"
     sub "Ь" "J"
-    sub "Ю" "Yu"
-    sub "Я" "Ya"
+    sub "Ю" "Ju"
+    sub "Я" "Ja"
 
     sub "а" "a"
     sub "б" "b"
@@ -183,8 +198,8 @@ stage main {
     sub "щ" "št"
     sub "ъ" "ǎ"
     sub "ь" "j"
-    sub "ю" "yu"
-    sub "я" "ya"
+    sub "ю" "ju"
+    sub "я" "ja"
 
   }
 
@@ -200,22 +215,22 @@ stage main {
 
   # YU
   sub {
-    from capture(maybe(upper)) + "Yu" + capture(upper)
-    to ref(1) + "YU" + ref(2)
+    from capture(maybe(upper)) + "Ju" + capture(upper)
+    to ref(1) + "JU" + ref(2)
   }
   sub {
-    from capture(upper) + "Yu" + capture(maybe(upper))
-    to ref(1) + "YU" + ref(2)
+    from capture(upper) + "Ju" + capture(maybe(upper))
+    to ref(1) + "JU" + ref(2)
   }
 
   # YA
   sub {
-    from capture(maybe(upper)) + "Ya" + capture(upper)
-    to ref(1) + "YA" + ref(2)
+    from capture(maybe(upper)) + "Ja" + capture(upper)
+    to ref(1) + "JA" + ref(2)
   }
   sub {
-    from capture(upper) + "Ya" + capture(maybe(upper))
-    to ref(1) + "YA" + ref(2)
+    from capture(upper) + "Ja" + capture(maybe(upper))
+    to ref(1) + "JA" + ref(2)
   }
 }
 

data/maps/un-ell-Grek-Latn-1987-phonetic.isc

@@ -25,6 +25,7 @@ metadata {
     note "Followed specification in treating final μπ as b, but final ντ as nd. That distinction is dubious. (In ELOT 743:1982, both d and nd are erroneously marked as initial, and no final is given.)"
     note "τζ is not correctly transcribed as dz: fixed"
     note "not reducing geminated consonants: fixed"
+    note "The claimed source — the phonetic-transcription version of the Greek table\n (Fifth UNCSGN paper submitted by Greece and Cyprus; ELOT 743:1982 column 5, as\n in ISO 843:1997 Annex B column 5) — is not in the reference library (the WGRS\n documents index lists the paper as 'digital copy not yet available'); the rows\n follow the map's own notes and are unchanged in the WO09 UNGEGN diff."
   }
 }
 

data/maps/un-ell-Grek-Latn-1987-tl.isc

@@ -14,10 +14,18 @@ metadata {
 
   notes {
     note "Transliteration version (reversible). Also appears in annex B of ISO 843:1997. Originates in v1.0 of UNGEGN transliteration, 1987. This likely was the intended system in ELOT 743:1982 for transliteration, but the reversible character diacritics are missing from the document, and the document does not indicate what there was supposed to be."
+    note "Diffed against the UN Romanization of Greek for Geographical Names (1987)\n table, transliteration version, as published in the WGRS Report on the Current\n Status of UN Romanization Systems for Geographical Names, version 4.0 (March\n 2016) (eki.ee/wgrs/rom1_el.htm). The transliteration version differs from the\n transcription only by the sub-macron of source note 1: αυ av̱/af̱, γγ ṉg,\n γξ ṉx, γχ ṉch, ευ ev̱/ef̱, η i̱, ηυ i̱v̱/i̱f̱, ω o̱ — all encoded by the dependency\n elot-ell-Grek-Latn-743-1982-tl and verified."
+
+    note "Glyph variant, recorded not changed: the source prints the sub-macron as\n U+0320 (combining minus sign below); the map uses U+0331 (combining macron\n below), the conventional sub-macron character of the 1987 resolution."
   }
 }
 
 tests {
+   "ναυτικό" ->  "naf̱tikó"
+   "Ευρώπη" ->  "Ev̱ró̱pi̱"
+   "Αγγλία" ->  "Aṉglía"
+   "η" ->  "i̱"
+   "ω" ->  "o̱"
    "Ένα πράμα μόνον με παρακίνησε κι εμένα να γράψω ότι τούτην την πατρίδα την έχομεν όλοι μαζί, και σοφοί κι αμαθείς και πλούσιοι και φτωχοί και πολιτικοί και στρατιωτικοί και οι πλέον μικρότεροι άνθρωποι· όσοι αγωνιστήκαμεν, αναλόγως ο καθείς, έχομεν να ζήσομεν εδώ. Το λοιπόν δουλέψαμεν όλοι μαζί, να την φυλάμεν κι όλοι μαζί και να μην λέγει ούτε ο δυνατός «εγώ» ούτε ο αδύνατος. Ξέρετε πότε να λέγει ο καθείς «εγώ»; Όταν αγωνιστεί μόνος του και φκιάσει ή χαλάσει, να λέγει «εγώ»· όταν όμως αγωνίζονται πολλοί και φκιάνουν, τότε να λένε «εμείς». Είμαστε εις το «εμείς» κι όχι εις το «εγώ». Και εις το εξής να μάθομεν γνώση, αν θέλομεν να φκιάσομεν χωριόν, να ζήσομεν όλοι μαζί.\n\nΓιάννης Μακρυγιάννης.\n" ->  "Éna práma mónon me parakíni̱se ki eména na grápso̱ óti toúti̱n ti̱n patrída ti̱n échomen óloi mazí, kai sofoí ki amatheís kai ploúsioi kai fto̱choí kai politikoí kai stratio̱tikoí kai oi pléon mikróteroi ánthro̱poi; ósoi ago̱nistí̱kamen, analógo̱s o katheís, échomen na zí̱somen edó̱. To loipón doulépsamen óloi mazí, na ti̱n fylámen ki óloi mazí kai na mi̱n légei oúte o dynatós «egó̱» oúte o adýnatos. Xérete póte na légei o katheís «egó̱»? Ótan ago̱nisteí mónos tou kai fkiásei í̱ chalásei, na légei «egó̱»; ótan ómo̱s ago̱nízontai polloí kai fkiánoun, tóte na léne «emeís». Eímaste eis to «emeís» ki óchi eis to «egó̱». Kai eis to exí̱s na máthomen gnó̱si̱, an thélomen na fkiásomen cho̱rión, na zí̱somen óloi mazí.\n\nGiánni̱s Makrygiánni̱s.\n"
 
   # UN card tests

data/maps/un-ell-Grek-Latn-1987-ts.isc

@@ -14,6 +14,7 @@ metadata {
 
   notes {
     note "Transcription version"
+    note "Diffed against the UN Romanization of Greek for Geographical Names (1987)\n table, transcription version, as published in the WGRS Report on the Current\n Status of UN Romanization Systems for Geographical Names, version 4.0 (March\n 2016) (eki.ee/wgrs/rom1_el.htm; also unstats wgrr4greek): all 24 rows, the\n eight context combinations (αυ av/af, ευ ev/ef, ηυ iv/if, γγ ng, γξ nx, γχ nch,\n μπ b/mp) with footnotes A-D, and notes 1-4 verified. The context-dependent\n variants are encoded by the dependency elot-ell-Grek-Latn-743-1982-ts, whose\n table implements the ELOT 743 rules the UN system is based on (res. V/19)."
   }
 }
 

data/maps/un-guj-Gujr-Latn-1972.isc

@@ -39,7 +39,9 @@ metadata {
     note "      Exception of ી : જી jī."
     note "      Exception of ૂ : રૂ rū."
     note "      ્‌ indicates pronunciation without a vowel."
+    note "Diffed against the UN Romanization of Gujarati for Geographical Names table\n as published in the WGRS Report on the Current Status of UN Romanization Systems\n for Geographical Names, version 4.0 (February 2013) (eki.ee/wgrs/rom1_gu.htm;\n also unstats wgrr4gujarati): all independent vowels, vowel signs, other symbols\n and consonants verified against the stage rules; no gaps or divergences."
   }
+
 }
 
 tests {

data/maps/un-hin-Deva-Latn-2016.isc

@@ -36,7 +36,13 @@ metadata {
     #    we found that actually when a word ends with a consonant it should be transliterated without the vowel in it.  #
     #    so we are applying this rule as if a word ends with a consonant, in this case we are removing the ending 'a'.  #
     #####################################################################################################################
+    note "Diffed against the UN Romanization of Hindi for Geographical Names table as\n published in the WGRS Report on the Current Status of UN Romanization Systems\n for Geographical Names, version 4.0 (February 2013) (eki.ee/wgrs/rom1_hi.htm;\n also unstats wgrr4hindi): all 13 independent vowels, 11 vowel signs, 3 other\n symbols, 33 consonants and the 7 dotted variants verified; no gaps or\n divergences."
+
+    note "Source-side erratum: rows I.7 and I.8 print ॠ (U+0960) for ṛ and ॠृ for ṝ;\n the map follows the linguistic standard (ऋ ṛ, ॠ ṝ) and is kept as printed in\n the source's value column."
+
+    note "The source note recommends dropping the inherent a word-finally in names\n (Kānapur, not Kānapura; जगत् Jagaṯ with a sub-macron on the final consonant);\n the map implements the a-drop, while the exceptional sub-macron of Jagaṯ is\n lexical and not encoded."
   }
+
 }
 
 tests {

data/maps/un-kan-Kana-Latn-2016.isc

@@ -28,7 +28,9 @@ metadata {
   notes {
     note "      While most consonants have a diacritic in the upper right corner of the character (like ಕ),\n      some do not, e.g. ಖ kha, ಜ ja, ನ na, ಬ ba, ಲ la."
     note "      Combinations with r as the first component are written by adding a special symbol after the\n      second consonant: ರ್ಗ rga."
+    note "Diffed against the UN Romanization of Kannada for Geographical Names table\n as published in the WGRS Report on the Current Status of UN Romanization Systems\n for Geographical Names, version 4.0 (February 2013) (eki.ee/wgrs/rom1_kn.pdf;\n also unstats wgrr4kannada): all independent vowels, vowel signs, other symbols\n and consonants verified. The source's ligature rows are compositional (first\n consonant bare + next consonant with inherent a); the map's rules reproduce\n them. No gaps or divergences."
   }
+
 }
 
 tests {

data/maps/un-mal-Mlym-Latn-1972.isc

@@ -33,10 +33,16 @@ metadata {
     note "      ടൗ In the older spelling: ൌ."
     note "      ട്‌  Indicates absence of the inherent short [a] vowel, e.g. ക്‌ k , സ്‌ s. Special, co-called chillu-forms of consonants with that symbol: ണ്‍ ṇ, ന്‍ n, ര്‍ r, ല്‍ l, ള്‍ ḷ."
     note "      റ  Special ligature: റ്റ ṭ̈."
+    note "Diffed against the UN Romanization of Malayalam for Geographical Names table\n as published in the WGRS Report on the Current Status of UN Romanization Systems\n for Geographical Names, version 4.0 (February 2013) (eki.ee/wgrs/rom1_ml.htm;\n also unstats wgrr4malayalam): all independent vowels, vowel signs, other symbols\n and consonants verified."
+
+    note "Fixed against the source: the au vowel sign ൗ (U+0D57) mapped to the empty\n string, so ടൗ rendered ṭa instead of ṭau; it now renders au per table II.13."
   }
+
 }
 
 tests {
+   "ടൗ" ->  "ṭau"
+   "ഔ" ->  "au"
    "ചൈനയ്ക്കെതിരെ ലഡാക്കിൽ സദാസജ്ജം; യുഎസിൽനിന്ന് ൭൨,൫൦൦ സിഗ്–൧൬ റൈഫിൾ" ->  "chainaykkĕtirĕ laḍākkil sadāsajjaṃ; yuĕsilninn 72,500 sig–16 ṟaiphiḷ"
    "സർഗഭൂമിക’യ്ക്കില്ല; ലളിതച്ചേച്ചി അങ്ങനെ പറഞ്ഞിട്ടുണ്ടാവില്ല: ആർഎൽവി രാമകൃഷ്ണൻ" ->  "sargabhūmika’ykkilla; laḷitachchechchi aṅṅanĕ paṟaññiṭṭuṇṭāvilla: ārĕlvi rāmakṛṣhṇan"
    "സ്വർണക്കടത്ത്‌: ഫൈസൽ ഫരീദും റബിന്‍സും ദുബായിൽ അറസ്റ്റിലായെന്ന്‌ എന്‍ഐഎ" ->  "svarṇakkaṭatt: phaisal pharīduṃ ṟabinsuṃ dubāyil aṟasṟṟilāyĕnn ĕnaiĕ"
@@ -55,7 +61,7 @@ tests {
 }
 
 aliases {
-  mlym_chars_1 = any("\u0d3e\u0d3f\u0d40\u0d41\u0d42\u0d43\u0d46\u0d47\u0d48\u0d4a\u0d4b\u0d4c\u0d4d")
+  mlym_chars_1 = any("\u0d3e\u0d3f\u0d40\u0d41\u0d42\u0d43\u0d46\u0d47\u0d48\u0d4a\u0d4b\u0d4c\u0d4d\u0d57")
 }
 
 stage main {
@@ -327,7 +333,7 @@ stage main {
 
     sub "्" ""
     sub "़" ""
-    sub "ൗ" ""
+    sub "ൗ" "au"
     sub "‍" ""
     sub "‌" ""
 

data/maps/un-mar-Deva-Latn-2016.isc

@@ -32,10 +32,18 @@ metadata {
     #    we found that actually when a word ends with a consonant it should be transliterated without the vowel in it.  #
     #    so we are applying this rule as if a word ends with a consonant, in this case we are removing the ending 'a'.  #
     #####################################################################################################################
+    note "Diffed against the UN Romanization of Marathi for Geographical Names table as\n published in the WGRS Report on the Current Status of UN Romanization Systems\n for Geographical Names, version 4.0 (March 2016) (eki.ee/wgrs/rom1_mr.htm;\n also unstats wgrr4marathi), which gives the Hindi table plus the Marathi\n additions: ळ ḷa, ऍ ă (abbreviated कॅ kă), the ऱ्य/ऱ्ह and -h ligatures, and the\n independent अि i … series."
+
+    note "Fixed against the source: the abbreviated ऍ form कॅ (U+0945) was not encoded\n and is now rendered ă; the corrupt 'note[3]' rule (from some + 0939) was\n removed. The ॲ ê and ऑ ô rows have no row in the source table and are kept as\n practical extensions; the ऱ row likewise extends the source's ligature-only\n use of ऱ."
   }
+
 }
 
 tests {
+   "कॅ" ->  "kă"
+   "ळ" ->  "ḷa"
+   "ऍ" ->  "ă"
+   "ऱ्या" ->  "ryā"
    "महाराष्ट्र" ->  "mahārāṣhṭr"
    "मुंबई" ->  "muṁbaī"
    "परिपक्क" ->  "paripakk"
@@ -65,15 +73,6 @@ stage main {
 
   run  map.devalatn.stage.main
 
-  # RULES
-  # note[3]
-  sub {
-    from some("\U") + "0939"
-    to "h"
-    before ""
-    after boundary
-  }
-
 
   # CHARACTERS
   parallel {
@@ -81,6 +80,8 @@ stage main {
     # Note 1
     sub "ळ" "ḷa"
     sub "ऍ" "ă"
+    # abbreviated form of ऍ (कॅ kă): consonant + U+0945
+    sub "\u0945" "ă"
 
     # Note 2
     sub "ऱ" "r"

data/maps/un-mkd-Cyrl-Latn-1977.isc

@@ -13,6 +13,12 @@ metadata {
     The system is used in The former Yugoslav Republic of Macedonia (see e.g. Report of The former Yugoslav Republic of Macedonia. Seventh United Nations Conference on the Standardization of Geographical Names. New York, 13-22 January 1998. Document E/CONF.91/CRP.23) and in international cartographic products.
     The Cyrillic alphabet is used. The romanization table is unambiguous and can be applied automatically. The system is on the whole reversible with the exception of the characters ѓ and ќ. These are romanized as g and k, respectively, before е and и. But also г and к are occasionally used before е and и, so, for example, the name Makedonija is reconstituted as Македонија, not Маќедонија in the Cyrillic source script.
   }
+
+  notes {
+    note "Diffed against the UN Romanization of Macedonian Cyrillic for Geographical\n Names (1977) table as published in the WGRS Report on the Current Status of UN\n Romanization Systems for Geographical Names, version 4.0 (February 2013)\n (eki.ee/wgrs/rom1_mk.htm; also unstats wgrr4macedonian): all 32 rows verified.\n The map's claimed URL (the older v2_2 site revision) presents the same 1977\n (III/11) table; no newer edition exists."
+
+    note "Row 32, the apostrophe letter ’, romanized ’, is now encoded explicitly;\n it previously passed through unromanized."
+  }
 }
 
 tests {
@@ -22,6 +28,11 @@ tests {
    "ЛУЃЕ луѓе" ->  "LUGE luge"
    "ЅВЕЗДА ѕвезда Ѕвезда" ->  "DZVEZDA dzvezda Dzvezda"
    "ЌАРУВАЊЕ ќарување" ->  "ĆARUVANJE ćaruvanje"
+   "Ѓорче Петров" ->  "Đorče Petrov"
+   "Ѓе" ->  "Ge"
+   "Ќесендре" ->  "Kesendre"
+   "Ќулаш" ->  "Ćulaš"
+   "п’етка" ->  "p’etka"
 
   # UN card tests
    "Скопје" ->  "Skopje"

data/maps/un-mon-Mong-Latn-general-2013.isc

@@ -33,6 +33,9 @@ metadata {
     note "Where two Roman equivalents are given, the second (in brackets) is\n used for recording the pronunciation of place-names while the first\n form is for general use."
     note "In the table only word-initial character variants are shown.\n Depending on the position in the word many variants of the characters\n are used as well as some ligatures. These features are not covered here."
     note "For technical reasons the characters of the Mongolian script are\n turned 90˚ anti-clockwise."
+    note "Diffed against the UN Romanization of Mongolian (in China) for Geographical\n Names table as published in the WGRS Report on the Current Status of UN\n Romanization Systems for Geographical Names, version 4.0 (September 2013)\n (eki.ee/wgrs/rom1_mnc.htm): all 27 rows verified against the dependency\n sasm-mon-Mong-Latn-general-1978; the general-use forms o and u of rows 26/27\n are the ones encoded here."
+
+    note "Source-side ambiguities, recorded not changed: the printed table lists ᠲ in\n both row 4 (d) and row 20 (t), and ᠺ in both row 7 (g) and row 11 (k); the map\n resolves ᠲ to t and ᠺ to k. The rows for ᠧ e, ᠩ ng, the Mongolian vowel separator\n and the space have no row in the printed table and are kept as practical\n extensions (the source note records that only word-initial variants are shown)."
   }
 
   implementation_notes {

data/maps/un-mon-Mong-Latn-phonetic-2013.isc

@@ -33,6 +33,9 @@ metadata {
     note "Where two Roman equivalents are given, the second (in brackets) is\n used for recording the pronunciation of place-names while the first\n form is for general use."
     note "In the table only word-initial character variants are shown.\n Depending on the position in the word many variants of the characters\n are used as well as some ligatures. These features are not covered here."
     note "For technical reasons the characters of the Mongolian script are\n turned 90˚ anti-clockwise."
+    note "Diffed against the UN Romanization of Mongolian (in China) for Geographical\n Names table as published in the WGRS Report on the Current Status of UN\n Romanization Systems for Geographical Names, version 4.0 (September 2013)\n (eki.ee/wgrs/rom1_mnc.htm): all 27 rows verified; rows 26/27 use the bracketed\n place-name-recording forms ô and û of source note 2, encoded by the dependency\n sasm-mon-Mong-Latn-phonetic-1978."
+
+    note "Source-side ambiguities, recorded not changed: the printed table lists ᠲ in\n both row 4 (d) and row 20 (t), and ᠺ in both row 7 (g) and row 11 (k); the map\n resolves ᠲ to t and ᠺ to k. The rows for ᠧ e, ᠩ ng, the Mongolian vowel separator\n and the space have no row in the printed table and are kept as practical\n extensions."
   }
 
   implementation_notes {

data/maps/un-nep-Deva-Latn-1972.isc

@@ -34,7 +34,9 @@ metadata {
     #    we found that actually when a word ends with a consonant it should be transliterated without the vowel in it.  #
     #    so we are applying this rule as if a word ends with a consonant, in this case we are removing the ending 'a'.  #
     #####################################################################################################################
+    note "Diffed against the UN Romanization System of Nepali for Geographical Names\n (1972) table as published in the WGRS Report on the Current Status of UN\n Romanization Systems for Geographical Names, version 4.0 (February 2013)\n (eki.ee/wgrs/rom1_ne.htm; also unstats wgrr4nepali): all rows verified; no gaps\n or divergences. The source's rows I.7/I.8 print ॠ for ṛ and ॠृ for ṝ (source-side\n glyph erratum); the map follows the linguistic standard."
   }
+
 }
 
 tests {

data/maps/un-nep-Deva-Latn-2013.isc

@@ -28,7 +28,9 @@ metadata {
 
     note "In normal spelling ṙ and ṙh are not marked, instead dotless ड ḍa, ढ ḍha are used to denote these sounds."
     note "      Ligatures may be formed with r as the first component: ~ह rha."
+    note "Diffed against the UN Romanization System of Nepali for Geographical Names\n (1972) table (WGRS Report, version 4.0, February 2013; eki.ee/wgrs/rom1_ne.htm;\n also unstats wgrr4nepali). Reissue note: no 2013 Nepali system exists — the UN\n system was approved only in 1972 (II/11); '2013' is the WGRS report revision\n year carried in the map's metadata. This map therefore delegates entirely to\n un-nep-Deva-Latn-1972 and encodes no separate edition."
   }
+
 }
 
 tests {

data/maps/un-ori-Orya-Latn-1972.isc

@@ -26,10 +26,18 @@ metadata {
 
   notes {
     note "Combinations with r as the first component are written with a special superscript symbol, e.g. ର୍କ rka."
+    note "Diffed against the UN Romanization of Oriya for Geographical Names table as\n published in the WGRS Report on the Current Status of UN Romanization Systems\n for Geographical Names, version 4.0 (February 2013) (eki.ee/wgrs/rom1_or.htm;\n also unstats wgrr4oriya): all independent vowels, vowel signs, other symbols\n and consonants verified."
+
+    note "Fixed against the source: the anusvara ଂ rendered m with a dot below instead\n of the printed ṁ (dot above); the candrabindu ଁ dropped its combining mark (m →\n m̐); and the alternate ai/au signs ୖ (U+0B56) and ୗ (U+0B57), printed in the\n source as (କୖ)/(କୗ), were unencoded and passed through raw. All four now follow\n table rows II.9/II.11 and III.1/III.2."
   }
+
 }
 
 tests {
+   "କଂ" ->  "kaṁ"
+   "କଁ" ->  "kam̐"
+   "କୖ" ->  "kai"
+   "କୗ" ->  "kau"
    "ର୍କ" ->  "rka"
    "ଓଡ଼ିଆ" ->  "oṙiā"
    "ଓଡ଼ିଶା" ->  "oṙishā"
@@ -37,10 +45,10 @@ tests {
    "ଆଇପିଏଲ୍‌-୧୩: ଦିଲ୍ଲୀ କ୍ୟାପିଟାଲ୍ସକୁ ୮୮ ରନ୍‌ ପରାସ୍ତ କଲା ସନରାଇଜର୍ସ ହାଇଦ୍ରାବାଦ" ->  "āipiel-13: dillī kyāpiṭālsaku 88 ran parāsta kalā sanarāijarsa hāidrābāda"
    "ପ୍ରେମ ସମ୍ପର୍କରେ ଭଟ୍ଟା: ରାଗରେ ପ୍ରେମିକାର ତଣ୍ଟି କାଟି ନିଜେ ବିଷ ପିଇଲା ପ୍ରେମିକ" ->  "prema samparkare bhaṭṭā: rāgare premikāra taṇṭi kāṭi nije biṣha piilā premika"
    "ପ୍ରେମ ସମ୍ପର୍କରେ ଭଟ୍ଟା: ରାଗରେ ପ୍ରେମିକାର ତଣ୍ଟି କାଟି ନିଜେ ବିଷ ପିଇଲା ପ୍ରେମିକ" ->  "prema samparkare bhaṭṭā: rāgare premikāra taṇṭi kāṭi nije biṣha piilā premika"
-   "ହୋଟେଲ, ଲଜ୍‌ରେ ରୁମ୍‌ ମିଳୁନି: ନେତା‌ଙ୍କ ନାଁରେ ଆଗୁଆ ହୋଇଯାଇଛି ବୁକିଂ" ->  "heāṭela, lajre rum miḷuni: netāṅka nāmre āguā heāiỵāichhhi bukiṃ"
+   "ହୋଟେଲ, ଲଜ୍‌ରେ ରୁମ୍‌ ମିଳୁନି: ନେତା‌ଙ୍କ ନାଁରେ ଆଗୁଆ ହୋଇଯାଇଛି ବୁକିଂ" ->  "heāṭela, lajre rum miḷuni: netāṅka nām̐re āguā heāiỵāichhhi bukiṁ"
    "ପର୍ଯ୍ୟଟକମାନଙ୍କ ନିମନ୍ତେ ନଭେମ୍ବର ୧ରୁ ଖୋଲିବ ଶିମିଳିପାଳ ଅଭୟାରଣ୍ୟ" ->  "parỵyaṭakamānaṅka nimante nabhembara 1ru kholiba shimiḷipāḷa abhayāraṇya"
-   "ପାରିବାରିକ ଅଶାନ୍ତିର କରୁଣ ପରିଣତି: କୂଅକୁ ଡେଇଁଲେ ମା’-ଝିଅ, ଝିଅ ମୃତ" ->  "pāribārika ashāntira karuṇa pariṇati: kūaku ḍeimle mā’-jhia, jhia mṛta"
-   "‘ଭ୍ରଷ୍ଟାଚାରର ବଂଶବାଦ’ ଏବେ ସାଜିଛି ଦେଶ ପାଇଁ ନୂଆ ସମସ୍ୟା; ପ୍ରଧାନମନ୍ତ୍ରୀ ମୋଦୀ" ->  "‘bhraṣhṭāchārara baṃshabāda’ ebe sājichhhi desha pāim nūā samasyā; pradhānamantrī modī"
+   "ପାରିବାରିକ ଅଶାନ୍ତିର କରୁଣ ପରିଣତି: କୂଅକୁ ଡେଇଁଲେ ମା’-ଝିଅ, ଝିଅ ମୃତ" ->  "pāribārika ashāntira karuṇa pariṇati: kūaku ḍeim̐le mā’-jhia, jhia mṛta"
+   "‘ଭ୍ରଷ୍ଟାଚାରର ବଂଶବାଦ’ ଏବେ ସାଜିଛି ଦେଶ ପାଇଁ ନୂଆ ସମସ୍ୟା; ପ୍ରଧାନମନ୍ତ୍ରୀ ମୋଦୀ" ->  "‘bhraṣhṭāchārara baṁshabāda’ ebe sājichhhi desha pāim̐ nūā samasyā; pradhānamantrī modī"
    "ପାହାଡ଼ି ଇଲାକାବାସୀଙ୍କ ଆଶାର ବତୀ ‘ପାର୍ବତୀ’" ->  "pāhāṙi ilākābāsīṅka āshāra batī ‘pārbatī’"
 
   # Note: UN card expects this to start with an uppercase letter, but that would break other tests
@@ -49,7 +57,7 @@ tests {
 }
 
 aliases {
-  orya_chars_1 = any("\u0b4d\u0b3e\u0b3f\u0b40\u0b41\u0b42\u0b43\u0b47\u0b48\u0b4b\u0b4c")
+  orya_chars_1 = any("\u0b4d\u0b3e\u0b3f\u0b40\u0b41\u0b42\u0b43\u0b47\u0b48\u0b4b\u0b4c\u0b56\u0b57")
 }
 
 stage main {
@@ -316,13 +324,13 @@ stage main {
     sub "କ୍ଷ" "kṣha"
 
     # Chandrabindu
-    sub "ଁ" "m"
+    sub "ଁ" "m̐"
 
     # Bisarga
     sub "ଃ" "ḥ"
 
     # Anusvāra
-    sub "ଂ" "ṃ"
+    sub "ଂ" "ṁ"
 
     # Medials # Needed for connecting constants
 
@@ -334,8 +342,10 @@ stage main {
     sub "ୃ" "ṛ"
     sub "େ" "e"
     sub "ୈ" "ai"
+    sub "ୖ" "ai"
     sub "ୋ" "o"
     sub "ୌ" "au"
+    sub "ୗ" "au"
 
     sub "्" ""
     sub "୍" ""

data/maps/un-pan-Guru-Latn-1972.isc

@@ -39,7 +39,9 @@ metadata {
     note "      (ਂ) Used in all other occasions, e.g. ਝੋਂਕਾ jhoṁkā."
     note "      (ੱ) Marks doubling of the following consonant: ਨਿੱਕਾ nikkā, ਲੱਭਣਾ labhbhṇā, ਕੁੱਤਾ kuttā, ਹਿੱਸਾ hissā, ਲੱਮਾ lammā."
     note "      Absence of the inherent vowel (-a) is not marked in the spelling in any way except for the combinations\n      with subscript characters and those which are doubled by ੱ (adhaka)."
+    note "Diffed against the UN Romanization of Panjabi for Geographical Names table\n as published in the WGRS Report on the Current Status of UN Romanization Systems\n for Geographical Names, version 4.0 (February 2013) (eki.ee/wgrs/rom1_pa.htm;\n also unstats wgrr4punjabi): all independent vowels, vowel signs, other symbols,\n consonants and adscript forms verified; no gaps or divergences."
   }
+
 }
 
 tests {

data/maps/un-prs-Arab-Latn-1967.isc

@@ -51,6 +51,7 @@ metadata {
 }
 
 tests {
+   "دَهَانِهٴ مَمبَر" ->  "Dahāneh-ye Mambar"
    "اَنجِيرة" ->  "Anjīrah"
    "اِيْوَانِي" ->  "Eyvānī"
    "آبَادَان" ->  "Ābādān"
@@ -110,15 +111,10 @@ stage main {
     sub {
     from "\u2013" + any("\u0649\u064a") + "\u0647"
     to "-īyeh"
-  }
-    sub {
-    from any("\u0654\u0674")
-    to "-e"
   }
     sub {
     from any("\u0654\u0674")
     to "-ye"
-    before any("\u064a\u0647")
   }
     sub {
     from "\u0650" + boundary
@@ -346,7 +342,7 @@ stage main {
 
   # POSTRULES
   sub {
-    from any("\u0061".."\uFFFF")
+    from any(["a", "b", "c", "d", "e", "f", "g", "h", "i", "j", "k", "l", "m", "n", "o", "p", "q", "r", "s", "t", "u", "v", "w", "x", "y", "z", "ā", "á", "é", "ī", "ū", "ḩ", "ş", "ţ", "z̄", "z̧", "ẕ"])
     to upcase
     before boundary
     not_before boundary + any("‘’'-")

data/maps/un-rus-Cyrl-Latn-1987.isc

@@ -25,6 +25,8 @@ metadata {
   }
 
   notes {
+    note "Diffed against the UN Romanization of Russian for Geographical Names\n (1987) table as published in the WGRS Report on the Current Status of UN\n Romanization Systems for Geographical Names, version 4.0 (March 2016)\n (eki.ee/wgrs/rom1_ru.htm; also unstats wgrr4russian): all 33 rows verified.\n Hard sign Ъ now renders \" (U+0022) per row 28 as printed; the map previously\n rendered ” (U+201D)."
+
     note "Cursive forms of some characters might be formed differently: Аа Бб\n Вв Гг Дд Ее Ёё Жж Зз Ии Йй Кк Лл Мм Нн Оо Пп Рр Сс Тт Уу Фф Хх Цц Чч Шш\n Щщ Ъъ Ыы Ьь Ээ Юю Яя."
 
     note "Fifth United Nations Conference on the Standardization of\n Geographical Names. Montreal, 18–31 August 1987. Vol. I. Report of the\n Conference, pp. 40–41."
@@ -60,7 +62,9 @@ tests {
    "Чeчeльник" ->  "Čečel’nik"
    "Шишкинo" ->  "Šiškino"
    "Щукинo" ->  "Ščukino"
-   "Пoдъячeвo" ->  "Pod”jačevo"
+   "Ъ" ->  "\""
+   "ъ" ->  "\""
+   "Пoдъячeвo" ->  "Pod\"jačevo"
    "Ыныкчaнский" ->  "Ynykčanskij"
    "Пaрaньгa" ->  "Paran’ga"
    "Кaзaнь" ->  "Kazan’"
@@ -70,7 +74,7 @@ tests {
    "Юхнoв" ->  "Juhnov"
    "Юрюзaнь" ->  "Jurjuzan’"
    "Ямaл" ->  "Jamal"
-   "Язъявaн" ->  "Jaz”javan"
+   "Язъявaн" ->  "Jaz\"javan"
    "Яя" ->  "Jaja"
    "Вязьмa" ->  "Vjaz’ma"
 
@@ -85,9 +89,9 @@ dependency "gost-rus-Cyrl-Latn-16876-71-1983" as cyrllatn
 stage main {
   # CHARACTERS
   parallel {
-    sub "\u042A" "”"
+    sub "\u042A" "\""
     sub "\u042C" "’"
-    sub "\u044A" "”"
+    sub "\u044A" "\""
     sub "\u044C" "’"
     sub "\u042E" "Ju"
     sub "\u044E" "ju"

data/maps/un-sin-Sinh-Latn-1972.isc

@@ -28,7 +28,11 @@ metadata {
     note "      ූ Exceptions: කූ kū, ගූ gū, තූ tū, භූ bhū, රූ rū, ශූ shū"
     note "      ් Marks absence of the inherent vowel. Special forms: ඛ් kh, ඞ් ṅ, ච් ch, ට් ṭ, ඩ් ḍ, ධ් dh, බ් b, ම් m, ව් v."
     note "      ඛ ඝ ඞ ච ඡ ඣ ඤ ඨ ඪ ථ ධ ඵ භ ශ ෂ These consonants are used only in Sanskrit and Pali loanwords."
+    note "The claimed source (Second UNCSGN, London 1972, vol. II, pp. 151-152 — the\n D. N. Sharma table) is not in the reference library; the rows follow the notes\n above and are unchanged in the WO09 UNGEGN diff. Reissue note: the library's\n current Sinhala page (WGRS 5.0, October 2021; eki.ee/wgrs/rom2_si.htm; also\n unstats wgrr5sinhala) states that no UN romanization system for Sinhala has\n been approved and documents the Sri Lanka Survey Department system approved\n 4 September 2018 (ISO 15919-based, amended); that 2018/2021 system is NOT\n encoded here."
+
+    note "Metadata erratum: the url field points at rom1_ml.htm (Malayalam); the\n Sinhala table is rom2_si. Left unchanged to keep the map's declared edition\n intact, and recorded here."
   }
+
 }
 
 tests {

data/maps/un-srp-Cyrl-Latn-1997.isc

@@ -6,13 +6,21 @@ metadata {
   language iso-639-2:srp
   source_script Cyrl
   destination_script Latn
-  name Romanization of Serbo-Crotian (1972)
+  name Romanization of Serbian (1977)
   url http://www.eki.ee/wgrs/rom1_sr.htm
   creation_date 1997
   description {    The United Nations recommended system was approved in 1977 (III/11), based on the Roman equivalent of the Serbo-Croatian Cyrillic alphabet that had long been employed in official gazetteers and maps of Yugoslavia. The table was published as an annex to the resolution1. The term relating to the language (Serbian instead of Serbo-Croatian) was modified in 2002 (resolution VIII/14).
     The system is used in Serbia, Montenegro, and Bosnia and Herzegovina. It is also used in international cartographic products.
     Serbian uses the Cyrillic script that is alphabetic. The romanization table is unambiguous and can be applied automatically. The system is reversible.
   }
+
+  notes {
+    note "Diffed against the UN Romanization of Serbian for Geographical Names (1977)\n table as published in the WGRS Report on the Current Status of UN Romanization\n Systems for Geographical Names, version 4.0 (March 2016)\n (eki.ee/wgrs/rom1_sr.htm; also unstats wgrr4serbian): all 30 rows verified\n against the system's stage rules; no gaps, no divergences."
+
+    note "The system was approved in 1977 (res. III/11); the name previously read\n 'Serbo-Crotian (1972)' — the language term was changed from Serbo-Croatian to\n Serbian by res. VIII/14 (2002). The id/filename year 1997 predates this diff\n and is retained."
+
+    note "The base rows are encoded through dependency alalc-srp-Cyrl-Latn-2013, whose\n parallel table is row-for-row identical to the UN 1977 Serbian table."
+  }
 }
 
 tests {