harfbuzz/src/gen-tag-table.py

1217 lines
36 KiB
Python
Raw Normal View History

#!/usr/bin/env python3
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
"""Generator of the mapping from OpenType tags to BCP 47 tags and vice
versa.
It creates a ``const LangTag[]``, matching the tags from the OpenType
languages system tag list to the language subtags of the BCP 47 language
subtag registry, with some manual adjustments. The mappings are
supplemented with macrolanguages' sublanguages and retired codes'
replacements, according to BCP 47 and some manual additions where BCP 47
omits a retired code entirely.
Also generated is a function, ``hb_ot_ambiguous_tag_to_language``,
intended for use by ``hb_ot_tag_to_language``. It maps OpenType tags
back to BCP 47 tags. Ambiguous OpenType tags (those that correspond to
multiple BCP 47 tags) are listed here, except when the alphabetically
first BCP 47 tag happens to be the chosen disambiguated tag. In that
case, the fallback behavior will choose the right tag anyway.
2020-05-28 12:31:15 +02:00
usage: ./gen-tag-table.py languagetags language-subtag-registry
Input files:
* https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags
* https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
"""
import collections
2021-03-28 16:09:13 +02:00
import html
2020-02-23 21:28:39 +01:00
from html.parser import HTMLParser
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
import itertools
import re
import sys
import unicodedata
if len (sys.argv) != 3:
2020-05-28 20:21:29 +02:00
sys.exit (__doc__)
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
def expect (condition, message=None):
if not condition:
if message is None:
raise AssertionError
raise AssertionError (message)
2021-03-28 16:09:13 +02:00
def write (s):
sys.stdout.flush ()
sys.stdout.buffer.write (s.encode ('utf-8'))
DEFAULT_LANGUAGE_SYSTEM = ''
# from https://www-01.sil.org/iso639-3/iso-639-3.tab
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
ISO_639_3_TO_1 = {
'aar': 'aa',
'abk': 'ab',
'afr': 'af',
'aka': 'ak',
'amh': 'am',
'ara': 'ar',
'arg': 'an',
'asm': 'as',
'ava': 'av',
'ave': 'ae',
'aym': 'ay',
'aze': 'az',
'bak': 'ba',
'bam': 'bm',
'bel': 'be',
'ben': 'bn',
'bis': 'bi',
'bod': 'bo',
'bos': 'bs',
'bre': 'br',
'bul': 'bg',
'cat': 'ca',
'ces': 'cs',
'cha': 'ch',
'che': 'ce',
'chu': 'cu',
'chv': 'cv',
'cor': 'kw',
'cos': 'co',
'cre': 'cr',
'cym': 'cy',
'dan': 'da',
'deu': 'de',
'div': 'dv',
'dzo': 'dz',
'ell': 'el',
'eng': 'en',
'epo': 'eo',
'est': 'et',
'eus': 'eu',
'ewe': 'ee',
'fao': 'fo',
'fas': 'fa',
'fij': 'fj',
'fin': 'fi',
'fra': 'fr',
'fry': 'fy',
'ful': 'ff',
'gla': 'gd',
'gle': 'ga',
'glg': 'gl',
'glv': 'gv',
'grn': 'gn',
'guj': 'gu',
'hat': 'ht',
'hau': 'ha',
'hbs': 'sh',
'heb': 'he',
'her': 'hz',
'hin': 'hi',
'hmo': 'ho',
'hrv': 'hr',
'hun': 'hu',
'hye': 'hy',
'ibo': 'ig',
'ido': 'io',
'iii': 'ii',
'iku': 'iu',
'ile': 'ie',
'ina': 'ia',
'ind': 'id',
'ipk': 'ik',
'isl': 'is',
'ita': 'it',
'jav': 'jv',
'jpn': 'ja',
'kal': 'kl',
'kan': 'kn',
'kas': 'ks',
'kat': 'ka',
'kau': 'kr',
'kaz': 'kk',
'khm': 'km',
'kik': 'ki',
'kin': 'rw',
'kir': 'ky',
'kom': 'kv',
'kon': 'kg',
'kor': 'ko',
'kua': 'kj',
'kur': 'ku',
'lao': 'lo',
'lat': 'la',
'lav': 'lv',
'lim': 'li',
'lin': 'ln',
'lit': 'lt',
'ltz': 'lb',
'lub': 'lu',
'lug': 'lg',
'mah': 'mh',
'mal': 'ml',
'mar': 'mr',
'mkd': 'mk',
'mlg': 'mg',
'mlt': 'mt',
'mol': 'mo',
'mon': 'mn',
'mri': 'mi',
'msa': 'ms',
'mya': 'my',
'nau': 'na',
'nav': 'nv',
'nbl': 'nr',
'nde': 'nd',
'ndo': 'ng',
'nep': 'ne',
'nld': 'nl',
'nno': 'nn',
'nob': 'nb',
'nor': 'no',
'nya': 'ny',
'oci': 'oc',
'oji': 'oj',
'ori': 'or',
'orm': 'om',
'oss': 'os',
'pan': 'pa',
'pli': 'pi',
'pol': 'pl',
'por': 'pt',
'pus': 'ps',
'que': 'qu',
'roh': 'rm',
'ron': 'ro',
'run': 'rn',
'rus': 'ru',
'sag': 'sg',
'san': 'sa',
'sin': 'si',
'slk': 'sk',
'slv': 'sl',
'sme': 'se',
'smo': 'sm',
'sna': 'sn',
'snd': 'sd',
'som': 'so',
'sot': 'st',
'spa': 'es',
'sqi': 'sq',
'srd': 'sc',
'srp': 'sr',
'ssw': 'ss',
'sun': 'su',
'swa': 'sw',
'swe': 'sv',
'tah': 'ty',
'tam': 'ta',
'tat': 'tt',
'tel': 'te',
'tgk': 'tg',
'tgl': 'tl',
'tha': 'th',
'tir': 'ti',
'ton': 'to',
'tsn': 'tn',
'tso': 'ts',
'tuk': 'tk',
'tur': 'tr',
'twi': 'tw',
'uig': 'ug',
'ukr': 'uk',
'urd': 'ur',
'uzb': 'uz',
'ven': 've',
'vie': 'vi',
'vol': 'vo',
'wln': 'wa',
'wol': 'wo',
'xho': 'xh',
'yid': 'yi',
'yor': 'yo',
'zha': 'za',
'zho': 'zh',
'zul': 'zu',
}
class LanguageTag (object):
"""A BCP 47 language tag.
Attributes:
subtags (List[str]): The list of subtags in this tag.
grandfathered (bool): Whether this tag is grandfathered. If
``true``, the entire lowercased tag is the ``language``
and the other subtag fields are empty.
language (str): The language subtag.
script (str): The script subtag.
region (str): The region subtag.
variant (str): The variant subtag.
Args:
tag (str): A BCP 47 language tag.
"""
def __init__ (self, tag):
global bcp_47
self.subtags = tag.lower ().split ('-')
self.grandfathered = tag.lower () in bcp_47.grandfathered
if self.grandfathered:
self.language = tag.lower ()
self.script = ''
self.region = ''
self.variant = ''
else:
self.language = self.subtags[0]
self.script = self._find_first (lambda s: len (s) == 4 and s[0] > '9', self.subtags)
self.region = self._find_first (lambda s: len (s) == 2 and s[0] > '9' or len (s) == 3 and s[0] <= '9', self.subtags[1:])
self.variant = self._find_first (lambda s: len (s) > 4 or len (s) == 4 and s[0] <= '9', self.subtags)
def __str__(self):
return '-'.join(self.subtags)
def __repr__ (self):
return 'LanguageTag(%r)' % str(self)
@staticmethod
def _find_first (function, sequence):
try:
return next (iter (filter (function, sequence)))
except StopIteration:
return None
def is_complex (self):
"""Return whether this tag is too complex to represent as a
``LangTag`` in the generated code.
Complex tags need to be handled in
``hb_ot_tags_from_complex_language``.
Returns:
Whether this tag is complex.
"""
return not (len (self.subtags) == 1
or self.grandfathered
and len (self.subtags[1]) != 3
and ot.from_bcp_47[self.subtags[0]] == ot.from_bcp_47[self.language])
def get_group (self):
"""Return the group into which this tag should be categorized in
``hb_ot_tags_from_complex_language``.
The group is the first letter of the tag, or ``'und'`` if this tag
should not be matched in a ``switch`` statement in the generated
code.
Returns:
This tag's group.
"""
return ('und'
if (self.language == 'und'
or self.variant in bcp_47.prefixes and len (bcp_47.prefixes[self.variant]) == 1)
else self.language[0])
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
class OpenTypeRegistryParser (HTMLParser):
"""A parser for the OpenType language system tag registry.
Attributes:
header (str): The "last updated" line of the registry.
names (Mapping[str, str]): A map of language system tags to the
names they are given in the registry.
ranks (DefaultDict[str, int]): A map of language system tags to
numbers. If a single BCP 47 tag corresponds to multiple
OpenType tags, the tags are ordered in increasing order by
rank. The rank is based on the number of BCP 47 tags
associated with a tag, though it may be manually modified.
to_bcp_47 (DefaultDict[str, AbstractSet[str]]): A map of
OpenType language system tags to sets of BCP 47 tags.
from_bcp_47 (DefaultDict[str, AbstractSet[str]]): ``to_bcp_47``
inverted. Its values start as unsorted sets;
``sort_languages`` converts them to sorted lists.
from_bcp_47_uninherited (Optional[Dict[str, AbstractSet[str]]]):
A copy of ``from_bcp_47``. It starts as ``None`` and is
populated at the beginning of the first call to
``inherit_from_macrolanguages``.
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
"""
def __init__ (self):
HTMLParser.__init__ (self)
self.header = ''
self.names = {}
self.ranks = collections.defaultdict (int)
self.to_bcp_47 = collections.defaultdict (set)
self.from_bcp_47 = collections.defaultdict (set)
self.from_bcp_47_uninherited = None
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
# Whether the parser is in a <td> element
self._td = False
# Whether the parser is after a <br> element within the current <tr> element
self._br = False
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
# The text of the <td> elements of the current <tr> element.
self._current_tr = []
def handle_starttag (self, tag, attrs):
if tag == 'br':
self._br = True
elif tag == 'meta':
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
for attr, value in attrs:
if attr == 'name' and value == 'updated_at':
self.header = self.get_starttag_text ()
break
elif tag == 'td':
self._td = True
self._current_tr.append ('')
elif tag == 'tr':
self._br = False
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
self._current_tr = []
def handle_endtag (self, tag):
if tag == 'td':
self._td = False
elif tag == 'tr' and self._current_tr:
expect (2 <= len (self._current_tr) <= 3)
name = self._current_tr[0].strip ()
tag = self._current_tr[1].strip ("\t\n\v\f\r '")
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
rank = 0
if len (tag) > 4:
expect (tag.endswith (' (deprecated)'), 'ill-formed OpenType tag: %s' % tag)
name += ' (deprecated)'
tag = tag.split (' ')[0]
rank = 1
self.names[tag] = re.sub (' languages$', '', name)
if not self._current_tr[2]:
return
iso_codes = self._current_tr[2].strip ()
self.to_bcp_47[tag].update (ISO_639_3_TO_1.get (code, code) for code in iso_codes.replace (' ', '').split (','))
rank += 2 * len (self.to_bcp_47[tag])
self.ranks[tag] = rank
def handle_data (self, data):
if self._td and not self._br:
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
self._current_tr[-1] += data
def handle_charref (self, name):
2021-03-28 16:09:13 +02:00
self.handle_data (html.unescape ('&#%s;' % name))
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
def handle_entityref (self, name):
2021-03-28 16:09:13 +02:00
self.handle_data (html.unescape ('&%s;' % name))
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
def parse (self, filename):
"""Parse the OpenType language system tag registry.
Args:
filename (str): The file name of the registry.
"""
2020-05-28 21:41:19 +02:00
with open (filename, encoding='utf-8') as f:
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
self.feed (f.read ())
expect (self.header)
for tag, iso_codes in self.to_bcp_47.items ():
for iso_code in iso_codes:
self.from_bcp_47[iso_code].add (tag)
def add_language (self, bcp_47_tag, ot_tag):
"""Add a language as if it were in the registry.
Args:
bcp_47_tag (str): A BCP 47 tag. If the tag is more than just
a language subtag, and if the language subtag is a
macrolanguage, then new languages are added corresponding
to the macrolanguages' individual languages with the
remainder of the tag appended.
ot_tag (str): An OpenType language system tag.
"""
global bcp_47
self.to_bcp_47[ot_tag].add (bcp_47_tag)
self.from_bcp_47[bcp_47_tag].add (ot_tag)
if bcp_47_tag.lower () not in bcp_47.grandfathered:
try:
[macrolanguage, suffix] = bcp_47_tag.split ('-', 1)
if macrolanguage in bcp_47.macrolanguages:
s = set ()
for language in bcp_47.macrolanguages[macrolanguage]:
if language.lower () not in bcp_47.grandfathered:
s.add ('%s-%s' % (language, suffix))
bcp_47.macrolanguages['%s-%s' % (macrolanguage, suffix)] = s
except ValueError:
pass
@staticmethod
def _remove_language (tag_1, dict_1, dict_2):
for tag_2 in dict_1.pop (tag_1):
dict_2[tag_2].remove (tag_1)
if not dict_2[tag_2]:
del dict_2[tag_2]
def remove_language_ot (self, ot_tag):
"""Remove an OpenType tag from the registry.
Args:
ot_tag (str): An OpenType tag.
"""
self._remove_language (ot_tag, self.to_bcp_47, self.from_bcp_47)
def remove_language_bcp_47 (self, bcp_47_tag):
"""Remove a BCP 47 tag from the registry.
Args:
bcp_47_tag (str): A BCP 47 tag.
"""
self._remove_language (bcp_47_tag, self.from_bcp_47, self.to_bcp_47)
def inherit_from_macrolanguages (self):
"""Copy mappings from macrolanguages to individual languages.
If a BCP 47 tag for an individual mapping has no OpenType
mapping but its macrolanguage does, the mapping is copied to
the individual language. For example, als (Tosk Albanian) has no
explicit mapping, so it inherits from sq (Albanian) the mapping
to SQI.
However, if an OpenType tag maps to a BCP 47 macrolanguage and
some but not all of its individual languages, the mapping is not
inherited from the macrolanguage to the missing individual
languages. For example, INUK (Nunavik Inuktitut) is mapped to
ike (Eastern Canadian Inuktitut) and iu (Inuktitut) but not to
ikt (Inuinnaqtun, which is an individual language of iu), so
this method does not add a mapping from ikt to INUK.
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
If a BCP 47 tag for a macrolanguage has no OpenType mapping but
some of its individual languages do, their mappings are copied
to the macrolanguage.
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
"""
global bcp_47
first_time = self.from_bcp_47_uninherited is None
if first_time:
self.from_bcp_47_uninherited = dict (self.from_bcp_47)
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
for macrolanguage, languages in dict (bcp_47.macrolanguages).items ():
ot_macrolanguages = {
ot_macrolanguage for ot_macrolanguage in self.from_bcp_47_uninherited.get (macrolanguage, set ())
}
blocked_ot_macrolanguages = set ()
if 'retired code' not in bcp_47.scopes.get (macrolanguage, ''):
for ot_macrolanguage in ot_macrolanguages:
round_trip_macrolanguages = {
l for l in self.to_bcp_47[ot_macrolanguage]
if 'retired code' not in bcp_47.scopes.get (l, '')
}
round_trip_languages = {
l for l in languages
if 'retired code' not in bcp_47.scopes.get (l, '')
}
intersection = round_trip_macrolanguages & round_trip_languages
if intersection and intersection != round_trip_languages:
blocked_ot_macrolanguages.add (ot_macrolanguage)
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
if ot_macrolanguages:
for ot_macrolanguage in ot_macrolanguages:
if ot_macrolanguage not in blocked_ot_macrolanguages:
for language in languages:
self.add_language (language, ot_macrolanguage)
if not blocked_ot_macrolanguages:
self.ranks[ot_macrolanguage] += 1
elif first_time:
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
for language in languages:
if language in self.from_bcp_47_uninherited:
ot_macrolanguages |= self.from_bcp_47_uninherited[language]
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
else:
ot_macrolanguages.clear ()
if not ot_macrolanguages:
break
for ot_macrolanguage in ot_macrolanguages:
self.add_language (macrolanguage, ot_macrolanguage)
def sort_languages (self):
"""Sort the values of ``from_bcp_47`` in ascending rank order."""
for language, tags in self.from_bcp_47.items ():
self.from_bcp_47[language] = sorted (tags,
key=lambda t: (self.ranks[t] + rank_delta (language, t), t))
ot = OpenTypeRegistryParser ()
class BCP47Parser (object):
"""A parser for the BCP 47 subtag registry.
Attributes:
header (str): The "File-Date" line of the registry.
names (Mapping[str, str]): A map of subtags to the names they
are given in the registry. Each value is a
``'\\n'``-separated list of names.
scopes (Mapping[str, str]): A map of language subtags to strings
suffixed to language names, including suffixes to explain
language scopes.
macrolanguages (DefaultDict[str, AbstractSet[str]]): A map of
language subtags to the sets of language subtags which
inherit from them. See
``OpenTypeRegistryParser.inherit_from_macrolanguages``.
prefixes (DefaultDict[str, AbstractSet[str]]): A map of variant
subtags to their prefixes.
grandfathered (AbstractSet[str]): The set of grandfathered tags,
normalized to lowercase.
"""
def __init__ (self):
self.header = ''
self.names = {}
self.scopes = {}
self.macrolanguages = collections.defaultdict (set)
self.prefixes = collections.defaultdict (set)
self.grandfathered = set ()
def parse (self, filename):
"""Parse the BCP 47 subtag registry.
Args:
filename (str): The file name of the registry.
"""
2020-05-28 21:41:19 +02:00
with open (filename, encoding='utf-8') as f:
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
subtag_type = None
subtag = None
deprecated = False
has_preferred_value = False
line_buffer = ''
for line in itertools.chain (f, ['']):
line = line.rstrip ()
if line.startswith (' '):
line_buffer += line[1:]
continue
line, line_buffer = line_buffer, line
if line.startswith ('Type: '):
subtag_type = line.split (' ')[1]
deprecated = False
has_preferred_value = False
elif line.startswith ('Subtag: ') or line.startswith ('Tag: '):
subtag = line.split (' ')[1]
if subtag_type == 'grandfathered':
self.grandfathered.add (subtag.lower ())
elif line.startswith ('Description: '):
description = line.split (' ', 1)[1].replace (' (individual language)', '')
description = re.sub (' (\(family\)|\((individual |macro)language\)|languages)$', '',
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
description)
if subtag in self.names:
self.names[subtag] += '\n' + description
else:
self.names[subtag] = description
elif subtag_type == 'language' or subtag_type == 'grandfathered':
if line.startswith ('Scope: '):
scope = line.split (' ')[1]
if scope == 'macrolanguage':
scope = ' [macrolanguage]'
elif scope == 'collection':
scope = ' [collection]'
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
else:
continue
self.scopes[subtag] = scope
elif line.startswith ('Deprecated: '):
self.scopes[subtag] = ' (retired code)' + self.scopes.get (subtag, '')
deprecated = True
elif deprecated and line.startswith ('Comments: see '):
# If a subtag is split into multiple replacement subtags,
# it essentially represents a macrolanguage.
for language in line.replace (',', '').split (' ')[2:]:
self._add_macrolanguage (subtag, language)
elif line.startswith ('Preferred-Value: '):
# If a subtag is deprecated in favor of a single replacement subtag,
# it is either a dialect or synonym of the preferred subtag. Either
# way, it is close enough to the truth to consider the replacement
# the macrolanguage of the deprecated language.
has_preferred_value = True
macrolanguage = line.split (' ')[1]
self._add_macrolanguage (macrolanguage, subtag)
elif not has_preferred_value and line.startswith ('Macrolanguage: '):
self._add_macrolanguage (line.split (' ')[1], subtag)
elif subtag_type == 'variant':
if line.startswith ('Deprecated: '):
self.scopes[subtag] = ' (retired code)' + self.scopes.get (subtag, '')
elif line.startswith ('Prefix: '):
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
self.prefixes[subtag].add (line.split (' ')[1])
elif line.startswith ('File-Date: '):
self.header = line
expect (self.header)
def _add_macrolanguage (self, macrolanguage, language):
global ot
if language not in ot.from_bcp_47:
for l in self.macrolanguages.get (language, set ()):
self._add_macrolanguage (macrolanguage, l)
if macrolanguage not in ot.from_bcp_47:
for ls in list (self.macrolanguages.values ()):
if macrolanguage in ls:
ls.add (language)
return
self.macrolanguages[macrolanguage].add (language)
def remove_extra_macrolanguages (self):
"""Make every language have at most one macrolanguage."""
inverted = collections.defaultdict (list)
for macrolanguage, languages in self.macrolanguages.items ():
for language in languages:
inverted[language].append (macrolanguage)
for language, macrolanguages in inverted.items ():
if len (macrolanguages) > 1:
macrolanguages.sort (key=lambda ml: len (self.macrolanguages[ml]))
biggest_macrolanguage = macrolanguages.pop ()
for macrolanguage in macrolanguages:
self._add_macrolanguage (biggest_macrolanguage, macrolanguage)
def _get_name_piece (self, subtag):
"""Return the first name of a subtag plus its scope suffix.
Args:
subtag (str): A BCP 47 subtag.
Returns:
The name form of ``subtag``.
"""
return self.names[subtag].split ('\n')[0] + self.scopes.get (subtag, '')
def get_name (self, lt):
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
"""Return the names of the subtags in a language tag.
Args:
lt (LanguageTag): A BCP 47 language tag.
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
Returns:
The name form of ``lt``.
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
"""
name = self._get_name_piece (lt.language)
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
if lt.script:
name += '; ' + self._get_name_piece (lt.script.title ())
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
if lt.region:
name += '; ' + self._get_name_piece (lt.region.upper ())
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
if lt.variant:
name += '; ' + self._get_name_piece (lt.variant)
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
return name
bcp_47 = BCP47Parser ()
ot.parse (sys.argv[1])
bcp_47.parse (sys.argv[2])
ot.add_language ('ary', 'MOR')
ot.add_language ('ath', 'ATH')
ot.add_language ('bai', 'BML')
ot.ranks['BAL'] = ot.ranks['KAR'] + 1
ot.add_language ('ber', 'BBR')
ot.remove_language_ot ('PGR')
ot.add_language ('el-polyton', 'PGR')
bcp_47.macrolanguages['et'] = {'ekk'}
bcp_47.names['flm'] = 'Falam Chin'
bcp_47.scopes['flm'] = ' (retired code)'
bcp_47.macrolanguages['flm'] = {'cfm'}
ot.ranks['FNE'] = ot.ranks['TNE'] + 1
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
ot.add_language ('und-fonipa', 'IPPH')
ot.add_language ('und-fonnapa', 'APPH')
ot.remove_language_ot ('IRT')
ot.add_language ('ga-Latg', 'IRT')
ot.add_language ('hy-arevmda', 'HYE')
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
ot.remove_language_ot ('KGE')
ot.add_language ('und-Geok', 'KGE')
bcp_47.macrolanguages['id'] = {'in'}
bcp_47.macrolanguages['ijo'] = {'ijc'}
ot.add_language ('kht', 'KHN')
ot.names['KHN'] = ot.names['KHT'] + ' (Microsoft fonts)'
ot.ranks['KHN'] = ot.ranks['KHT'] + 1
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
ot.ranks['LCR'] = ot.ranks['MCR'] + 1
ot.names['MAL'] = 'Malayalam Traditional'
ot.ranks['MLR'] += 1
bcp_47.names['mhv'] = 'Arakanese'
bcp_47.scopes['mhv'] = ' (retired code)'
ot.add_language ('mnw-TH', 'MONT')
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
ot.add_language ('no', 'NOR')
ot.add_language ('oc-provenc', 'PRO')
ot.remove_language_ot ('QUZ')
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
ot.add_language ('qu', 'QUZ')
ot.add_language ('qub', 'QWH')
ot.add_language ('qud', 'QVI')
ot.add_language ('qug', 'QVI')
ot.add_language ('qul', 'QUH')
ot.add_language ('qup', 'QVI')
ot.add_language ('qur', 'QWH')
ot.add_language ('qus', 'QUH')
ot.add_language ('quw', 'QVI')
ot.add_language ('qux', 'QWH')
ot.add_language ('qva', 'QWH')
ot.add_language ('qvh', 'QWH')
ot.add_language ('qvj', 'QVI')
ot.add_language ('qvl', 'QWH')
ot.add_language ('qvm', 'QWH')
ot.add_language ('qvn', 'QWH')
ot.add_language ('qvo', 'QVI')
ot.add_language ('qvp', 'QWH')
ot.add_language ('qvw', 'QWH')
ot.add_language ('qvz', 'QVI')
ot.add_language ('qwa', 'QWH')
ot.add_language ('qws', 'QWH')
ot.add_language ('qxa', 'QWH')
ot.add_language ('qxc', 'QWH')
ot.add_language ('qxh', 'QWH')
ot.add_language ('qxl', 'QVI')
ot.add_language ('qxn', 'QWH')
ot.add_language ('qxo', 'QWH')
ot.add_language ('qxr', 'QVI')
ot.add_language ('qxt', 'QWH')
ot.add_language ('qxw', 'QWH')
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
bcp_47.macrolanguages['ro-MD'].add ('mo')
ot.remove_language_ot ('SYRE')
ot.remove_language_ot ('SYRJ')
ot.remove_language_ot ('SYRN')
ot.add_language ('und-Syre', 'SYRE')
ot.add_language ('und-Syrj', 'SYRJ')
ot.add_language ('und-Syrn', 'SYRN')
bcp_47.names['xst'] = "Silt'e"
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
bcp_47.scopes['xst'] = ' (retired code)'
bcp_47.macrolanguages['xst'] = {'stv', 'wle'}
ot.add_language ('xwo', 'TOD')
ot.remove_language_ot ('ZHH')
ot.remove_language_ot ('ZHP')
ot.remove_language_ot ('ZHT')
ot.remove_language_ot ('ZHTM')
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
bcp_47.macrolanguages['zh'].remove ('lzh')
bcp_47.macrolanguages['zh'].remove ('yue')
ot.add_language ('zh-Hant-MO', 'ZHH')
ot.add_language ('zh-Hant-MO', 'ZHTM')
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
ot.add_language ('zh-Hant-HK', 'ZHH')
ot.add_language ('zh-Hans', 'ZHS')
ot.add_language ('zh-Hant', 'ZHT')
ot.add_language ('zh-HK', 'ZHH')
ot.add_language ('zh-MO', 'ZHH')
ot.add_language ('zh-MO', 'ZHTM')
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
ot.add_language ('zh-TW', 'ZHT')
ot.add_language ('lzh', 'ZHT')
ot.add_language ('lzh-Hans', 'ZHS')
ot.add_language ('yue', 'ZHH')
ot.add_language ('yue-Hans', 'ZHS')
bcp_47.macrolanguages['zom'] = {'yos'}
def rank_delta (bcp_47, ot):
"""Return a delta to apply to a BCP 47 tag's rank.
Most OpenType tags have a constant rank, but a few have ranks that
depend on the BCP 47 tag.
Args:
bcp_47 (str): A BCP 47 tag.
ot (str): An OpenType tag to.
Returns:
A number to add to ``ot``'s rank when sorting ``bcp_47``'s
OpenType equivalents.
"""
if bcp_47 == 'ak' and ot == 'AKA':
return -1
if bcp_47 == 'tw' and ot == 'TWI':
return -1
return 0
disambiguation = {
'ALT': 'alt',
'ARK': 'rki',
'ATH': 'ath',
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
'BHI': 'bhb',
'BLN': 'bjt',
'BTI': 'beb',
'CCHN': 'cco',
'CMR': 'swb',
'CPP': 'crp',
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
'CRR': 'crx',
'DUJ': 'dwu',
'ECR': 'crj',
'HAL': 'cfm',
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
'HND': 'hnd',
'HYE': 'hyw',
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
'KIS': 'kqs',
'KUI': 'uki',
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
'LRC': 'bqi',
'NDB': 'nd',
'NIS': 'njz',
'PLG': 'pce',
'PRO': 'pro',
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
'QIN': 'bgr',
'QUH': 'quh',
'QVI': 'qvi',
'QWH': 'qwh',
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
'SIG': 'stv',
'SRB': 'sr',
2021-03-08 19:12:47 +01:00
'SXT': 'xnj',
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
'ZHH': 'zh-HK',
'ZHS': 'zh-Hans',
'ZHT': 'zh-Hant',
'ZHTM': 'zh-MO',
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
}
ot.inherit_from_macrolanguages ()
bcp_47.remove_extra_macrolanguages ()
ot.inherit_from_macrolanguages ()
ot.names[DEFAULT_LANGUAGE_SYSTEM] = '*/'
ot.ranks[DEFAULT_LANGUAGE_SYSTEM] = max (ot.ranks.values ()) + 1
for tricky_ot_tag in filter (lambda tag: re.match ('[A-Z]{3}$', tag), ot.names):
possible_bcp_47_tag = tricky_ot_tag.lower ()
if possible_bcp_47_tag in bcp_47.names and not ot.from_bcp_47[possible_bcp_47_tag]:
ot.add_language (possible_bcp_47_tag, DEFAULT_LANGUAGE_SYSTEM)
bcp_47.macrolanguages[possible_bcp_47_tag] = set ()
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
ot.sort_languages ()
print ('/* == Start of generated table == */')
print ('/*')
print (' * The following table is generated by running:')
print (' *')
print (' * %s languagetags language-subtag-registry' % sys.argv[0])
print (' *')
print (' * on files with these headers:')
print (' *')
print (' * %s' % ot.header.strip ())
print (' * %s' % bcp_47.header)
print (' */')
print ()
print ('#ifndef HB_OT_TAG_TABLE_HH')
print ('#define HB_OT_TAG_TABLE_HH')
print ()
def hb_tag (tag):
"""Convert a tag to ``HB_TAG`` form.
Args:
tag (str): An OpenType tag.
Returns:
A snippet of C++ representing ``tag``.
"""
if tag == DEFAULT_LANGUAGE_SYSTEM:
return 'HB_TAG_NONE\t '
return "HB_TAG('%s','%s','%s','%s')" % tuple (('%-4s' % tag)[:4])
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
def get_variant_set (name):
"""Return a set of variant language names from a name.
Args:
name (str): A list of language names from the BCP 47 registry,
joined on ``'\\n'``.
Returns:
A set of normalized language names.
"""
return set (unicodedata.normalize ('NFD', n.replace ('\u2019', "'"))
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
.encode ('ASCII', 'ignore')
.strip ()
for n in re.split ('[\n(),]', name) if n)
def language_name_intersection (a, b):
"""Return the names in common between two language names.
Args:
a (str): A list of language names from the BCP 47 registry,
joined on ``'\\n'``.
b (str): A list of language names from the BCP 47 registry,
joined on ``'\\n'``.
Returns:
The normalized language names shared by ``a`` and ``b``.
"""
return get_variant_set (a).intersection (get_variant_set (b))
def get_matching_language_name (intersection, candidates):
return next (iter (c for c in candidates if not intersection.isdisjoint (get_variant_set (c))))
def same_tag (bcp_47_tag, ot_tags):
return len (bcp_47_tag) == 3 and len (ot_tags) == 1 and bcp_47_tag == ot_tags[0].lower ()
[ot-tags] Speed up hb_ot_tags_from_language() Part of https://github.com/harfbuzz/harfbuzz/issues/3591 "After that, bulk of the time I suppose is spent in binary-searching the language table. I suggest we split the language table in 2-letter and 3-letter tags, to speed-up the vast majority of cases that are 2-letter." benchmark-ot, before: ---------------------------------------------------------------------------------------------- Benchmark Time CPU Iterations ---------------------------------------------------------------------------------------------- BM_hb_ot_tags_from_script_and_language/COMMON zh_CN 112 ns 111 ns 6286271 BM_hb_ot_tags_from_script_and_language/COMMON en_US 60.6 ns 60.4 ns 11671176 BM_hb_ot_tags_from_script_and_language/LATIN en_US 61.3 ns 61.1 ns 11442645 BM_hb_ot_tags_from_script_and_language/COMMON none 4.75 ns 4.74 ns 146997235 BM_hb_ot_tags_from_script_and_language/LATIN none 4.65 ns 4.64 ns 150938747 After: ---------------------------------------------------------------------------------------------- Benchmark Time CPU Iterations ---------------------------------------------------------------------------------------------- BM_hb_ot_tags_from_script_and_language/COMMON zh_CN 89.5 ns 89.2 ns 7747649 BM_hb_ot_tags_from_script_and_language/COMMON en_US 38.5 ns 38.4 ns 18199432 BM_hb_ot_tags_from_script_and_language/LATIN en_US 39.0 ns 38.9 ns 18049238 BM_hb_ot_tags_from_script_and_language/COMMON none 4.53 ns 4.52 ns 154895110 BM_hb_ot_tags_from_script_and_language/LATIN none 4.54 ns 4.52 ns 154762105
2022-05-17 22:28:28 +02:00
for language_len in (2, 3):
if language_len == 3:
print ('#ifndef HB_NO_LANGUAGE_LONG')
[ot-tags] Speed up hb_ot_tags_from_language() Part of https://github.com/harfbuzz/harfbuzz/issues/3591 "After that, bulk of the time I suppose is spent in binary-searching the language table. I suggest we split the language table in 2-letter and 3-letter tags, to speed-up the vast majority of cases that are 2-letter." benchmark-ot, before: ---------------------------------------------------------------------------------------------- Benchmark Time CPU Iterations ---------------------------------------------------------------------------------------------- BM_hb_ot_tags_from_script_and_language/COMMON zh_CN 112 ns 111 ns 6286271 BM_hb_ot_tags_from_script_and_language/COMMON en_US 60.6 ns 60.4 ns 11671176 BM_hb_ot_tags_from_script_and_language/LATIN en_US 61.3 ns 61.1 ns 11442645 BM_hb_ot_tags_from_script_and_language/COMMON none 4.75 ns 4.74 ns 146997235 BM_hb_ot_tags_from_script_and_language/LATIN none 4.65 ns 4.64 ns 150938747 After: ---------------------------------------------------------------------------------------------- Benchmark Time CPU Iterations ---------------------------------------------------------------------------------------------- BM_hb_ot_tags_from_script_and_language/COMMON zh_CN 89.5 ns 89.2 ns 7747649 BM_hb_ot_tags_from_script_and_language/COMMON en_US 38.5 ns 38.4 ns 18199432 BM_hb_ot_tags_from_script_and_language/LATIN en_US 39.0 ns 38.9 ns 18049238 BM_hb_ot_tags_from_script_and_language/COMMON none 4.53 ns 4.52 ns 154895110 BM_hb_ot_tags_from_script_and_language/LATIN none 4.54 ns 4.52 ns 154762105
2022-05-17 22:28:28 +02:00
print ('static const LangTag ot_languages%d[] = {' % language_len)
for language, tags in sorted (ot.from_bcp_47.items ()):
if language == '' or '-' in language:
continue
if len(language) != language_len: continue
commented_out = same_tag (language, tags)
for i, tag in enumerate (tags, start=1):
[ot-tags] Further speed up language bsearch() Using an integer tag to bsearch, instead of string. Part of: https://github.com/harfbuzz/harfbuzz/issues/3591 Before: ------------------------------------------------------------------------------------------------ Benchmark Time CPU Iterations ------------------------------------------------------------------------------------------------ BM_hb_ot_tags_from_script_and_language/COMMON abcd_XY 8.11 ns 8.08 ns 87067795 BM_hb_ot_tags_from_script_and_language/COMMON zh_CN 53.6 ns 53.5 ns 13042418 BM_hb_ot_tags_from_script_and_language/COMMON en_US 24.2 ns 24.1 ns 29052731 BM_hb_ot_tags_from_script_and_language/LATIN en_US 24.4 ns 24.3 ns 28736769 BM_hb_ot_tags_from_script_and_language/COMMON none 4.43 ns 4.41 ns 160370413 BM_hb_ot_tags_from_script_and_language/LATIN none 4.35 ns 4.34 ns 160578191 After: ------------------------------------------------------------------------------------------------ Benchmark Time CPU Iterations ------------------------------------------------------------------------------------------------ BM_hb_ot_tags_from_script_and_language/COMMON abcd_XY 7.97 ns 7.95 ns 85208363 BM_hb_ot_tags_from_script_and_language/COMMON zh_CN 41.7 ns 41.6 ns 16945817 BM_hb_ot_tags_from_script_and_language/COMMON en_US 16.1 ns 16.0 ns 43613523 BM_hb_ot_tags_from_script_and_language/LATIN en_US 16.5 ns 16.4 ns 42568107 BM_hb_ot_tags_from_script_and_language/COMMON none 4.30 ns 4.29 ns 164055469 BM_hb_ot_tags_from_script_and_language/LATIN none 4.29 ns 4.27 ns 163793591
2022-05-17 23:51:41 +02:00
print ('%s{%s,\t%s},' % ('/*' if commented_out else ' ', hb_tag (language), hb_tag (tag)), end='')
[ot-tags] Speed up hb_ot_tags_from_language() Part of https://github.com/harfbuzz/harfbuzz/issues/3591 "After that, bulk of the time I suppose is spent in binary-searching the language table. I suggest we split the language table in 2-letter and 3-letter tags, to speed-up the vast majority of cases that are 2-letter." benchmark-ot, before: ---------------------------------------------------------------------------------------------- Benchmark Time CPU Iterations ---------------------------------------------------------------------------------------------- BM_hb_ot_tags_from_script_and_language/COMMON zh_CN 112 ns 111 ns 6286271 BM_hb_ot_tags_from_script_and_language/COMMON en_US 60.6 ns 60.4 ns 11671176 BM_hb_ot_tags_from_script_and_language/LATIN en_US 61.3 ns 61.1 ns 11442645 BM_hb_ot_tags_from_script_and_language/COMMON none 4.75 ns 4.74 ns 146997235 BM_hb_ot_tags_from_script_and_language/LATIN none 4.65 ns 4.64 ns 150938747 After: ---------------------------------------------------------------------------------------------- Benchmark Time CPU Iterations ---------------------------------------------------------------------------------------------- BM_hb_ot_tags_from_script_and_language/COMMON zh_CN 89.5 ns 89.2 ns 7747649 BM_hb_ot_tags_from_script_and_language/COMMON en_US 38.5 ns 38.4 ns 18199432 BM_hb_ot_tags_from_script_and_language/LATIN en_US 39.0 ns 38.9 ns 18049238 BM_hb_ot_tags_from_script_and_language/COMMON none 4.53 ns 4.52 ns 154895110 BM_hb_ot_tags_from_script_and_language/LATIN none 4.54 ns 4.52 ns 154762105
2022-05-17 22:28:28 +02:00
if commented_out:
print ('*/', end='')
print ('\t/* ', end='')
bcp_47_name = bcp_47.names.get (language, '')
bcp_47_name_candidates = bcp_47_name.split ('\n')
ot_name = ot.names[tag]
scope = bcp_47.scopes.get (language, '')
if tag == DEFAULT_LANGUAGE_SYSTEM:
write (f'{bcp_47_name_candidates[0]}{scope} != {ot.names[language.upper ()]}')
else:
[ot-tags] Speed up hb_ot_tags_from_language() Part of https://github.com/harfbuzz/harfbuzz/issues/3591 "After that, bulk of the time I suppose is spent in binary-searching the language table. I suggest we split the language table in 2-letter and 3-letter tags, to speed-up the vast majority of cases that are 2-letter." benchmark-ot, before: ---------------------------------------------------------------------------------------------- Benchmark Time CPU Iterations ---------------------------------------------------------------------------------------------- BM_hb_ot_tags_from_script_and_language/COMMON zh_CN 112 ns 111 ns 6286271 BM_hb_ot_tags_from_script_and_language/COMMON en_US 60.6 ns 60.4 ns 11671176 BM_hb_ot_tags_from_script_and_language/LATIN en_US 61.3 ns 61.1 ns 11442645 BM_hb_ot_tags_from_script_and_language/COMMON none 4.75 ns 4.74 ns 146997235 BM_hb_ot_tags_from_script_and_language/LATIN none 4.65 ns 4.64 ns 150938747 After: ---------------------------------------------------------------------------------------------- Benchmark Time CPU Iterations ---------------------------------------------------------------------------------------------- BM_hb_ot_tags_from_script_and_language/COMMON zh_CN 89.5 ns 89.2 ns 7747649 BM_hb_ot_tags_from_script_and_language/COMMON en_US 38.5 ns 38.4 ns 18199432 BM_hb_ot_tags_from_script_and_language/LATIN en_US 39.0 ns 38.9 ns 18049238 BM_hb_ot_tags_from_script_and_language/COMMON none 4.53 ns 4.52 ns 154895110 BM_hb_ot_tags_from_script_and_language/LATIN none 4.54 ns 4.52 ns 154762105
2022-05-17 22:28:28 +02:00
intersection = language_name_intersection (bcp_47_name, ot_name)
if not intersection:
write ('%s%s -> %s' % (bcp_47_name_candidates[0], scope, ot_name))
else:
name = get_matching_language_name (intersection, bcp_47_name_candidates)
bcp_47.names[language] = name
write ('%s%s' % (name if len (name) > len (ot_name) else ot_name, scope))
print (' */')
print ('};')
if language_len == 3:
print ('#endif')
[ot-tags] Speed up hb_ot_tags_from_language() Part of https://github.com/harfbuzz/harfbuzz/issues/3591 "After that, bulk of the time I suppose is spent in binary-searching the language table. I suggest we split the language table in 2-letter and 3-letter tags, to speed-up the vast majority of cases that are 2-letter." benchmark-ot, before: ---------------------------------------------------------------------------------------------- Benchmark Time CPU Iterations ---------------------------------------------------------------------------------------------- BM_hb_ot_tags_from_script_and_language/COMMON zh_CN 112 ns 111 ns 6286271 BM_hb_ot_tags_from_script_and_language/COMMON en_US 60.6 ns 60.4 ns 11671176 BM_hb_ot_tags_from_script_and_language/LATIN en_US 61.3 ns 61.1 ns 11442645 BM_hb_ot_tags_from_script_and_language/COMMON none 4.75 ns 4.74 ns 146997235 BM_hb_ot_tags_from_script_and_language/LATIN none 4.65 ns 4.64 ns 150938747 After: ---------------------------------------------------------------------------------------------- Benchmark Time CPU Iterations ---------------------------------------------------------------------------------------------- BM_hb_ot_tags_from_script_and_language/COMMON zh_CN 89.5 ns 89.2 ns 7747649 BM_hb_ot_tags_from_script_and_language/COMMON en_US 38.5 ns 38.4 ns 18199432 BM_hb_ot_tags_from_script_and_language/LATIN en_US 39.0 ns 38.9 ns 18049238 BM_hb_ot_tags_from_script_and_language/COMMON none 4.53 ns 4.52 ns 154895110 BM_hb_ot_tags_from_script_and_language/LATIN none 4.54 ns 4.52 ns 154762105
2022-05-17 22:28:28 +02:00
print ()
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
print ('/**')
print (' * hb_ot_tags_from_complex_language:')
print (' * @lang_str: a BCP 47 language tag to convert.')
print (' * @limit: a pointer to the end of the substring of @lang_str to consider for')
print (' * conversion.')
print (' * @count: maximum number of language tags to retrieve (IN) and actual number of')
print (' * language tags retrieved (OUT). If no tags are retrieved, it is not modified.')
print (' * @tags: array of size at least @language_count to store the language tag')
print (' * results')
print (' *')
print (' * Converts a multi-subtag BCP 47 language tag to language tags.')
print (' *')
print (' * Return value: Whether any language systems were retrieved.')
print (' **/')
print ('static inline bool')
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
print ('hb_ot_tags_from_complex_language (const char *lang_str,')
print ('\t\t\t\t const char *limit,')
print ('\t\t\t\t unsigned int *count /* IN/OUT */,')
print ('\t\t\t\t hb_tag_t *tags /* OUT */)')
print ('{')
def print_subtag_matches (subtag, string, new_line):
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
if subtag:
if new_line:
print ()
print ('\t&& ', end='')
print ('subtag_matches (%s, limit, "-%s", %i)' % (string, subtag, 1 + len (subtag)), end='')
complex_tags = collections.defaultdict (list)
for initial, group in itertools.groupby ((lt_tags for lt_tags in [
(LanguageTag (language), tags)
for language, tags in sorted (ot.from_bcp_47.items (),
key=lambda i: (-len (i[0]), i[0]))
] if lt_tags[0].is_complex ()),
key=lambda lt_tags: lt_tags[0].get_group ()):
complex_tags[initial] += group
2022-05-17 21:34:34 +02:00
# Calculate the min length of the subtags outside the switch
min_subtag_len = 100
for initial, items in sorted (complex_tags.items ()):
if initial != 'und':
continue
for lt, tags in items:
if not tags:
continue
subtag_len = 0
subtag_len += 1 + len (lt.script) if lt.script is not None else 0
subtag_len += 1 + len (lt.region) if lt.region is not None else 0
subtag_len += 1 + len (lt.variant) if lt.variant is not None else 0
2022-05-17 21:34:34 +02:00
min_subtag_len = min(subtag_len, min_subtag_len)
print (' if (limit - lang_str >= %d)' % (min_subtag_len + 2))
2022-05-17 21:34:34 +02:00
print (' {')
print (" const char *p = strchr (lang_str, '-');")
print (" if (!p || p >= limit || limit - p < %i) goto out;" % min_subtag_len)
for initial, items in sorted (complex_tags.items ()):
if initial != 'und':
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
continue
for lt, tags in items:
if not tags:
continue
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
if lt.variant in bcp_47.prefixes:
expect (next (iter (bcp_47.prefixes[lt.variant])) == lt.language,
'%s is not a valid prefix of %s' % (lt.language, lt.variant))
2022-05-17 21:34:34 +02:00
print (' if (', end='')
print_subtag_matches (lt.script, 'p', False)
print_subtag_matches (lt.region, 'p', False)
print_subtag_matches (lt.variant, 'p', False)
print (')')
2022-05-17 21:34:34 +02:00
print (' {')
write (' /* %s */' % bcp_47.get_name (lt))
print ()
if len (tags) == 1:
2022-05-17 21:34:34 +02:00
write (' tags[0] = %s; /* %s */' % (hb_tag (tags[0]), ot.names[tags[0]]))
print ()
2022-05-17 21:34:34 +02:00
print (' *count = 1;')
else:
print (' hb_tag_t possible_tags[] = {')
for tag in tags:
write (' %s, /* %s */' % (hb_tag (tag), ot.names[tag]))
print ()
2022-05-17 21:34:34 +02:00
print (' };')
print (' for (i = 0; i < %s && i < *count; i++)' % len (tags))
print ('\ttags[i] = possible_tags[i];')
2022-05-17 21:34:34 +02:00
print (' *count = i;')
print (' return true;')
print (' }')
print (' }')
print ('out:')
print (' switch (lang_str[0])')
print (' {')
for initial, items in sorted (complex_tags.items ()):
if initial == 'und':
continue
print (" case '%s':" % initial)
for lt, tags in items:
if not tags:
continue
print (' if (', end='')
script = lt.script
region = lt.region
if lt.grandfathered:
print ('0 == strcmp (&lang_str[1], "%s")' % lt.language[1:], end='')
else:
string_literal = lt.language[1:] + '-'
if script:
string_literal += script
script = None
if region:
string_literal += '-' + region
region = None
if string_literal[-1] == '-':
print ('0 == strncmp (&lang_str[1], "%s", %i)' % (string_literal, len (string_literal)), end='')
else:
2022-05-17 22:57:08 +02:00
print ('lang_matches (&lang_str[1], limit, "%s", %i)' % (string_literal, len (string_literal)), end='')
print_subtag_matches (script, 'lang_str', True)
print_subtag_matches (region, 'lang_str', True)
print_subtag_matches (lt.variant, 'lang_str', True)
print (')')
print (' {')
write (' /* %s */' % bcp_47.get_name (lt))
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
print ()
if len (tags) == 1:
write (' tags[0] = %s; /* %s */' % (hb_tag (tags[0]), ot.names[tags[0]]))
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
print ()
print (' *count = 1;')
else:
print (' unsigned int i;')
print (' hb_tag_t possible_tags[] = {')
for tag in tags:
write ('\t%s, /* %s */' % (hb_tag (tag), ot.names[tag]))
print ()
print (' };')
print (' for (i = 0; i < %s && i < *count; i++)' % len (tags))
print ('\ttags[i] = possible_tags[i];')
print (' *count = i;')
print (' return true;')
print (' }')
print (' break;')
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
print (' }')
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
print (' return false;')
print ('}')
print ()
print ('/**')
print (' * hb_ot_ambiguous_tag_to_language')
print (' * @tag: A language tag.')
print (' *')
print (' * Converts @tag to a BCP 47 language tag if it is ambiguous (it corresponds to')
print (' * many language tags) and the best tag is not the alphabetically first, or if')
print (' * the best tag consists of multiple subtags, or if the best tag does not appear')
print (' * in #ot_languages.')
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
print (' *')
print (' * Return value: The #hb_language_t corresponding to the BCP 47 language tag,')
print (' * or #HB_LANGUAGE_INVALID if @tag is not ambiguous.')
print (' **/')
print ('static inline hb_language_t')
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
print ('hb_ot_ambiguous_tag_to_language (hb_tag_t tag)')
print ('{')
print (' switch (tag)')
print (' {')
def verify_disambiguation_dict ():
"""Verify and normalize ``disambiguation``.
``disambiguation`` is a map of ambiguous OpenType language system
tags to the particular BCP 47 tags they correspond to. This function
checks that all its keys really are ambiguous and that each key's
value is valid for that key. It checks that no ambiguous tag is
missing, except when it can figure out which BCP 47 tag is the best
by itself.
It modifies ``disambiguation`` to remove keys whose values are the
same as those that the fallback would return anyway, and to add
ambiguous keys whose disambiguations it determined automatically.
Raises:
AssertionError: Verification failed.
"""
global bcp_47
global disambiguation
global ot
for ot_tag, bcp_47_tags in ot.to_bcp_47.items ():
if ot_tag == DEFAULT_LANGUAGE_SYSTEM:
primary_tags = []
else:
primary_tags = list (t for t in bcp_47_tags if t not in bcp_47.grandfathered and ot.from_bcp_47.get (t)[0] == ot_tag)
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
if len (primary_tags) == 1:
expect (ot_tag not in disambiguation, 'unnecessary disambiguation for OT tag: %s' % ot_tag)
if '-' in primary_tags[0]:
disambiguation[ot_tag] = primary_tags[0]
else:
first_tag = sorted (t for t in bcp_47_tags if t not in bcp_47.grandfathered and ot_tag in ot.from_bcp_47.get (t))[0]
if primary_tags[0] != first_tag:
disambiguation[ot_tag] = primary_tags[0]
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
elif len (primary_tags) == 0:
expect (ot_tag not in disambiguation, 'There is no possible valid disambiguation for %s' % ot_tag)
else:
original_languages = [t for t in primary_tags if t in ot.from_bcp_47_uninherited and 'retired code' not in bcp_47.scopes.get (t, '')]
if len (original_languages) == 1:
macrolanguages = original_languages
else:
macrolanguages = [t for t in primary_tags if bcp_47.scopes.get (t) == ' [macrolanguage]']
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
if len (macrolanguages) != 1:
macrolanguages = list (t for t in primary_tags if bcp_47.scopes.get (t) == ' [collection]')
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
if len (macrolanguages) != 1:
macrolanguages = list (t for t in primary_tags if 'retired code' not in bcp_47.scopes.get (t, ''))
if len (macrolanguages) != 1:
expect (ot_tag in disambiguation, 'ambiguous OT tag: %s %s' % (ot_tag, str (macrolanguages)))
expect (disambiguation[ot_tag] in bcp_47_tags,
'%s is not a valid disambiguation for %s' % (disambiguation[ot_tag], ot_tag))
elif ot_tag not in disambiguation:
disambiguation[ot_tag] = macrolanguages[0]
different_bcp_47_tags = sorted (t for t in bcp_47_tags if not same_tag (t, ot.from_bcp_47.get (t)))
if different_bcp_47_tags and disambiguation[ot_tag] == different_bcp_47_tags[0] and '-' not in disambiguation[ot_tag]:
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
del disambiguation[ot_tag]
for ot_tag in disambiguation.keys ():
expect (ot_tag in ot.to_bcp_47, 'unknown OT tag: %s' % ot_tag)
verify_disambiguation_dict ()
for ot_tag, bcp_47_tag in sorted (disambiguation.items ()):
write (' case %s: /* %s */' % (hb_tag (ot_tag), ot.names[ot_tag]))
print ()
write (' return hb_language_from_string (\"%s\", -1); /* %s */' % (bcp_47_tag, bcp_47.get_name (LanguageTag (bcp_47_tag))))
Autogenerate the BCP 47 to OpenType mappings The new script, gen-tag-table.py, generates `ot_languages` automatically from the [OpenType language system tag registry][ot] and the [IANA Language Subtag Registry][bcp47] with some manual modifications. If an OpenType tag maps to a BCP 47 macrolanguage, all the macrolanguage's individual languages are mapped to the same OpenType tag, except for individual languages with their own OpenType mappings. Deprecated BCP 47 tags are canonicalized. [ot]: https://docs.microsoft.com/en-us/typography/opentype/spec/languagetags [bcp47]: https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry Some OpenType tags correspond to multiple ISO 639 codes. The mapping from ISO 639 codes lists OpenType tags in priority order, such that more specific or more likely tags appear first. Some OpenType tags have no corresponding ISO 639 code in the registry so their mappings use BCP 47 subtags besides the language. For example, any BCP 47 tag with a fonipa variant subtag is mapped to 'IPPH', and 'IPPH' is mapped back to und-fonipa. Other OpenType tags have no corresponding ISO 639 code because it is not clear what they are for. HarfBuzz just ignores these tags. One such ignored tag is 'ZHP ' (Chinese Phonetic). It probably means zh-Latn. However, it is used in Microsoft JhengHei and Microsoft YaHei with the script tag 'hani', implying that it is not a romanization scheme after all. It would be simple enough to add this mapping to gen-tag-table.py once a definitive mapping is determined. The manual modifications are mainly either obvious mappings that the OpenType registry omits or mappings for compatibility with previous versions of HarfBuzz. Some of the old mappings were discarded, though, for homophonous language names. For example, OpenType maps 'KUI ' to kxu; previous versions of HarfBuzz also mapped it to kvd, because kvd and kxu both happen to be called "Kui". gen-tag-table.py also generates a function to convert multi-subtag tags like el-polyton and zh-HK to OpenType tags, replacing `ot_languages_zh` and the hard-coded list of special cases in `hb_ot_tags_from_language`. It also generates a function to convert OpenType tags to BCP 47, replacing the hard-coded list of special cases in `hb_ot_tag_to_language`.
2017-12-09 04:45:52 +01:00
print ()
print (' default:')
print (' return HB_LANGUAGE_INVALID;')
print (' }')
print ('}')
print ()
print ('#endif /* HB_OT_TAG_TABLE_HH */')
print ()
print ('/* == End of generated table == */')