diff options
Diffstat (limited to 'src/autofit/afblue.dat')
| -rw-r--r-- | src/autofit/afblue.dat | 295 |
1 files changed, 221 insertions, 74 deletions
diff --git a/src/autofit/afblue.dat b/src/autofit/afblue.dat index fd5684e..5ffb852 100644 --- a/src/autofit/afblue.dat +++ b/src/autofit/afblue.dat @@ -20,9 +20,8 @@ // labels separated by whitespace and followed by a colon (everything in a // single line); the first label gives the name of the enumeration template, // the second the name of the array template, and the third the name of the -// `maximum' template, holding the size of the largest array element. The -// script then fills the corresponding templates (indicated by `@' -// characters around the name). +// `maximum' template. The script then fills the corresponding templates +// (indicated by `@' characters around the name). // // A section contains one or more data records. Each data record consists // of two or more lines. The first line holds the enumeration name, and the @@ -30,15 +29,20 @@ // // There are two possible representations for array data. // -// - A string of characters in UTF-8 encoding enclosed in double quotes, -// using C syntax. There can be only one string per line, thus the -// starting and ending double quote must be the first and last character -// in the line, respectively, ignoring whitespace before and after the -// string. Space characters within the string are ignored too. If there -// are multiple strings (in multiple lines), they are concatenated to a -// single string. In the output, a string gets represented as a series of -// singles bytes, followed by a zero byte. The enumeration values simply -// hold byte offsets to the start of the corresponding strings. +// - A string of characters or character clusters (for example, representing +// Aksharas, Devanagari syllables) in UTF-8 encoding enclosed in double +// quotes, using C syntax, where the elements are separated by spaces. +// There can be only one string per line, thus the starting and ending +// double quote must be the first and last character in the line, +// respectively, ignoring whitespace before and after the string. If +// there are multiple strings (in multiple lines), they are concatenated +// to a single string. In the output, a string gets represented as a +// series of singles bytes, followed by a zero byte. The enumeration +// values simply hold byte offsets to the start of the corresponding +// strings. +// +// For strings, the `maximum' template holds the maximum number of +// non-space characters in all strings. // // - Data blocks enclosed in balanced braces, which get copied verbatim and // which can span multiple lines. The opening brace of a block must be @@ -47,6 +51,9 @@ // character after each block and counts the number of blocks to set the // enumeration values. // +// For data blocks, the `maximum' template holds the maximum number of +// array elements. +// // A section can contain either strings only or data blocks only. // // A comment line starts with `//'; it gets removed. A preprocessor @@ -69,20 +76,25 @@ AF_BLUE_STRING_ENUM AF_BLUE_STRINGS_ARRAY AF_BLUE_STRING_MAX_LEN: AF_BLUE_STRING_ARABIC_TOP "ا إ ل ك ط ظ" - AF_BLUE_STRING_ARABIC_JOIN + AF_BLUE_STRING_ARABIC_BOTTOM "ت ث ط ظ ك" + // We don't necessarily have access to medial forms via Unicode in case + // Arabic presentational forms are missing. The only character that is + // guaranteed to have the same vertical position with joining (this is, + // non-isolated) forms is U+0640, ARABIC TATWEEL, which must join both + // round and flat curves. + AF_BLUE_STRING_ARABIC_JOIN + "ـ" AF_BLUE_STRING_CYRILLIC_CAPITAL_TOP - "БВЕПЗОСЭ" + "Б В Е П З О С Э" AF_BLUE_STRING_CYRILLIC_CAPITAL_BOTTOM - "БВЕШЗОСЭ" + "Б В Е Ш З О С Э" AF_BLUE_STRING_CYRILLIC_SMALL - "хпншезос" + "х п н ш е з о с" AF_BLUE_STRING_CYRILLIC_SMALL_DESCENDER - "руф" + "р у ф" - // we separate the letters with spaces to avoid ligatures; - // this is just for convenience to simplify reading AF_BLUE_STRING_DEVANAGARI_BASE "क म अ आ थ ध भ श" AF_BLUE_STRING_DEVANAGARI_TOP @@ -96,36 +108,95 @@ AF_BLUE_STRING_ENUM AF_BLUE_STRINGS_ARRAY AF_BLUE_STRING_MAX_LEN: "ु ृ" AF_BLUE_STRING_GREEK_CAPITAL_TOP - "ΓΒΕΖΘΟΩ" + "Γ Β Ε Ζ Θ Ο Ω" AF_BLUE_STRING_GREEK_CAPITAL_BOTTOM - "ΒΔΖΞΘΟ" + "Β Δ Ζ Ξ Θ Ο" AF_BLUE_STRING_GREEK_SMALL_BETA_TOP - "βθδζλξ" + "β θ δ ζ λ ξ" AF_BLUE_STRING_GREEK_SMALL - "αειοπστω" + "α ε ι ο π σ τ ω" AF_BLUE_STRING_GREEK_SMALL_DESCENDER - "βγημρφχψ" + "β γ η μ ρ φ χ ψ" AF_BLUE_STRING_HEBREW_TOP - "בדהחךכםס" + "ב ד ה ח ך כ ם ס" AF_BLUE_STRING_HEBREW_BOTTOM - "בטכםסצ" + "ב ט כ ם ס צ" AF_BLUE_STRING_HEBREW_DESCENDER - "קךןףץ" + "ק ך ן ף ץ" + + AF_BLUE_STRING_KHMER_TOP + "ខ ទ ន ឧ ឩ ា" + AF_BLUE_STRING_KHMER_SUBSCRIPT_TOP + "ក្ក ក្ខ ក្គ ក្ថ" + AF_BLUE_STRING_KHMER_BOTTOM + "ខ ឃ ច ឋ ប ម យ ឲ" + AF_BLUE_STRING_KHMER_DESCENDER + "ត្រ រៀ ឲ្យ អឿ" + AF_BLUE_STRING_KHMER_LARGE_DESCENDER + "ន្ត្រៃ ង្ខ្យ ក្បៀ ច្រៀ ន្តឿ ល្បឿ" + + AF_BLUE_STRING_KHMER_SYMBOLS_WAXING_TOP + "᧠ ᧡" + AF_BLUE_STRING_KHMER_SYMBOLS_WANING_BOTTOM + "᧶ ᧹" + + AF_BLUE_STRING_LAO_TOP + "າ ດ ອ ມ ລ ວ ຣ ງ" + AF_BLUE_STRING_LAO_BOTTOM + "າ ອ ບ ຍ ຣ ຮ ວ ຢ" + AF_BLUE_STRING_LAO_ASCENDER + "ປ ຢ ຟ ຝ" + AF_BLUE_STRING_LAO_LARGE_ASCENDER + "ໂ ໄ ໃ" + AF_BLUE_STRING_LAO_DESCENDER + "ງ ຊ ຖ ຽ ໆ ຯ" AF_BLUE_STRING_LATIN_CAPITAL_TOP - "THEZOCQS" + "T H E Z O C Q S" AF_BLUE_STRING_LATIN_CAPITAL_BOTTOM - "HEZLOCUS" + "H E Z L O C U S" AF_BLUE_STRING_LATIN_SMALL_F_TOP - "fijkdbh" + "f i j k d b h" AF_BLUE_STRING_LATIN_SMALL - "xzroesc" + "x z r o e s c" AF_BLUE_STRING_LATIN_SMALL_DESCENDER - "pqgjy" + "p q g j y" + + // we assume that both the subscript and superscript ranges + // don't contain oldstyle digits (actually, most fonts probably + // have digits only in those ranges) + AF_BLUE_STRING_LATIN_SUBS_CAPITAL_TOP + "₀ ₃ ₅ ₇ ₈" + AF_BLUE_STRING_LATIN_SUBS_CAPITAL_BOTTOM + "₀ ₁ ₂ ₃ ₈" + AF_BLUE_STRING_LATIN_SUBS_SMALL_F_TOP + "ᵢ ⱼ ₕ ₖ ₗ" + AF_BLUE_STRING_LATIN_SUBS_SMALL + "ₐ ₑ ₒ ₓ ₙ ₛ ᵥ ᵤ ᵣ" + AF_BLUE_STRING_LATIN_SUBS_SMALL_DESCENDER + "ᵦ ᵧ ᵨ ᵩ ₚ" + + AF_BLUE_STRING_LATIN_SUPS_CAPITAL_TOP + "⁰ ³ ⁵ ⁷ ᵀ ᴴ ᴱ ᴼ" + AF_BLUE_STRING_LATIN_SUPS_CAPITAL_BOTTOM + "⁰ ¹ ² ³ ᴱ ᴸ ᴼ ᵁ" + AF_BLUE_STRING_LATIN_SUPS_SMALL_F_TOP + "ᵇ ᵈ ᵏ ʰ ʲ ᶠ ⁱ" + AF_BLUE_STRING_LATIN_SUPS_SMALL + "ᵉ ᵒ ʳ ˢ ˣ ᶜ ᶻ" + AF_BLUE_STRING_LATIN_SUPS_SMALL_DESCENDER + "ᵖ ʸ ᵍ" + + AF_BLUE_STRING_MYANMAR_TOP + "ခ ဂ င ဒ ဝ ၥ ၊ ။" + AF_BLUE_STRING_MYANMAR_BOTTOM + "င ဎ ဒ ပ ဗ ဝ ၊ ။" + AF_BLUE_STRING_MYANMAR_ASCENDER + "ဩ ြ ၍ ၏ ၆ ါ ိ" + AF_BLUE_STRING_MYANMAR_DESCENDER + "ဉ ည ဥ ဩ ဨ ၂ ၅ ၉" - // we separate the letters with spaces to avoid ligatures; - // this is just for convenience to simplify reading AF_BLUE_STRING_TELUGU_TOP "ఇ ఌ ఙ ఞ ణ ఱ ౯" AF_BLUE_STRING_TELUGU_BOTTOM @@ -150,44 +221,44 @@ AF_BLUE_STRING_ENUM AF_BLUE_STRINGS_ARRAY AF_BLUE_STRING_MAX_LEN: #ifdef AF_CONFIG_OPTION_CJK AF_BLUE_STRING_CJK_TOP - "他们你來們到和地" - "对對就席我时時會" - "来為能舰說说这這" - "齊 |" - "军同已愿既星是景" - "民照现現理用置要" - "軍那配里開雷露面" - "顾" + "他 们 你 來 們 到 和 地" + " 对 對 就 席 我 时 時 會" + " 来 為 能 舰 說 说 这 這" + " 齊 |" + " 军 同 已 愿 既 星 是 景" + " 民 照 现 現 理 用 置 要" + " 軍 那 配 里 開 雷 露 面" + " 顾" AF_BLUE_STRING_CJK_BOTTOM - "个为人他以们你來" - "個們到和大对對就" - "我时時有来為要說" - "说 |" - "主些因它想意理生" - "當看着置者自著裡" - "过还进進過道還里" - "面" + "个 为 人 他 以 们 你 來" + " 個 們 到 和 大 对 對 就" + " 我 时 時 有 来 為 要 說" + " 说 |" + " 主 些 因 它 想 意 理 生" + " 當 看 着 置 者 自 著 裡" + " 过 还 进 進 過 道 還 里" + " 面" #ifdef AF_CONFIG_OPTION_CJK_BLUE_HANI_VERT AF_BLUE_STRING_CJK_LEFT - "些们你來們到和地" - "她将將就年得情最" - "样樣理能說说这這" - "通 |" - "即吗吧听呢品响嗎" - "师師收断斷明眼間" - "间际陈限除陳随際" - "隨" + " 些 们 你 來 們 到 和 地" + " 她 将 將 就 年 得 情 最" + " 样 樣 理 能 說 说 这 這" + " 通 |" + " 即 吗 吧 听 呢 品 响 嗎" + " 师 師 收 断 斷 明 眼 間" + " 间 际 陈 限 除 陳 随 際" + " 隨" AF_BLUE_STRING_CJK_RIGHT - "事前學将將情想或" - "政斯新样樣民沒没" - "然特现現球第經谁" - "起 |" - "例別别制动動吗嗎" - "增指明朝期构物确" - "种調调費费那都間" - "间" + "事 前 學 将 將 情 想 或" + " 政 斯 新 样 樣 民 沒 没" + " 然 特 现 現 球 第 經 谁" + " 起 |" + " 例 別 别 制 动 動 吗 嗎" + " 增 指 明 朝 期 构 物 确" + " 种 調 调 費 费 那 都 間" + " 间" #endif /* AF_CONFIG_OPTION_CJK_BLUE_HANI_VERT */ @@ -219,6 +290,14 @@ AF_BLUE_STRING_ENUM AF_BLUE_STRINGS_ARRAY AF_BLUE_STRING_MAX_LEN: // respectively. Only horizontal blue zones (i.e., adjusting vertical // coordinate values) are supported. // +// Some scripts like Khmer need character composition to get all necessary +// blue zones, since Unicode only provides an abstract data model that +// doesn't represent all possible glyph shapes. For such character +// clusters, the HarfBuzz library is used to convert them into the +// corresponding glyphs. The largest glyph element (where `largest' can be +// either `largest ascender' or `largest descender') then defines the +// corresponding flat or round extremum. +// // For the latin auto-hinter, the overshoot should be larger than the // reference for top zones, and vice versa for bottom zones. // @@ -229,6 +308,21 @@ AF_BLUE_STRING_ENUM AF_BLUE_STRINGS_ARRAY AF_BLUE_STRING_MAX_LEN: // // If not set, take the minimum values. // +// Mutually exclusive with `LATIN_SUB_TOP'. +// +// LATIN_SUB_TOP +// For all glyphs of a character cluster, compute the maximum flat +// and round coordinate values of each component, then take the +// smallest of the maximum values. The idea is to get the top of +// subscript glyphs, as used in Khmer, for example. Note that +// this mechanism doesn't work for ordinary ligatures. +// +// This flags indicates a secondary blue zone: It gets removed if +// there is a non-LATIN_SUB_TOP blue zone at the same coordinate +// value (after scaling). +// +// Mutually exclusive with `LATIN_TOP'. +// // LATIN_NEUTRAL // Ignore round extrema and define the blue zone with flat values only. // Both top and bottom of contours can match. This is useful for @@ -264,11 +358,11 @@ AF_BLUE_STRING_ENUM AF_BLUE_STRINGS_ARRAY AF_BLUE_STRING_MAX_LEN: // // Characters in a blue string are *not* automatically classified. Instead, // first come the characters used for the overshoot value, then the -// character `|', then the characters used for the reference value. The -// blue zone is then set up by the mean values of all reference values and -// all overshoot values, respectively. Both horizontal and vertical blue -// zones (i.e., adjusting vertical and horizontal coordinate values, -// respectively) are supported. +// character `|', then the characters used for the reference value +// (everything separated by space characters). The blue zone is then set up +// by the mean values of all reference values and all overshoot values, +// respectively. Both horizontal and vertical blue zones (i.e., adjusting +// vertical and horizontal coordinate values, respectively) are supported. // // For the cjk auto-hinter, the overshoot should be smaller than the // reference for top zones, and vice versa for bottom zones. @@ -289,9 +383,10 @@ AF_BLUE_STRING_ENUM AF_BLUE_STRINGS_ARRAY AF_BLUE_STRING_MAX_LEN: AF_BLUE_STRINGSET_ENUM AF_BLUE_STRINGSETS_ARRAY AF_BLUE_STRINGSET_MAX_LEN: AF_BLUE_STRINGSET_ARAB - { AF_BLUE_STRING_ARABIC_TOP, AF_BLUE_PROPERTY_LATIN_TOP } - { AF_BLUE_STRING_ARABIC_JOIN, 0 } - { AF_BLUE_STRING_MAX, 0 } + { AF_BLUE_STRING_ARABIC_TOP, AF_BLUE_PROPERTY_LATIN_TOP } + { AF_BLUE_STRING_ARABIC_BOTTOM, 0 } + { AF_BLUE_STRING_ARABIC_JOIN, AF_BLUE_PROPERTY_LATIN_NEUTRAL } + { AF_BLUE_STRING_MAX, 0 } AF_BLUE_STRINGSET_CYRL { AF_BLUE_STRING_CYRILLIC_CAPITAL_TOP, AF_BLUE_PROPERTY_LATIN_TOP } @@ -329,6 +424,30 @@ AF_BLUE_STRINGSET_ENUM AF_BLUE_STRINGSETS_ARRAY AF_BLUE_STRINGSET_MAX_LEN: { AF_BLUE_STRING_HEBREW_DESCENDER, 0 } { AF_BLUE_STRING_MAX, 0 } + AF_BLUE_STRINGSET_KHMR + { AF_BLUE_STRING_KHMER_TOP, AF_BLUE_PROPERTY_LATIN_TOP | + AF_BLUE_PROPERTY_LATIN_X_HEIGHT } + { AF_BLUE_STRING_KHMER_SUBSCRIPT_TOP, AF_BLUE_PROPERTY_LATIN_SUB_TOP } + { AF_BLUE_STRING_KHMER_BOTTOM, 0 } + { AF_BLUE_STRING_KHMER_DESCENDER, 0 } + { AF_BLUE_STRING_KHMER_LARGE_DESCENDER, 0 } + { AF_BLUE_STRING_MAX, 0 } + + AF_BLUE_STRINGSET_KHMS + { AF_BLUE_STRING_KHMER_SYMBOLS_WAXING_TOP, AF_BLUE_PROPERTY_LATIN_TOP | + AF_BLUE_PROPERTY_LATIN_X_HEIGHT } + { AF_BLUE_STRING_KHMER_SYMBOLS_WANING_BOTTOM, 0 } + { AF_BLUE_STRING_MAX, 0 } + + AF_BLUE_STRINGSET_LAO + { AF_BLUE_STRING_LAO_TOP, AF_BLUE_PROPERTY_LATIN_TOP | + AF_BLUE_PROPERTY_LATIN_X_HEIGHT } + { AF_BLUE_STRING_LAO_BOTTOM, 0 } + { AF_BLUE_STRING_LAO_ASCENDER, AF_BLUE_PROPERTY_LATIN_TOP } + { AF_BLUE_STRING_LAO_LARGE_ASCENDER, AF_BLUE_PROPERTY_LATIN_TOP } + { AF_BLUE_STRING_LAO_DESCENDER, 0 } + { AF_BLUE_STRING_MAX, 0 } + AF_BLUE_STRINGSET_LATN { AF_BLUE_STRING_LATIN_CAPITAL_TOP, AF_BLUE_PROPERTY_LATIN_TOP } { AF_BLUE_STRING_LATIN_CAPITAL_BOTTOM, 0 } @@ -339,6 +458,34 @@ AF_BLUE_STRINGSET_ENUM AF_BLUE_STRINGSETS_ARRAY AF_BLUE_STRINGSET_MAX_LEN: { AF_BLUE_STRING_LATIN_SMALL_DESCENDER, 0 } { AF_BLUE_STRING_MAX, 0 } + AF_BLUE_STRINGSET_LATB + { AF_BLUE_STRING_LATIN_SUBS_CAPITAL_TOP, AF_BLUE_PROPERTY_LATIN_TOP } + { AF_BLUE_STRING_LATIN_SUBS_CAPITAL_BOTTOM, 0 } + { AF_BLUE_STRING_LATIN_SUBS_SMALL_F_TOP, AF_BLUE_PROPERTY_LATIN_TOP } + { AF_BLUE_STRING_LATIN_SUBS_SMALL, AF_BLUE_PROPERTY_LATIN_TOP | + AF_BLUE_PROPERTY_LATIN_X_HEIGHT } + { AF_BLUE_STRING_LATIN_SUBS_SMALL, 0 } + { AF_BLUE_STRING_LATIN_SUBS_SMALL_DESCENDER, 0 } + { AF_BLUE_STRING_MAX, 0 } + + AF_BLUE_STRINGSET_LATP + { AF_BLUE_STRING_LATIN_SUPS_CAPITAL_TOP, AF_BLUE_PROPERTY_LATIN_TOP } + { AF_BLUE_STRING_LATIN_SUPS_CAPITAL_BOTTOM, 0 } + { AF_BLUE_STRING_LATIN_SUPS_SMALL_F_TOP, AF_BLUE_PROPERTY_LATIN_TOP } + { AF_BLUE_STRING_LATIN_SUPS_SMALL, AF_BLUE_PROPERTY_LATIN_TOP | + AF_BLUE_PROPERTY_LATIN_X_HEIGHT } + { AF_BLUE_STRING_LATIN_SUPS_SMALL, 0 } + { AF_BLUE_STRING_LATIN_SUPS_SMALL_DESCENDER, 0 } + { AF_BLUE_STRING_MAX, 0 } + + AF_BLUE_STRINGSET_MYMR + { AF_BLUE_STRING_MYANMAR_TOP, AF_BLUE_PROPERTY_LATIN_TOP | + AF_BLUE_PROPERTY_LATIN_X_HEIGHT } + { AF_BLUE_STRING_MYANMAR_BOTTOM, 0 } + { AF_BLUE_STRING_MYANMAR_ASCENDER, AF_BLUE_PROPERTY_LATIN_TOP } + { AF_BLUE_STRING_MYANMAR_DESCENDER, 0 } + { AF_BLUE_STRING_MAX, 0 } + AF_BLUE_STRINGSET_TELU { AF_BLUE_STRING_TELUGU_TOP, AF_BLUE_PROPERTY_LATIN_TOP } { AF_BLUE_STRING_TELUGU_BOTTOM, 0 } |