(font (nil nil unicode-bmp :otf=knda=rphf)))
(category
- ;; C: general consonant
+ ;; C: consonant (except for RA)
;; R: RA
+ ;; n: NUKTA
;; H: HALANT
- ;; u: MATRA (above)
- ;; b: MATRA (below)
- ;; p: MATRA (post)
- ;; t: MATRA (two-part)
- ;; T: MATRA (three-part)
+ ;; u: vowel sign (above)
+ ;; b: vowel sign (below)
+ ;; p: vowel sign (post)
+ ;; l: length mark
;; a: vowel modifier (post)
;; V: independent vowel
;; N: ZWNJ (ZERO WIDTH NON-JOINER)
;; J: ZWJ (ZERO WIDTH JOINER)
- ;; E: ELSE
- ;;
- (0x200C ?N) ; ZWNJ
- (0x200D ?J) ; ZWJ
- (0x0964 ?E) ; DANDA
- (0x0965 ?E) ; DOUBLE DANDA
- (0x0C80 0x0CFF ?E) ; ELSE
- (0x0C82 0x0C83 ?a) ; SIGN ANUSVARA, VISARGA
+ ;; X: generic
+ (0x0C80 0x0CFF ?X) ; generic
+ (0x0C82 0x0C83 ?a) ; SIGN ANUSVARA .. VISARGA
(0x0C85 0x0C94 ?V) ; LETTER A .. LETTER AU
(0x0C95 0x0CB9 ?C) ; LETTER KA .. LETTER HA
(0x0CB0 ?R) ; LETTER RA
(0x0CBC ?n) ; SIGN NUKTA
- (0x0CBE ?p) ; VOWEL SIGN AA (post)
- (0x0CBF ?u) ; VOWEL SIGN I (above)
- (0x0CC0 ?t) ; VOWEL SIGN II (two-part)
- (0x0CC1 0x0CC4 ?p) ; VOWEL SIGN U, UU, R, RR (post)
- (0x0CC6 ?u) ; VOWEL SIGN E (above)
- (0x0CC7 0x0CCA ?t) ; VOWEL SIGN EE, AI, O (two-part)
- (0x0CCB ?T) ; VOWEL SIGN OO (three-part)
- (0x0CCC ?u) ; VOWEL SIGN AU (above)
- (0x0CCD ?H) ; SIGN VIRAMA (HALANT)
- (0x0CD5 0x0CD6 ?l) ; LENGTH MARK (length post)
+ (0x0CBE ?p) ; VOWEL SIGN AA
+ (0x0CBF ?u) ; VOWEL SIGN I
+ (0x0CC1 0x0CC4 ?p) ; VOWEL SIGN U .. RR
+ (0x0CC6 ?u) ; VOWEL SIGN E
+ (0x0CCC ?u) ; VOWEL SIGN AU
+ (0x0CCD ?H) ; SIGN VIRAMA
+ (0x0CD5 0x0CD6 ?l) ; LENGTH MARK .. AI LENGTH MARK
(0x0CDE ?C) ; LETTER FA (LLLA)
- (0x0CE0 0x0CE1 ?V) ; LETTER VOCALIC RR, LL
- (0x0CE2 0x0CE3 ?b) ; VOWEL SIGN VOCALIC L, LL
- (0x0CFD ?Z) ; internal use
- (0x0CFE ?Y) ; internal use
- (0x0CFF ?X) ; internal use
+ (0x0CE0 0x0CE1 ?V) ; LETTER VOCALIC RR .. LL
+ (0x0CE2 0x0CE3 ?b) ; VOWEL SIGN VOCALIC L .. LL
+ (0x0964 0x0965 ?X) ; DANDA .. DOUBLE DANDA
+ (0x200C ?N) ; ZWNJ
+ (0x200D ?J) ; ZWJ
)
-;; Step 1 : Syllable identification and Halant movement.
+;; Preprocessing
(generator
(0
(cond
- ;; A syllable containing a three-part vowel sign.
- ("([CR]J?n?(H[CR]n?)*)(T)(a)?"
- < | (1 move-base-Halant) (3 0x0CC6 0x0CC2 0x0CD5) (4 =) | >)
-
- ;; A syllable containing a two-part vowel sign.
- ("([CR]J?n?(H[CR]n?)*)(t)(a)?"
- < | (1 move-base-Halant) (3 two-part) (4 =) | >)
-
- ;; A syllable ending with vowel signs and/or a vowel modifier.
- ("([CR]J?n?(H[CR]n?)*)([ubpl]+a?|a)"
- < | (1 move-base-Halant) (3 = *) | >)
+ ;; Decompose two-part and three-part vowels
+ ((0x0CC0)
+ 0x0CBF 0x0CD5)
+ ((0x0CC7)
+ 0x0CC6 0x0CD5)
+ ((0x0CC8)
+ 0x0CC6 0x0CD6)
+ ((0x0CCA)
+ 0x0CC6 0x0CC2)
+ ((0x0CCB)
+ 0x0CC6 0x0CC2 0x0CD5)
+ ;; Move ZWJ between RA and HARANT.
+ ("(R)(J)(H)"
+ (2 =) (1 =) (3 =))
+ ("." =))
+ *))
- ;; Forced Halant form.
- ("([CR]n?H)N"
- < | (1 move-base-Halant) | >)
+;; Syllable identification.
+;; Apply 'nukt' and 'akhn' before reordering.
- ;; A syllable without a vowel sign nor a vowel modifier.
- ("[CR]J?n?(H[CR]n?)*H?"
- < | move-base-Halant | >)
+(generator
+ (0
+ (cond
+ ;; A syllable with an above vowel sign.
+ ;;1 2 3 4
+ ("(J)?([CR]n?(H[CR]n?)*)(up?l?a?)"
+ < | (1 =) (2 otf:knda=nukt,akhn+) (4 = *) | >)
+
+ ;; A syllable with a post- or below vowel sign.
+ ;;1 2 3 4 5
+ ("(J)?([CR]n?(H[CR]n?)*)(p|b)(a)?"
+ < | (1 =) (2 otf:knda=nukt,akhn+) (4 =) (5 =) | >)
+
+ ;; A syllable with a vowel modifier, but without vowel signs.
+ ;;1 2 3 4
+ ("(J)?([CR]n?(H[CR]n?)*)(a)"
+ < | (1 =) (2 otf:knda=nukt,akhn+) (4 =) | >)
+
+ ;; No vowel signs, no vowel modifiers.
+ ;;1 2 3 4
+ ("(J)?([CR]n?(H[CR]n?)*)(H|HN)?"
+ < | (1 =) (2 otf:knda=nukt,akhn+) (4 = *) | >)
;; A syllable starting with an independent vowel.
("Va?"
- < | = * | > )
+ < | = * | >)
("." =))
- *)
-
- ;; Move Halant on the base consonant to the tail if the syllable ends
- ;; with a consonant.
- ;; A leading 0x0CFF means Halant movement.
- ;; A leading 0x0CFE means a syllable ending with a Halant.
- ;; A leading 0x0CFD means Halant & Reph movements.
- (move-base-Halant
- (cond
- (".*H$"
- 0x0CFE = *)
- ("(RH)([CR]n?)$"
- 0x0CFD (2 = *) (1 = =))
- ("(RH)([CR]n?)(H)(.*)"
- 0x0CFD (2 = *) (4 = *) (3 =) (1 = =))
- ("(R)J(H)(.*)"
- 0x0CFF (1 =) (3 = *) (2 =))
- ("(Cn?)(H)(.*)"
- 0x0CFF (1 = *) (3 = *) (2 =))
- (".*"
- = *)))
-
- ;; Divide a two-part Matra into elements.
- (two-part
- (cond
- ((0x0CC0) 0x0CBF 0x0CD5)
- ((0x0CC7) 0x0CC6 0x0CD5)
- ((0x0CC8) 0x0CC6 0x0CD6)
- ((0x0CCA) 0x0CC6 0x0CC2)))
- )
+ *))
-;; Step 2 : Move Matra if applicable.
-;; The base consonant in a Halant-ending syllable is changed into Halant form.
+;; Reordering. The base consonant is always the first one, because
+;; all Kannada consonants have a below form.
(generator
(0
(cond
- ;; 1 2 3 4 5 6 7 8
- (" Z([CR]n?)([CRnH]*)(RH)(u)?(b)?(p)?(l*)(a)? "
- | (1 = *) (4 =) (5 =) (6 =) (2 = *) (7 = *) (3 otf:knda=rphf) (8 =) |)
- ;; 1 2 3 4 5 6 7
- (" X([CR]n?)([CRnH]*)(u)?(b)?(p)?(l*)(a)? "
- | (1 = *) (3 =) (4 =) (5 =) (2 = *) (6 = *) (7 =) |)
- (" Y([CR]n?H)([^ ]*) "
- ;; Kedage-n.ttf lacks the haln feature, so "otf:knda=haln" does not work.
- | (1 otf:knda=~rphf,~blwf,*) (2 = *) |)
+ ;; 1 2 34 5 6 7 8
+ (" (RH)?(.)((H.)*)(u|b)?(p)?(l)?(a)? "
+ | (2 =) (5 =) (6 =) (3 move-h) (7 =) (1 otf:knda=rphf+) (8 =) |)
+
+ ;; A syllable ending with a halant.
+ ;; 1 2 34
+ (" (RH)?(.H)((.H)*)N? "
+ ;; WORKAROUND : Kedage fonts produce halant form with 'psts'.
+ | (2 otf:knda=haln,psts+)(3 = *) (1 otf:knda=rphf+) |)
+
+ ;; No reph movement if preceded by a ZWJ.
+ ;; 1 23 4 5 6
+ (" J(.)((H.)*)(u|b)?(p)?(l?a?) "
+ | (1 =) (4 =) (5 =) (2 move-h) (6 = *) |)
+
+ ;; 1 23
+ (" J(.H)((.H)*)N? "
+ | (1 otf:knda=haln,psts+) (2 = *) |)
+
("." =))
- *))
+ *)
-;; Step 3 : Drive OTF tables.
+ (move-h
+ (cond
+ ("(H)(.+)"
+ (2 = *) (1 =)))))
+
+;; Apply other OTF features.
(generator
(0
(cond
- (" ([^ ]*) "
- (1 otf:knda=~rphf,*))
+ (" ([^ ]+) "
+ (1 otf:knda=blwf,abvs,blws,psts))
+
("."
[ otf:knda=+ ]))
*))
projects / m17n / m17n-db.git / commitdiff