Merge branch 'main' of https://git.matthewbutterick.com/mbutterick/typesetting

Intercapped (aka camelCase) words are treated like a list of subwords joined by hyphens, but only if the pieces meet the usual min-length threshold. But don’t use this splitting technique when unhyphenating.

hyphenate/hyphenate/private/core.rkt

@@ -162,14 +162,32 @@
 (define (joiner->string joiner) (format "~a" joiner))
 
 
-(define (apply-proc proc x [omit-string (λ (x) #f)] [omit-txexpr (λ (x) #f)] [joiner default-joiner])
+(define (apply-proc proc x
+                    [omit-string (λ (x) #f)]
+                    [omit-txexpr (λ (x) #f)]
+                    [joiner default-joiner]
+                    #:intercap-min-length [intercap-min-length #false])
   (let loop ([x x])
     (cond
       [(and (string? x) (not (omit-string x)))
-       ;; handle intercapped words as capitalized pieces
-       (define letter-before-uc #px"(?<=\\p{Ll})(?=\\p{Lu}\\p{Ll})") ; match xXx but not xXX or XXX
-       (string-join (for/list ([x (in-list (string-split x letter-before-uc))])
-                      (proc x)) (joiner->string joiner))]
+       (define words
+         (cond
+           [intercap-min-length
+            ;; handle intercapped words as a list of subwords,
+            ;; subject to the intercap-min-length
+            (define zero-length-quantifier "")
+            (define letter-before-uc
+              ;; match xXx but not xXX or XXX
+              (pregexp (format "(?<=\\p{L}{~a})(?=\\p{Lu}\\p{Ll}{~a})"
+                               (if (> intercap-min-length 0)
+                                   intercap-min-length
+                                   zero-length-quantifier)
+                               (if (> intercap-min-length 1)
+                                   (sub1 intercap-min-length)
+                                   zero-length-quantifier))))
+            (string-split x letter-before-uc)]
+           [else (list x)]))
+       (string-join (map proc words) (joiner->string joiner))]
       [(and (txexpr? x) (not (omit-txexpr x))) 
        (make-txexpr (get-tag x) (get-attrs x) (map loop (get-elements x)))]
       [else x])))
@@ -199,7 +217,8 @@
               [else word])]))
   (define (insert-hyphens text) (regexp-replace* word-pattern text replacer))
   (begin0
-    (apply-proc insert-hyphens x omit-string? omit-txexpr? joiner)
+    (apply-proc insert-hyphens x omit-string? omit-txexpr? joiner
+                #:intercap-min-length min-length)
     ;; deleting from the main cache is cheaper than having to do two cache lookups for every word
     ;; (missing words will just be regenerated later)
     (for-each (λ (ee) (remove-exception-word word-cache ee)) extra-exceptions)))

hyphenate/hyphenate/tests.rkt

@@ -22,7 +22,9 @@
 (check-equal? (unhyphenate "poly\u00ADmor\u00ADphism") "polymorphism")
 (check-equal? (hyphenate "polymorphism" #\-) "poly-mor-phism")
 (check-equal? (hyphenate "compotumi" #\-) "com-po-tu-mi")
-(check-equal? (hyphenate "CompOtumi" #\-) "Comp-Otu-mi")
+(check-equal? (hyphenate "CompOtumi" #\-) "Com-pO-tu-mi")
+(check-equal? (hyphenate "iMagnificence" #\-) "iMag-nif-i-cence")
+(check-equal? (hyphenate "CompOtumi" #:min-length 4 #\-) "Comp-Otu-mi")
 (check-equal? (hyphenate "polymorphism" "foo") "polyfoomorfoophism")
 (check-equal? (unhyphenate "polyfoomorfoophism" "foo") "polymorphism")
 (check-equal? (hyphenate "circular polymorphism squandering") "cir\u00ADcu\u00ADlar poly\u00ADmor\u00ADphism squan\u00ADder\u00ADing")

hyphenate/hyphenate/us/exceptions.rkt

@@ -4,6 +4,12 @@ as-so-ciate
 as-so-ciates
 dec-li-na-tion
 dis-tri-bu-tion
+iMac
+iMacs
+iPad
+iPads
+iPhone
+iPhones
 law-yer
 law-yer-ing
 law-yer-ly
@@ -24,3 +30,4 @@ ref-or-ma-tion
 ret-ri-bu-tion
 ta-ble
 ty-pog-ra-pher
+ty-pog-ra-phers