aboutsummaryrefslogtreecommitdiff
path: root/gnu
diff options
context:
space:
mode:
authorDanny Milosavljevic <dannym@scratchpost.org>2020-05-03 21:40:04 +0200
committerDanny Milosavljevic <dannym@scratchpost.org>2020-05-03 23:04:57 +0200
commitbb357c509e1c017e1fef5aa5f4d05beea0c25157 (patch)
treed4c20addb90b3d58a907644e874c4cd7b2086273 /gnu
parent96e399ee30251a0767a45b2bde3f052b25c62714 (diff)
downloadguix-bb357c509e1c017e1fef5aa5f4d05beea0c25157.tar.gz
guix-bb357c509e1c017e1fef5aa5f4d05beea0c25157.zip
file-systems: Fix UTF-16 handling in initrd.
Follow-up to f73f4b3a2d7a313a6cb1667bd69205ea4b09f57c. * gnu/build/file-systems.scm (bytevector->u16-list): New procedure. (utf16->string): New procedure.
Diffstat (limited to 'gnu')
-rw-r--r--gnu/build/file-systems.scm25
1 files changed, 25 insertions, 0 deletions
diff --git a/gnu/build/file-systems.scm b/gnu/build/file-systems.scm
index 4ac672d96f..b920e8fc62 100644
--- a/gnu/build/file-systems.scm
+++ b/gnu/build/file-systems.scm
@@ -110,6 +110,31 @@ NUL terminator, return the size of the bytevector."
(loop (+ index 2)))
length))))
+(define* (bytevector->u16-list bv endianness #:optional (index 0))
+ (if (< index (bytevector-length bv))
+ (cons (bytevector-u16-ref bv index endianness)
+ (bytevector->u16-list bv endianness (+ index 2)))
+ '()))
+
+;; The initrd doesn't have iconv data, so do the conversion ourselves.
+(define (utf16->string bv endianness)
+ (list->string
+ (map integer->char
+ (reverse
+ (let loop ((remainder (bytevector->u16-list bv endianness))
+ (result '()))
+ (match remainder
+ (() result)
+ ((a) (cons a result))
+ ((a b x ...)
+ (if (and (>= a #xD800) (< a #xDC00) ; high surrogate
+ (>= b #xDC00) (< b #xE000)) ; low surrogate
+ (loop x (cons (+ #x10000
+ (* #x400 (- a #xD800))
+ (- b #xDC00))
+ result))
+ (loop (cons b x) (cons a result))))))))))
+
(define (null-terminated-utf16->string bv endianness)
(utf16->string (sub-bytevector bv 0 (bytevector-utf16-length bv))
endianness))