unicode (or just plain byte) regular expression positions?

[Tim Meehan]

Say that I have a strange character group that I want to find in a binary file.

I wanted to use something like this:

(define needle (list->string (map integer->char (list #xab #xcd #xef))))

(define needle-offset

  (call-with-input-file "big_binary_blob.bin"

    #:mode 'binary

    (λ (p)

      (regexp-match-positions (regexp needle) p))))

The "regexp-match-positions" returns #f (even though I know that needle is in there, I put it there). Is there a better way to go about this? The binary blob is about 100 MiB or so, if that helps.

[Jon Zeppieri]

You're searching for a certain unicode codepoint sequence (U+00AB,
U+00CD, U+00EF) in a string, but I think you're trying to search for a
byte sequence in a byte string. You can read in the file as bytes and
use a byte regexp. So:

(define needle (list->bytes (list #xab #xcd #xef)))

(define needle-offset
(call-with-input-file "big_binary_blob.bin"
#:mode 'binary
(λ (p)

(regexp-match-positions (byte-regexp needle) p))))