Re: [clisp-list] strings and byte arrays (was set-dispatch-macro-character and recursive read weirdness)

[Sam S. <sd...@po...>]

thanks for your bug report.
please file a new self-contained bug report with a reproducible test 
case at http://sourceforge.net/tracker/?func=add&group_id=1355&atid=101355
you might also want to look at
http://clisp.cons.org/impnotes/faq.html#faq-help
http://clisp.cons.org/impnotes/clisp.html#bugs

root wrote:
> I seem to have run into a bug in clisp.
> 
> I have a function that converts an string to a byte array.
> I hand it a string of hex characters and these are converted
> into byte hex strings.
> 
> (defun string-to-byte-array (bytestring)
>  (let (stack result)
>   (do ((p 0))
>       ((>= p (length bytestring)) nil)
>    (multiple-value-setq (m n)
>     (parse-integer bytestring :start p :radix 16 :junk-allowed t))
>    (when m (push m stack))
>    (setf p n))
>   (setq result 
>    (make-array (length stack) :element-type 'unsigned-byte
>                               :initial-contents (nreverse stack)))
>   (format t "string-to-byte-array ~a ~a ~a~%" (type-of result)
>    (length bytestring) (length result))
>   result))
> 
> This works fine for small strings. However I have a string that
> is roughly 92000 bytes it creates a simple-vector of roughly 30000 bytes.
> 
> The actual output of the format function is:
>   
>   string-to-byte-array (SIMPLE-VECTOR 30720) 92159 30720
> 
> I cannot use either svref or aref on the result because CLISP
> complains that the result is not a SIMPLE-VECTOR.
> 
> t