Fwd: [trivial-utf-8-devel] error on sbcl

[Marijn Haverbeke]

Hi Anton,

The problem seems to be that I'm using too high an optimisation
setting, which causes SBCL to go ahead and assume you are passing in a
vector of unsigned bytes, while you are not.
(trivial-utf-8:utf-8-bytes-to-string (coerce *body* '(unsigned-byte
8))) does work as it should. I'll see if I can make it check the type
of its argument.

Anyway, I've updated the library, if you re-download it your first
example should work. (Hm, maybe I should start doing versions.)


Best,
Marijn



On Mon, Jul 14, 2008 at 3:41 PM, Anton Vodonosov <avodonosov at yandex.ru> wrote:
> Hello.
>
> The following code when used with SBCL signals an error:
>
>  (defparameter *body*
>    #(110 115 58 104 116 116 112 58 47 47
>      115 112 101 99 115 46 111 112 101 110
>      105 100 46 110 101 116 47 97 117 116
>      104 47 50 46 48))
>
>  (trivial-utf-8:utf-8-bytes-to-string *body*)
>
> Invalid byte at start of character: 0xB8
>   [Condition of type UTF-8-DECODING-ERROR]
>
> Backtrace:
>  0: (TRIVIAL-UTF-8::UTF-8-STRING-LENGTH #<unavailable argument>)
>  1: (UTF-8-BYTES-TO-STRING #<unavailable argument>)
>  ...
>
> With CLISP all is ok, the string is decoded as
> ns:http://specs.openid.net/auth/2.0
>
> Best regards,
> -Anton
>
> _______________________________________________
> trivial-utf-8-devel mailing list
> trivial-utf-8-devel at common-lisp.net
> http://common-lisp.net/cgi-bin/mailman/listinfo/trivial-utf-8-devel
>