Custom Query (1030 matches)

(/ (complex 1d160 1)) overflows

Answer should be #C(1.0d-160 -9.99988867182683d-321)

Hi,

When the a package is present, say, HUNCHENTOOT, the following works

(defpackage :foo

(:use :cl :hunchentoot))

(in-package :foo)

..

However, as start to use more packages during the course of hacking, I sometimes :USE non-existent package or misspell its name.

(defpackage :foo

(:use :cl :hunchentoot :frob))

(in-package :foo)

..

On Linux, I get the correct behaviour:

The name "FROB" does not designate any package.

[Condition of type SB-KERNEL:SIMPLE-PACKAGE-ERROR]

But on Win32, things get unwieldy: I first get the error, but if I try to correct it by fixing the spelling or loading an ASDF system, all forms refuse to evaluate. For example, after fixing the typo, I get this:

Undefined function :USE called with arguments (:CL

:CL-WHO :POSTMODERN :SIMPLE-DATE :HUNCHENTOOT) .

[Condition of type CCL::UNDEFINED-FUNCTION-CALL]

Even evaluating () returns ==> COMMON-LISP:NIL

The only way I know how to repair things without restarting CCL is to do this:

FOO> (cl-user::in-package :cl-user) #<Package "COMMON-LISP-USER"> CL-USER> (in-package :foo) #<Package "FOO">

Attached is a stack trace.

CL-USER> (lisp-implementation-version) "Version 1.4 (WindowsX8632)"

Welcome to Clozure Common Lisp Version 1.6-dev-r13678M  (DarwinX8664)!
? (defparameter *s* (make-socket :type :datagram))
*S*
?  (send-to *s* "hey" 3 :remote-host "127.0.0.1" :remote-port 12345)
> Error: value "hey" is not of the expected type (OR (ARRAY CHARACTER)
>                                                    (ARRAY (UNSIGNED-BYTE 8))
>                                                    (ARRAY (SIGNED-BYTE 8))).
> While executing: VERIFY-SOCKET-BUFFER, in process listener(1).
> Type :POP to abort, :R for a list of available restarts.
> Type :? for other options.
1 > 

verify-socket-buffer is actually checking for an 8-bit subtag, and with this new-fangled unicode thing, that doesn't include strings.

My first inclination is to say that i/o on datagram sockets is done in octects. This requires the user to use encode-string-to-octects/decode-string-from-octets explicitly.

It might also be work to pay attention to the socket object's external format and encode the string on the user's behalf, but if Gilgamesh wants to send a cuneiform string "foo", then the actual buffer length in octets isn't going to be 3.