Re: CVS: sbcl-page platform-support-platforms.lisp, 1.39, 1.40

<jsnell <at> users.sourceforge.net> wrote:
> Modified Files:
> 	platform-support-platforms.lisp 
> Log Message:
> Roll back the broken x86/Linux, x86-64/Linux, x86/Darwin and ppc/Darwin
> binaries from 0.9.14 to earlier ones. Please upload new versions of them
> done using a binary-distribution.sh from HEAD instead of the one included
> with 0.9.14. (The one included with 0.9.14 does not include sbcl-pwd.sh)

Bringing this up on sbcl-devel, since trying to communicating this issue
just via sbcl-commits and #lisp was apparenly a complete failure 

--

-- 
Juho Snellman

Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

applying patches

What stage of applying patches for UCS2 from Ivan Boldyrev and my old 
patch for non-ascii c-string?
What to do and where to run? :)
Publish the final version for the current sbcl snapshot once again?

I am tired to wait. :)

Thanks!

--

-- 
WBR, Yaroslav Kavenchuk.

Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

Re: applying patches

On 5. Jul 2006, at 11:50, Yaroslav Kavenchuk wrote:

> What stage of applying patches for UCS2 from Ivan Boldyrev and my old
> patch for non-ascii c-string?
> Publish the final version for the current sbcl snapshot once again?

I am compiling an ucs2 tree right now.  I didn't want to do this  
earlier because it takes a while on my laptop and I had no power on  
the various airports I inhabited today.  Once ucs2 is in the tree,  
you might want to resend your patch.

Cheers,

Rudi

Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

_______________________________________________
Sbcl-devel mailing list
Sbcl-devel <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sbcl-devel

Re: [PATCH] UCS-2 and UTF-16

On 29. Jun 2006, at 13:27, Ivan Boldyrev wrote:

> Oops!  Sorry, this patch have only UCS-2.  UTF-16 sources are lost :(
> I will write it again.
>
> Yaroslav also noticed some glitch in processing :NULL-TERMINATE
> keyword of STRING-TO-OCTETS: the byte array must be terminated by two
> zero bytes.  See patch below.

Many thanks, committed as 0.9.14.8.

Cheers,

Rudi

Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

_______________________________________________
Sbcl-devel mailing list
Sbcl-devel <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sbcl-devel

Re: applying patches

Rudi Schlatte wrote:
> I am compiling an ucs2 tree right now.  I didn't want to do this  
> earlier because it takes a while on my laptop and I had no power on  
> the various airports I inhabited today.  Once ucs2 is in the tree,  
> you might want to resend your patch.

Many thanks and excuse me: global reconstruction after applying ucs2 
patch... :/

Thanks!

--

-- 
WBR, Yaroslav Kavenchuk.

Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

Re: [PATCH] UCS-2 and UTF-16

On 9527 day of my life Rudi Schlatte wrote:
>> Oops!  Sorry, this patch have only UCS-2.  UTF-16 sources are lost :(
>> I will write it again.
>
> Many thanks, committed as 0.9.14.8.

Wonderful, but...  Why did you rename file?

New version has UCS-2 and UCS-4 generated from same macro.  So
ucs-2.lisp is not good name :)

--

-- 
Ivan Boldyrev

                      Ok people, move along, there's nothing to see here.

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

_______________________________________________
Sbcl-devel mailing list
Sbcl-devel <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sbcl-devel

[PATCH] UCS-4 and UTF-16

ucs-2.lisp should be deleted, because new file superseds it (code for
UCS-2 and UCS-4 is generated by same macros).  If you want to rename
the file before committing, choose more generic name (like
"unicode.lisp" etc.).

--

-- 
Ivan Boldyrev

       Assembly of a Japanese bicycle requires greatest peace of spirit.

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

_______________________________________________
Sbcl-devel mailing list
Sbcl-devel <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sbcl-devel

access to foreign vector allocates memory, how to avoid that?

Hi guys,
  this is my first post so please be patient with me ;)
  I am building my own glue over sockint::send/recvfrom, because I
want to remove any foreign memory allocation when doing send/recv. I
have a struct like this:

(defstruct (raw-connection (:conc-name rcon-))
  socket
  buffer
  buffer-length
  buffer-dimension)

(defun create-raw-connection (socket buffer-dimension)
  (make-raw-connection :socket socket
		       :buffer (sb-alien:make-alien (array (sb-alien:unsigned 8) 1)
buffer-dimension)
		       :buffer-length 0
		       :buffer-dimension buffer-dimension))

(defun destroy-raw-connection (rcon)
  (when (socket-open-p (rcon-socket rcon))
    (socket-close (rcon-socket rcon)))
  (unless (null (rcon-buffer rcon))
    (sb-alien:free-alien (rcon-buffer rcon))
    (setf (rcon-buffer rcon) nil)))

(defun send (rcon buffer length &optional (flags 0))
  (let ((len (sb-bsd-sockets::with-vector-sap (buffer-sap buffer)
	       (sockint::send (socket-file-descriptor (rcon-socket rcon))
			      buffer-sap

Re: access to foreign vector allocates memory, how to avoid that?

"karol skocik" <karol.skocik <at> gmail.com> writes:

> I am building my own glue over sockint::send/recvfrom, because I
> want to remove any foreign memory allocation when doing send/recv. I
> have a struct like this:

You may want to look at using a vector of (UNSIGNED-BYTE 8) instead of
an alien array, and use WITH-PINNED-OBJECTS and VECTOR-SAP to pass a
pointer to the buffer to SEND and RECV:

;; something like (untested):
(let ((buf (make-array len :element-type '(unsigned-byte 8))))
  (sb-sys:with-pinned-objects (buf)
    (let ((sap (sb-sys:vector-sap buf)))
      (sockint::send fd sap len flags))))

Otherwise, to avoid the allocation with SB-ALIEN, you probably need to
add some type and/or OPTIMIZE declarations...

James

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

Re: access to foreign vector allocates memory, how to avoid that?

"karol skocik" <karol.skocik <at> gmail.com> writes:

> (defstruct (raw-connection (:conc-name rcon-))
>   socket
>   buffer
>   buffer-length
>   buffer-dimension)
>
> (defun create-raw-connection (socket buffer-dimension)
>   (make-raw-connection :socket socket
> 		       :buffer (sb-alien:make-alien (array (sb-alien:unsigned 8) 1)
> buffer-dimension)
> 		       :buffer-length 0
> 		       :buffer-dimension buffer-dimension))
[...]
> (defun copy-from-rcon-buf (rcon buffer length)
>   (let ((rbuf-sap (sb-alien:deref (rcon-buffer rcon))))
>     (loop :for i :from 0 :below length
> 	  :do (setf (aref buffer i) (sb-alien:deref rbuf-sap i)))))
>
> Here ^^^^^^^^^^^^^^^^^^^^^ is my problem, when I run test code from REPL :
[...]
> ;   could not optimize away %SAP-ALIEN: forced to do runtime
> ; allocation of alien-value structure
[...]
> this line :
>  (let ((rbuf-sap (sb-alien:deref (rcon-buffer rcon))))
>
> in copy-from-rcon-buf makes problem (allocation).