gmane.lisp.scheme.gauche

Re: [Planet Debian] David Bremner: Exporting Debianpackaging patches from git, (redux)*

Jens Thiele — 2013-05-01T13:15:47

sorry, wrong list

------------------------------------------------------------------------------
Introducing AppDynamics Lite, a free troubleshooting tool for Java/.NET
Get 100% visibility into your production application - at no cost.
Code-level diagnostics for performance bottlenecks with <2% overhead
Download for free and get started troubleshooting in minutes.
http://p.sf.net/sfu/appdyn_d2d_ap1

[Planet Debian] David Bremner: Exporting Debianpackaging patches from git, (redux)*

Jens Thiele — 2013-05-01T13:10:15

hmm
there really isn't a nice and simple packaging workflow using git?

------------------------------------------------------------------------------
Introducing AppDynamics Lite, a free troubleshooting tool for Java/.NET
Get 100% visibility into your production application - at no cost.
Code-level diagnostics for performance bottlenecks with <2% overhead
Download for free and get started troubleshooting in minutes.
http://p.sf.net/sfu/appdyn_d2d_ap1_______________________________________________
Gauche-devel mailing list
Gauche-devel< at >lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/gauche-devel

generize -> , ->>

shuji yamamoto — 2013-04-30T07:12:40

Hi all.
I wrote macro generized ->,->> of clojure.

(define-macro (-*> x form . more)
(if (pair? more)
`(-*> (-*> ,x ,form) ,< at >more)
(if (pair? form)
(receive (head tail) (break (cut eq? '<*> <>) form)
(if (null? tail)
`(,< at >head ,x)
`(,< at >head ,x ,< at >(cdr tail) ) ) )
`(,form ,x) ) ) )

#|
gosh> (-*> 1 (+ 12 <*>))
13
gosh> (-*> 1 (cut + 12 <*>))
#<closure #f>
gosh> (-*> 1 (cut + 12 <*>) (<*>) )
13
gosh> (-*> 1 (cut list 12 <*>) (<*>) )
(12 1)
gosh> (-*> 1 (cut list 12 <*> 67 ) (<*>) )
(12 1 67)
gosh> (-*> (quotient&remainder 13 5) (receive (x y) <*> (print x y)))
23
#<undef>
gosh> (-*> (quotient&remainder 13 5) (receive (x y) <*> (list x y)))
(2 3)

gosh> (-*> 'gauche (find-module <*>))
#<module gauche>
gosh> (-*> 'gauche (find-module <*>) (~ <*> 'table ) )
#<hash-table eq? 0x96f4ed8>
gosh> (-*> 'gauche (find-module <*>) (~ <*> 'table ) (~ <*> 'cond-list))
#<gloc gauche#cond-list>
gosh> (-*> 'gauche (find-module <*>) (global-variable-ref <*> 'cond-list))
#<macro cond-list>

gosh> (-*> 'gauche find-module )
#<module

Repository URL changed

Shiro Kawai — 2013-04-17T02:34:12

Recently Sourceforge.net rolled out a new infrastructure
and is asking projects to upgrade, which requires repo URL
to be changed.  I fetch this opportunity to migrate Gauche
and Gauche-gl repo to github.

  Gauche:    https://github.com/shirok/Gauche
  Gauche-gl: https://github.com/shirok/Gauche-gl

These are now 'official' repositories; the ones on sf.net
won't be updated and will eventually be deleted.  Note that
I still use sf.net for release file distribution and mailing
lists.

If you've been tracking Gauche HEAD, the easiest way is to
clone from the new repo.   Alternatively, if you have your
local stuff in your repo to carry over, a quick hack is to
edit '[remote "origin"]' section of .git/config like this:

[remote "origin"]
fetch = +refs/heads/*:refs/remotes/origin/*
url = git://github.com/shirok/Gauche.git

Then you should be able to do 'git pull' to get changes
from the new repo.

--shiro


------------------------------------------------------------------------------
Precog is a next-generatio

Re: (expt 2 1000000000000)

Shiro Kawai — 2013-04-09T12:13:04

Wtf?

gosh> (ash 1 1000000000000)
0

Ooh.

number.c: 3055: ScmObj Scm_Ash(ScmObj x, int cnt)
                                         ^^^

Aah.


------------------------------------------------------------------------------
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter

(expt 2 1000000000000)

Jens Thiele — 2013-04-09T11:16:00

gosh> (expt 2 1000000000000)
0

------------------------------------------------------------------------------
Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
http://www2.precog.com/precogplatform/slashdotnewsletter

Re: write-ber-integer ignores given port

Shiro Kawai — 2013-03-20T01:26:28

It is indeed a bug.  Patch applied.  Thanks.

--shiro

From: nozomi <nozomi< at >isozaki.jp>
Subject: [Gauche-devel] write-ber-integer ignores given port
Date: Wed, 20 Mar 2013 08:41:26 +0900


------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_d2d_mar

write-ber-integer ignores given port

nozomi — 2013-03-19T23:41:26

Hello,

Though I gave a file port to write-ber-integer, it wrote to the stdout.
Probably it is a bug.

Here is the patch to fix it.

diff --git ext/binary/io.scm ext/binary/io.scm
index bd825af..54ae77b 100644
--- ext/binary/io.scm
+++ ext/binary/io.scm
< at >< at > -219,8 +219,8 < at >< at >
         [start (ash number -7)])
     (unless (zero? start)
       (let loop ([n start])
-        (cond [(< n 128) (write-u8 (logior n #b10000000))]
+        (cond [(< n 128) (write-u8 (logior n #b10000000) port)]
               [else (loop (ash n -7)) ;; write high bytes first
-                    (write-u8 (logior (logand n #b01111111) #
b10000000))])))
-    (write-u8 final)))
+                    (write-u8 (logior (logand n #b01111111) #
b10000000) port)])))
+    (write-u8 final port)))


This is the first time for me to send a patch. I'd be glad if you 
indicate some mistakes
on this e-mail (if there is).

Regards

--
Nozomi Isozaki <nozomi< at >isozaki.jp>


------------------------------------------------------------------------------
Ev

Re: macro question

Shiro Kawai — 2013-03-13T09:41:52

I think it's a bug in Gauche's macro expander.  It should work as
you expected.

Rewriting the current hygienic macro expander is on my todo list
for several years already, and this put one more reason to do it
sooner.

From: Stephen Lewis <lewis+gauche< at >freeshell.org>
Subject: [Gauche-devel] macro question
Date: Tue, 12 Mar 2013 22:54:58 -0700

------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_d2d_mar

macro question

Stephen Lewis — 2013-03-13T05:54:58

Shiro,

I am still learning Scheme macros so I may be misunderstanding something,
but when I tried the following:

(define-syntax sl-macro
  (syntax-rules ()
                ((sl-macro _expr)
                 (let-syntax
                   ((sl-rule-1
                      (syntax-rules ()
                                    ((_ expr)
                                     (sl-rule-2 expr))
                                    ))
                    (sl-rule-2
                      (syntax-rules ()
                                    ((_ expr) (print expr)))))
                   (sl-rule-1 _expr)))))

(sl-macro (+ 1 2 3))

I get the error:
gosh> (sl-macro (+ 1 2 3))
*** ERROR: unbound variable: sl-rule-2

Which is understandable because 'sl-rule-2' is not in scope in sl-rule-1.
However when I replace the 'let-syntax' with 'letrec-syntax' the problem
is still there,

(define-syntax sl-macro
  (syntax-rules ()
((sl-macro _expr)
 (letrec-syntax
   ((sl-rule-1
      (syntax-rules ()
    ((_ expr)

Re: (make-u8vector (ash 1 32))

Jens Thiele — 2013-02-20T22:55:00


sounds great, exactly what i am looking for :-)
(using a c-wrapper based hack at the moment)

------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_d2d_feb

Re: (make-u8vector (ash 1 32))

Shiro Kawai — 2013-02-20T18:46:34

Argggh.  The old curse of ILP32.

I think, when everybody had 4GB limit, I just didn't seriously 
consider these cases since it wasn't feasible anyway.  But yeah,
nowadays you might want to use such a large buffer.

Strictly speaking, changing arg from 'int size' to 'size_t size'
would break ABI compatibility, I think.  That is, exising binaries
calling libgauche may pass 'size' arg without clearing the upper
32 bits of the register.  I'm not sure, though.

I'm inclined to make this change now, but if anybody knows that
this change would cause serious ABI incompatibility issue, please
let me know.

Btw, Scm_Make**Vector() are now in src/vector.c; it's hidden in
cpp macros so grep won't show them.

--shiro

From: Jens Thiele <karme< at >karme.de>
Subject: Re: [Gauche-devel] (make-u8vector (ash 1 32))
Date: Wed, 20 Feb 2013 16:16:57 +0100


------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDyn

Re: (make-u8vector (ash 1 32))

Shiro Kawai — 2013-02-20T19:08:45

I have a plan to add mmap interface that's accessible as
uvectors.  The "owner" field of uvector is inteded to keep
the pointer to some descriptor of mapped memory (so that
the mem can be unmapped when all the references are dropped).

From: Jens Thiele <karme< at >karme.de>
Subject: Re: [Gauche-devel] (make-u8vector (ash 1 32))
Date: Wed, 20 Feb 2013 19:56:23 +0100

------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_d2d_feb

Re: (make-u8vector (ash 1 32))

Jens Thiele — 2013-02-20T18:56:23


i ended up there trying to use uvector and a huge memory mapped file


thanks! and yes it was grep :-)

------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_d2d_feb

Re: (make-u8vector (ash 1 32))

Jens Thiele — 2013-02-20T15:16:57

further playing around in that direction:
let's make sure we run out of memory early:
$ulimit -v 100000 -m 100000
$gosh -ugauche.uvector
gosh> (make-u8vector (ash 1 29))
GC Warning: Out of Memory! Heap size: 2 MiB. Returning NULL!
out of memory (536870912).  aborting...

ok, fine

$ export LC_ALL=C; gosh -ugauche.uvector
gosh> (uvector-alias <u8vector> (make-u64vector (ash 1 30)))
Segmentation fault

(first thought stored size is in bytes but now i think it is the number
of elements => crash likely not caused by uvector-alias...)

$ export LC_ALL=C; gosh -ugauche.uvector
gosh> (make-u64vector (ash 1 30))
Segmentation fault

stack trace:
#0  0x00007fa48f2d8ee8 in Scm_MakeU64Vector ()
from /usr/lib/libgauche-0.9.so.0

hmm, can't find the code at the moment, it was generated but removed
again?

------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/

(make-u8vector (ash 1 32))

Jens Thiele — 2013-02-20T14:33:52

gosh> (make-u8vector (ash 1 32))
#u8()
gosh> (make-u8vector (ash 1 60))
#u8()

:-)

looking at the uvector functions size is an int (32bit on amd64/x86_64)
internally size is stored as ScmWord => intptr_t which is 64bit on amd64/x86_64
shouldn't the uvector functions use something like size_t then?

greetings
karme

------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_d2d_feb

Re: stream question / unbound memory?

Jens Thiele — 2013-01-13T15:34:29


only did some quick tests
seems to work
thanks!

------------------------------------------------------------------------------
Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS,
MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current
with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft
MVPs and experts. ON SALE this month only -- learn more at:
http://p.sf.net/sfu/learnmore_123012

Re: gosh crashes with big arg values in SICP expmod function

Shiro Kawai — 2013-01-09T13:06:14

Just pushed commit 7d19820 that fixed this bug.
It was a Gauche compiler bug that made stack estimation incorrect,
and in this particular case it caused stack overrun during deep
recursion of expmod and square.  The bug hasn't been surfaced since
satisfying criteria to cause the overrun was quite tricky.

--shiro


From: alpha gamma <mnfp13< at >gmail.com>
Subject: [Gauche-devel] gosh crashes with big arg values in SICP expmod function
Date: Sat, 5 Jan 2013 00:59:29 +0100



------------------------------------------------------------------------------
Master Java SE, Java EE, Eclipse, Spring, Hibernate, JavaScript, jQuery
and much more. Keep your Java skills current with LearnJavaNow -
200+ hours of step-by-step video tutorials by Java experts.
SALE $49.99 this month only -- learn more at:
http://p.sf.net/sfu/learnmore_122612

Re: stream question / unbound memory?

Shiro Kawai — 2013-01-09T12:04:21

Hi Jens,

I think I plugged the leak by commit d6a56bb.  It was actually
a dumb mistake introduced when I made force thread-safe---the
code installs dynamic handler but never rewound it.  The leak
test now runs in bounded memory.

From: Jens Thiele <karme< at >karme.de>
Subject: Re: [Gauche-devel] stream question / unbound memory?
Date: Thu, 29 Nov 2012 15:34:53 +0100

------------------------------------------------------------------------------
Master Java SE, Java EE, Eclipse, Spring, Hibernate, JavaScript, jQuery
and much more. Keep your Java skills current with LearnJavaNow -
200+ hours of step-by-step video tutorials by Java experts.
SALE $49.99 this month only -- learn more at:
http://p.sf.net/sfu/learnmore_122612

Re: gosh crashes with big arg values in SICP expmod function

Shiro Kawai — 2013-01-05T04:10:44

Confirmed.  Actually it works on 0.9.2, so it's a regression.
I'm looking at it.

--shiro

From: alpha gamma <mnfp13< at >gmail.com>
Subject: [Gauche-devel] gosh crashes with big arg values in SICP expmod function
Date: Sat, 5 Jan 2013 00:59:29 +0100


------------------------------------------------------------------------------
Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS,
MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current
with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft
MVPs and experts. SALE $99.99 this month only -- learn more at:
http://p.sf.net/sfu/learnmore_122912

gosh crashes with big arg values in SICP expmodfunction

alpha gamma — 2013-01-04T23:59:29

Hello,

I see a gosh crash when executing the expmod function (as defined in the
SICP book
in "The Fermat Test" paragraph) when passing some big number to it.

This is the code:

gosh> (define (expmod base exp m)
  (cond ((= exp 0) 1)
        ((even? exp)
         (remainder (square (expmod base (/ exp 2) m))
                    m))
        (else
         (remainder (* base (expmod base (- exp 1) m))
                    m))))
expmod
gosh> expmod
#<closure expmod>
gosh> (expmod
39946212748484926716547576613536335244212076288673919135945266186958748103426425703814309198601642350315000896671931766728383932165170481242281502692377651264308084928779825974300461250516314879806400112473529994805032016172865793130893434344597739942980810628981974738091062467640291490240751539719709498119
4579711411739303932781641950823536362260984847560207996105061048524744666028182744596635729822182619157340621001866326570968870025285337835239875057209431981746998334958692562298763199271303752731161229378629533372192349752082022955