q-lang-users — general discussion, questions, suggestions, etc.

Re: [q-lang-users] ANN: Q 7.0 release candidate

[John C. <co...@cc...>]

Albert Graef scripsit:

> >Thanks for responding so quickly.  Unfortunately, these aren't the
> >clearest of instructions.  I tried adding "$(LIBICONV)" to the definition
> 
> Hmm, that should probably be $(LTLIBICONV), can you post the two 
> definitions please?

They are defined identically as "-liconv".

> Hmm, looks like automake tries to execute autoconf (maybe wants to get 
> the autoconf version?) and doesn't find it. Is autoconf/automake working 
> with other programs? You could also try to fix the corresponding 
> Makefile.in's instead, add the proper link options manually and then 
> rerun configure, if automake doesn't work for you. (But see below.)

I've never tried to run automake before, so I don't really know.  At most
I've installed patches into Makefile.in.

> Yes, clearly the proper link option for libiconv is missing. So I'll 
> have to add a corresponding LIBADD line in libq/Makefile.am and also 
> modify the LIBADD lines in the other Makefile.am's I mentioned. If you 
> can wait a little more, I'll  try to figure out what the right @XYZ@ 
> variables are which have to be included for libiconv and libgettext, and 
> then prepare a new tarball.

This is a new box and my first attempt to install 7.0, so there's no hurry.

-- 
Possession is said to be nine points of the law,		John Cowan
but that's not saying how many points the law might have.	co...@cc...
	--Thomas A. Cowan (law professor and my father

Re: [q-lang-users] ANN: Q 7.0 release candidate

[Albert G. <Dr....@t-...>]

John Cowan wrote:
> Thanks for responding so quickly.  Unfortunately, these aren't the
> clearest of instructions.  I tried adding "$(LIBICONV)" to the definition

Hmm, that should probably be $(LTLIBICONV), can you post the two 
definitions please?

> of libq_la_LDFLAGS in libq/Makefile.am, but then was roadblocked:
> "automake" crashes with "automake: `configure.ac' or `configure.in'

Yep, for libq/Makefile.am, automake needs to be run in the main source 
dir. Or you can try to just run autogen.sh, that should rebuild the 
entire configury (but probably that won't work either, given that 
automake gives you the troubles you mentioned below).

> is required", and trying it in the q-7.0 root directory crashes with
> "sh: autoconf: command not found".

Hmm, looks like automake tries to execute autoconf (maybe wants to get 
the autoconf version?) and doesn't find it. Is autoconf/automake working 
with other programs? You could also try to fix the corresponding 
Makefile.in's instead, add the proper link options manually and then 
rerun configure, if automake doesn't work for you. (But see below.)

> Here's the current state of what make (in the root) says:
> /bin/sh ../libtool --tag=CC --mode=link gcc  -g -O2   -o libq.la -rpath /usr/local/lib -no-undefined -version-info 7:0:0 libq.lo  -lcrypt -lutil -lm 
> rm -fr  .libs/libq.dll.a
> gcc -shared  .libs/libq.o  -lcrypt -lutil  -o .libs/cygq-7.dll -Wl,--image-base=0x10000000 -Wl,--out-implib,.libs/libq.dll.a
> Creating library file: .libs/libq.dll.a
> .libs/libq.o: In function `from_utf8':
> /opt/q-7.0/libq/libq.c:508: undefined reference to `_libiconv_open'
> /opt/q-7.0/libq/libq.c:539: undefined reference to `_libiconv'
> /opt/q-7.0/libq/libq.c:563: undefined reference to `_libiconv'
> /opt/q-7.0/libq/libq.c:568: undefined reference to `_libiconv_close'

Yes, clearly the proper link option for libiconv is missing. So I'll 
have to add a corresponding LIBADD line in libq/Makefile.am and also 
modify the LIBADD lines in the other Makefile.am's I mentioned. If you 
can wait a little more, I'll  try to figure out what the right @XYZ@ 
variables are which have to be included for libiconv and libgettext, and 
then prepare a new tarball.

Albert

-- 
Dr. Albert Gr"af
Dept. of Music-Informatics, University of Mainz, Germany
Email:  Dr....@t-..., ag...@mu...
WWW:    http://www.musikinformatik.uni-mainz.de/ag

Re: [q-lang-users] ANN: Q 7.0 release candidate

[John C. <co...@cc...>]

Albert Graef scripsit:

> looks like I forgot about LIBICONV because here on Linux I have iconv in 
> libc. In principle, adding $(LIBICONV) or $(LTLIBICONV) to the link 
> lines of compiler and interpreter (src/Makefile.am), libq 
> (libq/Makefile.am) and clib (modules/clib/Makefile.am) should do the 
> trick. Something similar will be needed to make clib::gettext and 
> friends work. If that doesn't help then please post the complete make 
> log so that we can figure out what exactly goes wrong.

Thanks for responding so quickly.  Unfortunately, these aren't the
clearest of instructions.  I tried adding "$(LIBICONV)" to the definition
of libq_la_LDFLAGS in libq/Makefile.am, but then was roadblocked:
"automake" crashes with "automake: `configure.ac' or `configure.in'
is required", and trying it in the q-7.0 root directory crashes with
"sh: autoconf: command not found".

Attempts to do "make Makefile.in" report that it is up to date, and
"make Makefile" says there is no such target.

Here's the current state of what make (in the root) says:

$ make
make  all-recursive
make[1]: Entering directory `/opt/q-7.0'
Making all in .
make[2]: Entering directory `/opt/q-7.0'
make[2]: Leaving directory `/opt/q-7.0'
Making all in libltdl
make[2]: Entering directory `/opt/q-7.0/libltdl'
make  all-am
make[3]: Entering directory `/opt/q-7.0/libltdl'
make[3]: Leaving directory `/opt/q-7.0/libltdl'
make[2]: Leaving directory `/opt/q-7.0/libltdl'
Making all in glob
make[2]: Entering directory `/opt/q-7.0/glob'
make[2]: Nothing to be done for `all'.
make[2]: Leaving directory `/opt/q-7.0/glob'
Making all in regex
make[2]: Entering directory `/opt/q-7.0/regex'
make[2]: Nothing to be done for `all'.
make[2]: Leaving directory `/opt/q-7.0/regex'
Making all in doc
make[2]: Entering directory `/opt/q-7.0/doc'
make[2]: Nothing to be done for `all'.
make[2]: Leaving directory `/opt/q-7.0/doc'
Making all in etc
make[2]: Entering directory `/opt/q-7.0/etc'
make[2]: Nothing to be done for `all'.
make[2]: Leaving directory `/opt/q-7.0/etc'
Making all in examples
make[2]: Entering directory `/opt/q-7.0/examples'
make[2]: Nothing to be done for `all'.
make[2]: Leaving directory `/opt/q-7.0/examples'
Making all in libq
make[2]: Entering directory `/opt/q-7.0/libq'
/bin/sh ../libtool --tag=CC --mode=link gcc  -g -O2   -o libq.la -rpath /usr/local/lib -no-undefined -version-info 7:0:0 libq.lo  -lcrypt -lutil -lm 
rm -fr  .libs/libq.dll.a
gcc -shared  .libs/libq.o  -lcrypt -lutil  -o .libs/cygq-7.dll -Wl,--image-base=0x10000000 -Wl,--out-implib,.libs/libq.dll.a
Creating library file: .libs/libq.dll.a
.libs/libq.o: In function `from_utf8':
/opt/q-7.0/libq/libq.c:508: undefined reference to `_libiconv_open'
/opt/q-7.0/libq/libq.c:539: undefined reference to `_libiconv'
/opt/q-7.0/libq/libq.c:563: undefined reference to `_libiconv'
/opt/q-7.0/libq/libq.c:568: undefined reference to `_libiconv_close'
.libs/libq.o: In function `to_utf8':
/opt/q-7.0/libq/libq.c:587: undefined reference to `_libiconv_open'
/opt/q-7.0/libq/libq.c:618: undefined reference to `_libiconv'
/opt/q-7.0/libq/libq.c:623: undefined reference to `_libiconv_close'
.libs/libq.o: In function `file_from_utf8':
/opt/q-7.0/libq/libq.c:676: undefined reference to `_libiconv'
/opt/q-7.0/libq/libq.c:643: undefined reference to `_libiconv_open'
.libs/libq.o: In function `file_to_utf8':
/opt/q-7.0/libq/libq.c:728: undefined reference to `_libiconv'
/opt/q-7.0/libq/libq.c:698: undefined reference to `_libiconv_open'
.libs/libq.o: In function `file_encoding':
/opt/q-7.0/libq/libq.c:778: undefined reference to `_libiconv_close'
/opt/q-7.0/libq/libq.c:765: undefined reference to `_libiconv_close'
/opt/q-7.0/libq/libq.c:750: undefined reference to `_libiconv_open'
/opt/q-7.0/libq/libq.c:751: undefined reference to `_libiconv_open'
/opt/q-7.0/libq/libq.c:754: undefined reference to `_libiconv_close'
/opt/q-7.0/libq/libq.c:756: undefined reference to `_libiconv_close'
collect2: ld returned 1 exit status
make[2]: *** [libq.la] Error 1
make[2]: Leaving directory `/opt/q-7.0/libq'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/opt/q-7.0'
make: *** [all] Error 2

-- 
Babies are born as a result of the              John Cowan
mating between men and women, and most          http://www.reutershealth.com
men and women enjoy mating.                     http://www.ccil.org/~cowan
    --Isaac Asimov in Earth: Our Crowded Spaceship      co...@cc...

[q-lang-users] Is this the Q we all know and love?

[WW <w_...@ea...>]

http://www.tiobe.com/tpci.htm - at number 49.

Regards,

Walter

Re: [q-lang-users] ANN: Q 7.0 release candidate

[Albert G. <Dr....@t-...>]

John Cowan wrote:
> My attempt to build 7.0 on Cygwin failed because _libiconv_open and
> _libiconv_close weren't found when trying to build libq.la.  I have
> libiconv installed under Cygwin, so patching around this is probably a
> matter of inserting a strategic "-liconv" into the Makefile somewhere.
> It's already mentioned in the defines for LIBICONV, LTLIBICONV, and
> XML_LIBS; adding it to LIBS and LDFLAGS didn't help any.
> 
> I presume the fix is simple, though.  Can anyone help?

Hi John,

looks like I forgot about LIBICONV because here on Linux I have iconv in 
libc. In principle, adding $(LIBICONV) or $(LTLIBICONV) to the link 
lines of compiler and interpreter (src/Makefile.am), libq 
(libq/Makefile.am) and clib (modules/clib/Makefile.am) should do the 
trick. Something similar will be needed to make clib::gettext and 
friends work. If that doesn't help then please post the complete make 
log so that we can figure out what exactly goes wrong.

Cheers,
Albert

-- 
Dr. Albert Gr"af
Dept. of Music-Informatics, University of Mainz, Germany
Email:  Dr....@t-..., ag...@mu...
WWW:    http://www.musikinformatik.uni-mainz.de/ag

Re: [q-lang-users] ANN: Q 7.0 release candidate

[John C. <co...@cc...>]

Albert Graef scripsit:

> it took much too long, but at long last I finished unicode support in 
> the Q interpreter and the core modules. 

Hurrah!

> Windows users: sorry, but you will have to wait a little more, I hope I 
> get around porting the new release this week. I also want to check other 
> platforms such as OS X asap, and I still have to overhaul the add-on 
> modules (multimedia library etc.) to make them unicode-aware. (Another 
> big item on my TODO list which didn't make it into the 7.0 release yet 
> is 64 bit support, I'm afraid that this will have to wait until Q 7.1.)

My attempt to build 7.0 on Cygwin failed because _libiconv_open and
_libiconv_close weren't found when trying to build libq.la.  I have
libiconv installed under Cygwin, so patching around this is probably a
matter of inserting a strategic "-liconv" into the Makefile somewhere.
It's already mentioned in the defines for LIBICONV, LTLIBICONV, and
XML_LIBS; adding it to LIBS and LDFLAGS didn't help any.

I presume the fix is simple, though.  Can anyone help?

-- 
You escaped them by the will-death              John Cowan
and the Way of the Black Wheel.                 co...@cc...
I could not.  --Great-Souled Sam                http://www.ccil.org/~cowan

[q-lang-users] =?ISO-2022-JP?B?GyRCIXoheSF6OiMkXiRHJTUlJCVIJE5BKiRTGyhC?= =?ISO-2022-JP?B?GyRCSn00VjBjJEMkRiReJDskcyQrISkheiF5IXobKEI=?= 13

[<in...@nn...>]

///////////////////////////////////////////////////////////////////////

$B!z!y!z:#$^$G=P2q$$7O%5%$%H$NA*$SJ}4V0c$C$F$^$;$s$+!)!z!y!z(B

///////////////////////////////////////////////////////////////////////







$B%a%kM'Jg=8$N%5%$%H$G$O%a!<%k$@$1$G<B:]$K$O=P2q$($^$;$s!#(B

$B3d$j@Z$C$?4X78!"3d$j@Z$i$J$$4X78!"$I$A$i$b=P2q$C$F$+$i$N;O$^$j$G$9!#(B



$B!yL5NA%]%$%s%H#1#2#0#0#01_J,!y(B

$B!z$b$A$m$sL5NA%]%$%s%H$G%"%I%l%98r49$b#O#K$G$9!z(B





$B!z2<5-%5%$%H$G$N%@%i%a$O5,Ls0cH?$G$9$h!z(B

http://nncom.info/wty/











////////////////////////////////////////////////////

$BITMW$NJ}$O%3%A%i$^$G(B

in...@ok...

[q-lang-users] ANN: Q 7.0 release candidate

[Albert G. <Dr....@t-...>]

Hi,

it took much too long, but at long last I finished unicode support in 
the Q interpreter and the core modules. As a bonus, I also threw in a 
new GUI builder which generates native Q code and is based on Peter G. 
Baum's Gnocl, a very nice Tcl extension for building GNOME/GTK+ 
applications (http://gnocl.sf.net).

I have already been testing the new release on my Linux development box 
for some weeks, and the unicode stuff seems to work o.k., at least in 
utf8 and latin1 locales. But due to the complex nature of the changes, 
I'm first releasing a candidate, in the hope that at least some of you 
can do some more testing. In particular, it would be nice if those of 
you who work in non-Western locales could give the new release a whirl 
(do we have any Q programmers from the Far East on the list?), so that 
bugs can be weeded out before the final release.

Windows users: sorry, but you will have to wait a little more, I hope I 
get around porting the new release this week. I also want to check other 
platforms such as OS X asap, and I still have to overhaul the add-on 
modules (multimedia library etc.) to make them unicode-aware. (Another 
big item on my TODO list which didn't make it into the 7.0 release yet 
is 64 bit support, I'm afraid that this will have to wait until Q 7.1.)

The q-7.0rc1 tarball is available for download here: 
http://sourceforge.net/project/showfiles.php?group_id=96881&package_id=103965

And you can find the release notes here:
http://sourceforge.net/project/shownotes.php?release_id=391246&group_id=96881

A screenshot of the new GUI builder is available here:
http://q-lang.sourceforge.net/gqbuilder.png

Enjoy! :)
Albert

-- 
Dr. Albert Gr"af
Dept. of Music-Informatics, University of Mainz, Germany
Email:  Dr....@t-..., ag...@mu...
WWW:    http://www.musikinformatik.uni-mainz.de/ag

[q-lang-users] =?ISO-2022-JP?B?GyRCJDRNeE1RTS1GcSQmOGY6QiQkJF4kOSEjJV0lJCVzJUhESTJDGyhC?=

[<con...@uu...>]

$B5.J}MM$H2q$$$?$$$H$$$&=w@-2q0wMM$,$*$j$^$9!#(B

$B"'$*JV;v$3$A$i$+$i"'(B

http://311.send-mail.info/kanzen_muryo.php?id=35q76jhfgd8gl7yjs



$B$$$D$b$4MxMQD:$$$F@?$KM-Fq$&8f:B$$$^$9!#(B

$B%5%$%H%j%K%e!<%"%k$KH<$$!":FEPO?<jB3$-$,I,MW$G$9!#(B

$B$^$?!"#1#0#0#0#01_J,$NL5NA%]%$%s%H$b$4MQ0U$7$F$*$j$^$9$N$G$<$R:FEPO?<jB3$-$r:Q$^$;$F$/$@$5$$!#(B

http://311.send-mail.info/kanzen_muryo.php?id=i479fgd8glfgh75kwhg



$B"(%o%s%/%j%C%/!&ITEv@A5aEy$G$O$"$j$^$;$s$N$G!"0B?4$7$F$4MxMQ$/$@$5$$!#(B



$B:FEPO?$,I,MW$N$J$$J}$O(B

con...@uu...

Re: [q-lang-users] modules and visibility (Fixed?)

[Albert G. <Dr....@t-...>]

Arnoldo Jose Muller Molina wrote:
> Just after sending the e-mail it occurred to me that Q wasn't publishing
> the symbols I was using when matching, so I changed t2.q to:

Exactly. :) In the interpreter you're always in the namespace of the 
main script you're running. You were running t2, t1 doesn't export those 
symbols to t2 => same identifier will refer to a different symbol in the 
namespace of the main script (that symbol is created by the interpreter 
on the fly). Exactly like it's supposed to be.  When in doubt, try the 
whos command:

==> whos fcondition

fcondition      not a defined symbol

If the interpreter actually knows the symbol then whos will show you the 
module in which it is defined.

==> whos map

map             function symbol defined in /usr/share/q/lib/stdlib.q
                 map $1 $2

> public fcondition;
> public parameterRef;
> public methodRef;
> public fieldRef;
> public ftype;
> public localRef;

Of course you could just write:

public fcondition, parameterRef, methodRef, fieldRef, ftype, localRef;

> Now it works... Would this be the best solution?

Well, it looks like your symbols are actually supposed to be constants, 
so why not turn them into an enumeration type:

public type FSpec = const fcondition, parameterRef, methodRef, fieldRef, 
ftype, localRef;

Then you can just write:

fspec _:FSpec = true;
fspec _       = false otherwise;

As another nice side effect, you can now also compare the FSpec values 
using (=), (<) etc. :)

HTH,
Albert

-- 
Dr. Albert Gr"af
Dept. of Music-Informatics, University of Mainz, Germany
Email:  Dr....@t-..., ag...@mu...
WWW:    http://www.musikinformatik.uni-mainz.de/ag

[q-lang-users] modules and visibility (Fixed?)

[Arnoldo J. M. M. <ar...@da...>]

Hello All,

Just after sending the e-mail it occurred to me that Q wasn't publishing
the symbols I was using when matching, so I changed t2.q to:

/////////////////////
public fspec;

// change: (* x *)V
public fcondition;
public parameterRef;
public methodRef;
public fieldRef;
public ftype;
public localRef;
// change: (* x *)V

fspec(fcondition) = true;
fspec(parameterRef) = true;
fspec(methodRef) = true;
fspec(fieldRef) = true;
fspec(ftype) = true;
fspec(localRef) = true;
fspec(_) = false;
/////////////////////

Now it works... Would this be the best solution? 

Thanks again!

Best regards,

AJMM.


On Sun, 2006-01-08 at 13:02 +0900, Arnoldo Jose Muller Molina wrote:
> Dear All,
> 
> Thank you so much for all your previous help. Thanks to your advice I've
> been able to understand Q better. Every day I like Q more and more!
> (* x *)V
> 
> 
> I have two files:
> t1.q:
> ////////////////////////////
> public fspec;
> 
> fspec(fcondition) = true;
> fspec(parameterRef) = true;
> fspec(methodRef) = true;
> fspec(fieldRef) = true;
> fspec(ftype) = true;
> fspec(localRef) = true;
> fspec(_) = false;
> ////////////////////////////
> 
> 
> and t2.q:
> ////////////////////////////
> import "t1";
> ////////////////////////////
> 
> When I execute the interpreter with  "q t1.q" and enter the following:
> ==> fspec(localRef)
> true
> 
> I will of course get true, 
> 
> When I execute the interpreter with "q t2.q" and enter:
> ==> fspec(localRef);
> false
> 
> I unexpectedly get false...
> 
> Using "debug on" I get the following:
> 
> ==> fspec(localRef);
>   0>  t1.q, line 10: fspec localRef  ==>  false
> (type ? for help)
> :
> **  fspec localRef  ==>  false
> false
> 
> (Line 10 is the last rule of the t1.q file)
> 
> What am I doing wrong? If I comment out line 10, Q will not execute the
> rule fspec, it's like the fspec rule disappears.
> (T _ T)
> 
> Thanks in advance!
> 
>  
> 
> 
> 

[q-lang-users] modules and visibility

[Arnoldo J. M. M. <ar...@da...>]

Dear All,

Thank you so much for all your previous help. Thanks to your advice I've
been able to understand Q better. Every day I like Q more and more!
(* x *)V


I have two files:
t1.q:
////////////////////////////
public fspec;

fspec(fcondition) = true;
fspec(parameterRef) = true;
fspec(methodRef) = true;
fspec(fieldRef) = true;
fspec(ftype) = true;
fspec(localRef) = true;
fspec(_) = false;
////////////////////////////


and t2.q:
////////////////////////////
import "t1";
////////////////////////////

When I execute the interpreter with  "q t1.q" and enter the following:
==> fspec(localRef)
true

I will of course get true, 

When I execute the interpreter with "q t2.q" and enter:
==> fspec(localRef);
false

I unexpectedly get false...

Using "debug on" I get the following:

==> fspec(localRef);
  0>  t1.q, line 10: fspec localRef  ==>  false
(type ? for help)
:
**  fspec localRef  ==>  false
false

(Line 10 is the last rule of the t1.q file)

What am I doing wrong? If I comment out line 10, Q will not execute the
rule fspec, it's like the fspec rule disappears.
(T _ T)

Thanks in advance!

 

[q-lang-users] RE: q-lang-users digest, Vol 1 #140 - 2 msgs

[Sheehy, M. (MH) <mar...@ma...>]

Thanks very much for your detailed response.

It is greatly appreciated.
Greg


-----Original Message-----
From: q-l...@li...
[mailto:q-l...@li...]
Sent: Wednesday, 28 December 2005 3:29 PM
To: q-l...@li...
Subject: q-lang-users digest, Vol 1 #140 - 2 msgs


Send q-lang-users mailing list submissions to
	q-l...@li...

To subscribe or unsubscribe via the World Wide Web, visit
	https://lists.sourceforge.net/lists/listinfo/q-lang-users
or, via email, send a message with subject or body 'help' to
	q-l...@li...

You can reach the person managing the list at
	q-l...@li...

When replying, please edit your Subject line so it is more specific
than "Re: Contents of q-lang-users digest..."


Today's Topics:

   1. Re: newbie question: Efficiency & other stuff (Albert Graef)
   2. Re: Newbie: Backtracking again (Albert Graef)

--__--__--

Message: 1
Date: Tue, 27 Dec 2005 09:40:52 +0100
From: Albert Graef <Dr....@t-...>
To:  q-l...@li...
Subject: Re: [q-lang-users] newbie question: Efficiency & other stuff
Reply-To: q-l...@li...

Arnoldo Jose Muller Molina wrote:
> dmatch(E1,E2) = (#expand(E1) + #expand(E2)) - (2 * dmatchAux(expand
> (E1),expand(E2)));

As Q functions may have side effects, it is not possible to optimize 
away shared subexpressions like this. That's what where clauses are for. 
Try something like:

dmatch(E1,E2) = (#X1 + #X2) - (2 * dmatchAux(X1,X2))
   where X1 = expand(E1), X2 = expand(E2);

> I have executed some stats commands after executing this rule, with and
> without the "(#expand(E1) + #expand(E2))" and it seems that q is
> calculating expand(E1) and expand(E2) 2 times (1). 
> 
> 1) Is this true?
> 
> 2) if that is so, is there a way of implementing a "let" so I can
> compute only once the results of expand(_)? Of course I could create an
> elaborate chain of rules and allow this, but is there a way that doesn't
> involve creating several rules? 
> 
> Thank you!
> 
> --------------------------
> (1) (expand(...) is a deterministic rule, it will always return the same
> result for the same input)

-- 
Dr. Albert Gr"af
Dept. of Music-Informatics, University of Mainz, Germany
Email:  Dr....@t-..., ag...@mu...
WWW:    http://www.musikinformatik.uni-mainz.de/ag


--__--__--

Message: 2
Date: Tue, 27 Dec 2005 10:47:55 +0100
From: Albert Graef <Dr....@t-...>
To:  q-l...@li...
CC:  Gre...@ma...
Subject: Re: [q-lang-users] Newbie: Backtracking again
Reply-To: q-l...@li...

Mele, Greg (MH) wrote:
> I dont understand the queens algorithm enough to see how backtracking 
> works by the fail instruction.
>  
> I would like to have a prolog-like system:
> ie the classic genealogical database
>  
>  
> father(a,b).
> father(b,c).
> father(c,d).
>  
> ancestor(X,Y) :- father (X,Z), ancestor (Z,Y). 

Prolog uses both unification and backtracking to solve problems like 
this. Of course you could do something similar in Q, but it would 
essentially amount to writing a simple Prolog interpreter in Q. I won't 
go into this here, but I'll try to explain the workings of the fail 
construct in the following.

The fail instruction is only useful to implement backtracking, without 
the unification part. It just tells the interpreter that the current 
equation has failed, so that it moves on to the next applicable 
equation. E.g., if you have two equations like this:

foo = writes "foo#1\n" || fail;
foo = writes "foo#2\n";

and you evaluate `foo' then the interpreter first starts evaluating the 
right-hand side of the first rule. After printing out "foo#1", it will 
evaluate the `fail', causing the first rule to fail. Now the second rule 
gets executed, printing "foo#2". So the end result would be the 
following printout:

foo#1
foo#2

(I also often use this trick for debugging purposes. E.g., if you have a 
function bar X with a bunch of defining equations, then simply putting 
an equation like bar X = printf "bar invoked with arg %s\n" (str X) || 
fail; in front of those equations allows you to print bar's argument 
whenever it is invoked, after which the bar X call is evaluated as usual.)

Ok, now how does the queens example work? Let's take a look at the 
crucial definition of the search function (which is invoked as search N 
1 1 [] from queens N):

search N I J P	= write P || writes "\n" if I>N;
		= search N (I+1) 1 (P++[(I,J)]) || fail if safe (I,J) P;
		= search N I (J+1) P if J<N;
		= () otherwise;

N is the size of the board (number of rows/columns), I the current row, 
J the current column and P the list of places (row/column pairs) of the 
queens determined so far (in rows 0..I-1 of the board).

The first equation simply checks whether we've already found a solution 
(I>N implies that #P=N) which is printed out.

The second equation is where the backtracking actually happens. If it is 
safe to place the next queen at position (I,J), w.r.t. the places P we 
already have, then we add (I,J) to P and invoke the algorithm 
recursively starting at column 1 in the next row. After returning from 
the recursive call, the entire subspace of the solution space with the 
given placements P++[(I,J)] has been covered. Now we let the equation 
fail, so that the interpreter moves on to the third equation to try 
other columns for the Ith row.

In any case (i.e., also if the second equation is skipped because (I,J) 
is not safe), the third equation moves on to the next column J+1 and 
invokes the algorithm recursively. This only happens if J<N, otherwise 
there are no other potential placements for the Ith row in which case 
the fourth equation returns ().

Note that the third equation is tail-recursive, i.e., it is essentially 
just a loop which iterates over the different columns for the same row 
I. The second equation is where the algorithm recursively descends into 
the solution subspace for a given initial placement.

Well, I hope that this clarifies the inner workings of the algorithm a 
bit. Tracing the algorithm for a small value of N (either on paper or 
with the interpreter's debugger) also helps. ;-)

Cheers,
Albert

-- 
Dr. Albert Gr"af
Dept. of Music-Informatics, University of Mainz, Germany
Email:  Dr....@t-..., ag...@mu...
WWW:    http://www.musikinformatik.uni-mainz.de/ag



--__--__--

_______________________________________________
q-lang-users mailing list
q-l...@li...
https://lists.sourceforge.net/lists/listinfo/q-lang-users


End of q-lang-users Digest
THIS E-MAIL IS CONFIDENTIAL.  If you have received this e-mail in error,
please notify us by return e-mail and delete the document.  If you are not
the intended recipient you are hereby notified that any disclosure, copying,
distribution or taking any action in reliance on the contents of this
information is strictly prohibited and may be unlawful.  Eastern Health is
not liable for the proper and complete transmission of the information
contained in this communication or for any delay in its receipt.

Re: [q-lang-users] Newbie: Backtracking again

[Albert G. <Dr....@t-...>]

Mele, Greg (MH) wrote:
> I dont understand the queens algorithm enough to see how backtracking 
> works by the fail instruction.
>  
> I would like to have a prolog-like system:
> ie the classic genealogical database
>  
>  
> father(a,b).
> father(b,c).
> father(c,d).
>  
> ancestor(X,Y) :- father (X,Z), ancestor (Z,Y). 

Prolog uses both unification and backtracking to solve problems like 
this. Of course you could do something similar in Q, but it would 
essentially amount to writing a simple Prolog interpreter in Q. I won't 
go into this here, but I'll try to explain the workings of the fail 
construct in the following.

The fail instruction is only useful to implement backtracking, without 
the unification part. It just tells the interpreter that the current 
equation has failed, so that it moves on to the next applicable 
equation. E.g., if you have two equations like this:

foo = writes "foo#1\n" || fail;
foo = writes "foo#2\n";

and you evaluate `foo' then the interpreter first starts evaluating the 
right-hand side of the first rule. After printing out "foo#1", it will 
evaluate the `fail', causing the first rule to fail. Now the second rule 
gets executed, printing "foo#2". So the end result would be the 
following printout:

foo#1
foo#2

(I also often use this trick for debugging purposes. E.g., if you have a 
function bar X with a bunch of defining equations, then simply putting 
an equation like bar X = printf "bar invoked with arg %s\n" (str X) || 
fail; in front of those equations allows you to print bar's argument 
whenever it is invoked, after which the bar X call is evaluated as usual.)

Ok, now how does the queens example work? Let's take a look at the 
crucial definition of the search function (which is invoked as search N 
1 1 [] from queens N):

search N I J P	= write P || writes "\n" if I>N;
		= search N (I+1) 1 (P++[(I,J)]) || fail if safe (I,J) P;
		= search N I (J+1) P if J<N;
		= () otherwise;

N is the size of the board (number of rows/columns), I the current row, 
J the current column and P the list of places (row/column pairs) of the 
queens determined so far (in rows 0..I-1 of the board).

The first equation simply checks whether we've already found a solution 
(I>N implies that #P=N) which is printed out.

The second equation is where the backtracking actually happens. If it is 
safe to place the next queen at position (I,J), w.r.t. the places P we 
already have, then we add (I,J) to P and invoke the algorithm 
recursively starting at column 1 in the next row. After returning from 
the recursive call, the entire subspace of the solution space with the 
given placements P++[(I,J)] has been covered. Now we let the equation 
fail, so that the interpreter moves on to the third equation to try 
other columns for the Ith row.

In any case (i.e., also if the second equation is skipped because (I,J) 
is not safe), the third equation moves on to the next column J+1 and 
invokes the algorithm recursively. This only happens if J<N, otherwise 
there are no other potential placements for the Ith row in which case 
the fourth equation returns ().

Note that the third equation is tail-recursive, i.e., it is essentially 
just a loop which iterates over the different columns for the same row 
I. The second equation is where the algorithm recursively descends into 
the solution subspace for a given initial placement.

Well, I hope that this clarifies the inner workings of the algorithm a 
bit. Tracing the algorithm for a small value of N (either on paper or 
with the interpreter's debugger) also helps. ;-)

Cheers,
Albert

-- 
Dr. Albert Gr"af
Dept. of Music-Informatics, University of Mainz, Germany
Email:  Dr....@t-..., ag...@mu...
WWW:    http://www.musikinformatik.uni-mainz.de/ag

Re: [q-lang-users] newbie question: Efficiency & other stuff

[Albert G. <Dr....@t-...>]

Arnoldo Jose Muller Molina wrote:
> dmatch(E1,E2) = (#expand(E1) + #expand(E2)) - (2 * dmatchAux(expand
> (E1),expand(E2)));

As Q functions may have side effects, it is not possible to optimize 
away shared subexpressions like this. That's what where clauses are for. 
Try something like:

dmatch(E1,E2) = (#X1 + #X2) - (2 * dmatchAux(X1,X2))
   where X1 = expand(E1), X2 = expand(E2);

> I have executed some stats commands after executing this rule, with and
> without the "(#expand(E1) + #expand(E2))" and it seems that q is
> calculating expand(E1) and expand(E2) 2 times (1). 
> 
> 1) Is this true?
> 
> 2) if that is so, is there a way of implementing a "let" so I can
> compute only once the results of expand(_)? Of course I could create an
> elaborate chain of rules and allow this, but is there a way that doesn't
> involve creating several rules? 
> 
> Thank you!
> 
> --------------------------
> (1) (expand(...) is a deterministic rule, it will always return the same
> result for the same input)

-- 
Dr. Albert Gr"af
Dept. of Music-Informatics, University of Mainz, Germany
Email:  Dr....@t-..., ag...@mu...
WWW:    http://www.musikinformatik.uni-mainz.de/ag

[q-lang-users] newbie question: Efficiency & other stuff

[Arnoldo J. M. M. <ar...@da...>]

Hello All,

I have this piece of code:

dmatch(E1,E2) = (#expand(E1) + #expand(E2)) - (2 * dmatchAux(expand
(E1),expand(E2)));

I have executed some stats commands after executing this rule, with and
without the "(#expand(E1) + #expand(E2))" and it seems that q is
calculating expand(E1) and expand(E2) 2 times (1). 

1) Is this true?

2) if that is so, is there a way of implementing a "let" so I can
compute only once the results of expand(_)? Of course I could create an
elaborate chain of rules and allow this, but is there a way that doesn't
involve creating several rules? 

Thank you!

--------------------------
(1) (expand(...) is a deterministic rule, it will always return the same
result for the same input)

[q-lang-users] Newbie: Backtracking again

[Mele, G. (MH) <Gre...@ma...>]

Thanks in advance.
 
I dont understand the queens algorithm enough to see how backtracking works
by the fail instruction.
 
I would like to have a prolog-like system:
ie the classic genealogical database
 
 
father(a,b).
father(b,c).
father(c,d).
 
ancestor(X,Y) :- father (X,Z), ancestor (Z,Y). 
 
Is this possible?
Or is there another simple example of backtracking using the fail
instruction that I could ponder over?
 
 
Greg
 
 
 
THIS E-MAIL IS CONFIDENTIAL.  If you have received this e-mail in error,
please notify us by return e-mail and delete the document.  If you are not
the intended recipient you are hereby notified that any disclosure, copying,
distribution or taking any action in reliance on the contents of this
information is strictly prohibited and may be unlawful.  Eastern Health is
not liable for the proper and complete transmission of the information
contained in this communication or for any delay in its receipt.

[q-lang-users] New file releases

[Albert G. <Dr....@t-...>]

Hi all,

as some of you might already have noticed, I have uploaded quite a few 
new packages on q-lang.sf.net. All packages have been updated for SuSE 
10.0, and new releases are available for SuperCollider (CVS 2005-11-10) 
and Faust (0.9.5). Moreover, a bugfix release of Q-Faust (1.1-2, source 
and binary RPMs) has been uploaded. As usual, you can get all the good 
new stuff from the Download page (http://q-lang.sf.net/download.html).

Please also note that I currently have no means to update the Fedora 
packages, so these are not listed on the Download page any longer. (You 
can still find the old FC1 packages in the download area at 
http://sourceforge.net/project/showfiles.php?group_id=96881, though, and 
I'll upload new Fedora packages as soon as I get around installing the 
latest FC release. In the meantime, if anyone can help providing 
up-to-date packages from some reasonably recent FC release, please let 
me know.)

Cheers,
Albert

-- 
Dr. Albert Gr"af
Dept. of Music-Informatics, University of Mainz, Germany
Email:  Dr....@t-..., ag...@mu...
WWW:    http://www.musikwissenschaft.uni-mainz.de/~ag

Re: [q-lang-users] sqlite module

[Albert G. <Dr....@t-...>]

Albert Graef wrote:
> here is an sqlite3 module which I quickly hacked together over the 
> weekend. Hope it's useful for some. Unpack, run make and you should be 
> set. A little example is included.
> 
> I'll officially release this later when I have some time...

Release 1.0 is now available from http://q-lang.sf.net.

-- 
Dr. Albert Gr"af
Dept. of Music-Informatics, University of Mainz, Germany
Email:  Dr....@t-..., ag...@mu...
WWW:    http://www.musikwissenschaft.uni-mainz.de/~ag

[q-lang-users] sqlite module

[Albert G. <Dr....@t-...>]

Hi,

here is an sqlite3 module which I quickly hacked together over the 
weekend. Hope it's useful for some. Unpack, run make and you should be 
set. A little example is included.

I'll officially release this later when I have some time...

Cheers,
Albert

-- 
Dr. Albert Gr"af
Dept. of Music-Informatics, University of Mainz, Germany
Email:  Dr....@t-..., ag...@mu...
WWW:    http://www.musikwissenschaft.uni-mainz.de/~ag

Re: [q-lang-users] Re: modules in c++

[Albert G. <Dr....@t-...>]

Brian Wilfley wrote:
> I just realized you wanted to know about g++. Its version is 3.4.2.

Ok, thanks. That seems to be a fairly new version. I just wanted to make 
sure that I don't break any recent setups with the patch.

-- 
Dr. Albert Gr"af
Dept. of Music-Informatics, University of Mainz, Germany
Email:  Dr....@t-..., ag...@mu...
WWW:    http://www.musikwissenschaft.uni-mainz.de/~ag

Re: [q-lang-users] Re: modules in c++

[Brian W. <bwi...@tr...>]

I just realized you wanted to know about g++. Its version is 3.4.2.

bpw

On 10/13/05, Albert Graef <Dr....@t-...> wrote:
> Hi Brian,
>
> I'm about to commit your suggested patch to cvs now. Just to make sure:
> Which version of mingw/g++ did you use? 2.95.x? 3.x?
>
> Brian Wilfley wrote:
> > Here's what I've figured out:
> >
> > 1) The quoted errors can be "fixed" with the following patch to libq.h:
> > ----------------- patch.libq.h ----------------------
> > *** libq.1.6.h        Wed Oct 20 01:09:42 2004
> > --- libq.h    Fri Aug 26 15:38:04 2005
> > ***************
> > *** 127,135 ****
> > --- 127,141 ----
> >
> >   /* Predefined function and type symbols. */
> >
> > + #ifdef __cplusplus
> > + __DLLIMPORT  int truesym, falsesym, nilsym, voidsym;
> > + __DLLIMPORT  int inttype, floattype, booltype, strtype, filetype,
> > +   listtype, tupletype;
> > + #else
> >   __DLLIMPORT const int truesym, falsesym, nilsym, voidsym;
> >   __DLLIMPORT const int inttype, floattype, booltype, strtype, filetype=
,
> >     listtype, tupletype;
> > + #endif
> >
> >   /* Expression construction. */
>
> --
> Dr. Albert Gr"af
> Dept. of Music-Informatics, University of Mainz, Germany
> Email:  Dr....@t-..., ag...@mu...
> WWW:    http://www.musikwissenschaft.uni-mainz.de/~ag
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by:
> Power Architecture Resource Center: Free content, downloads, discussions,
> and more. http://solutions.newsforge.com/ibmarch.tmpl
> _______________________________________________
> q-lang-users mailing list
> q-l...@li...
> https://lists.sourceforge.net/lists/listinfo/q-lang-users
>


--
Brian P. Wilfley, Ph.D.
Principal Scientist
Voice: 650.331.3476 x102
Fax: 650.887.2205
E-mail: bwi...@tr...
Triple Ring Technologies, Inc.
1850 Embarcadero Road
Palo Alto, CA 94303

Re: [q-lang-users] Re: modules in c++

[Brian W. <bwi...@tr...>]

Hi Albert,

I'm very sorry for the delay in responding. These emails get sorted
into folders, and I've not checked this one in a bit.

I rummaged around for some version information. I found the following:

in c:\MinGW\include\_mingw.h,
#define __MINGW32_VERSION 3.7

And running running the "Add/remove Programs" control panel gives
version 1.0.10 for MSYS.

Let me know if there's anything else I can help with.

Brian

On 10/13/05, Albert Graef <Dr....@t-...> wrote:
> Hi Brian,
>
> I'm about to commit your suggested patch to cvs now. Just to make sure:
> Which version of mingw/g++ did you use? 2.95.x? 3.x?
>
> Brian Wilfley wrote:
> > Here's what I've figured out:
> >
> > 1) The quoted errors can be "fixed" with the following patch to libq.h:
> > ----------------- patch.libq.h ----------------------
> > *** libq.1.6.h        Wed Oct 20 01:09:42 2004
> > --- libq.h    Fri Aug 26 15:38:04 2005
> > ***************
> > *** 127,135 ****
> > --- 127,141 ----
> >
> >   /* Predefined function and type symbols. */
> >
> > + #ifdef __cplusplus
> > + __DLLIMPORT  int truesym, falsesym, nilsym, voidsym;
> > + __DLLIMPORT  int inttype, floattype, booltype, strtype, filetype,
> > +   listtype, tupletype;
> > + #else
> >   __DLLIMPORT const int truesym, falsesym, nilsym, voidsym;
> >   __DLLIMPORT const int inttype, floattype, booltype, strtype, filetype=
,
> >     listtype, tupletype;
> > + #endif
> >
> >   /* Expression construction. */
>
> --
> Dr. Albert Gr"af
> Dept. of Music-Informatics, University of Mainz, Germany
> Email:  Dr....@t-..., ag...@mu...
> WWW:    http://www.musikwissenschaft.uni-mainz.de/~ag
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by:
> Power Architecture Resource Center: Free content, downloads, discussions,
> and more. http://solutions.newsforge.com/ibmarch.tmpl
> _______________________________________________
> q-lang-users mailing list
> q-l...@li...
> https://lists.sourceforge.net/lists/listinfo/q-lang-users
>


--
Brian P. Wilfley, Ph.D.
Principal Scientist
Voice: 650.331.3476 x102
Fax: 650.887.2205
E-mail: bwi...@tr...
Triple Ring Technologies, Inc.
1850 Embarcadero Road
Palo Alto, CA 94303

Re: [q-lang-users] newbie: backtracking?

[Albert G. <Dr....@t-...>]

Arnoldo Jose Muller Molina wrote:
> I was reading the manual and it says the matching process explicitly
> takes the first (from top to bottom) equation that matches (or the
> equation with the highest priority) . Is there a way to make q backtrack
> all the possible matchings instead of just the first? 

Yes. The builtin fail function tells the interpreter to abort the 
current equation and try the next applicable one. Please see the "8 
queens" example in examples/queens2.q and section 10.6 in the manual 
(http://q-lang.sourceforge.net/qdoc/qdoc_10.html#SEC54).

There's also a purely functional way to do this, namely to enumerate the 
entire solution space as a stream (infinite list) and then traverse that 
stream to extract the wanted solutions. See examples/queens.q for an 
example. (But the "dirty" version using fail is usually much faster.)

-- 
Dr. Albert Gr"af
Dept. of Music-Informatics, University of Mainz, Germany
Email:  Dr....@t-..., ag...@mu...
WWW:    http://www.musikwissenschaft.uni-mainz.de/~ag

[q-lang-users] newbie: backtracking?

[Arnoldo J. M. M. <ar...@da...>]

Hi All:

I was reading the manual and it says the matching process explicitly
takes the first (from top to bottom) equation that matches (or the
equation with the highest priority) . Is there a way to make q backtrack
all the possible matchings instead of just the first? 

Thanks in advance,

AM.