Ciao-users September 2002

ciao-users@cliplab.org

5 participants
5 discussions

Solaris environment, compile problems, setenv() and friends
by Vasilios Hoffman 02 Oct '02

02 Oct '02

Hi, Sorry to bother you folks, but I'm building ciao-1.8p2 on a solaris/sparc 5.7 platform with gcc 2.95.3, and there is a surprising error. The make fails because of the following error: gcc wam.o format.o inout.o misc.o startgoal.o ciao.o nondet.o objareas.o streams.o timing.o wambuiltin.o wamfunction.o unix_utils.o attr.o dynlink.o tasks.o locks.o prolog_tasks.o own_malloc_linear.o alloc.o heapgc.o indexing.o initial.o registers.o stacks.o support.o term_support.o wamsupport.o qread.o qget.o builtin.o interrupt.o gauge.o bignum.o profile.o ciao_prolog.o start.o version.o sockets_c.o random.o main.o -ldl -lm -lnsl -lpthread -lrt -lsocket -o ciaoengine Undefined first referenced symbol in file setenv unix_utils.o ld: fatal: Symbol referencing errors. No output written to ciaoengine which is definately because unix_utils.c is using setenv() -- and that's not present in solaris C libraries (at least through solaris 8, I haven't checked solaris 9). It's not present in a lot of C libraries, so I'm very surprised that it's used when it is detecting a "SolarisSparc" system -- which it does, btw, for example: if test ! -d /usr/local/src/ciao-1.8p2/bin/SolarisSparc ; then \ mkdir /usr/local/src/ciao-1.8p2/bin/SolarisSparc ; chmod 775 /usr/local/src/ciao-1.8p2/bin/SolarisSparc ; fi make[2]: Leaving directory `/usr/local/src/ciao-1.8p2' cd /usr/local/src/ciao-1.8p2/bin/SolarisSparc && \ ln -s ../../engine/*.[ch] . && \ ln -s ../../engine/Makefile . && \ rm -f configure.h make createincludedir is right in the beginning of the output of make. So, any ideas? All I did was edit SETTINGS and run make. The system is fairly gnu-ified, and it is used for compiling a diverse assortment of things, so most tools are there and/or fairly up-to-date (gcc2.95.3 versus newer gcc3+ is a topic for another day). If there's some patch I need, or something I need to edit, please let me know! I don't want to mess around any more, because it's such an obvious bug that I have to believe it's something stupid on my end... Thanks, -V Hoffman Vasilios Hoffman Junior Unix Administrator Wesleyan University voice (860) 685-3142 fax (860) 685-2401 ============================================================================== Message: Address: Action: help majordomo(a)clip.dia.fi.upm.es Info. on useful commands subscribe ciao-users-request(a)clip.dia.fi.upm.es Subscribe to this list unsubscribe ciao-users-request(a)clip.dia.fi.upm.es Unsubscribe from this list <whatever> ciao-users(a)clip.dia.fi.upm.es Send message to list ----------------------------------------------------------------------------- Archived messages: http://www.clip.dia.fi.upm.es/Mail/ciao-users/ -----------------------------------------------------------------------------

2 3

Re: horse_riding and MYCIN
by John O'Gorman 30 Sep '02

30 Sep '02

Thanks to Paco Bueno's changes to mycin_tr.pl, I can now use the example horse_riding. As explained by Paco, mycin is a package. At compile-time it rewrites its clauses into a different form. This provokes the question: Is is possible to prompt for some of the properties (like sunny(yes) cf 0.6) and then submit them from the top level to mycin. I am trying to build a small expert system (for claims processing). So I want to use something like the Solve program in Art of Prolog (p 312) to get the answers to a series of questions - assert a set of statements of the form: property(x) cf CF, then query: ?- file:mycin(payout(x),CF). Can this be done simply and directly? Or will I have to write code to 1. Prompt for the answers (where necessary) 2. Build a file (like horse_riding.pl) and compile it. 3. use_module the new file. 4. delete the file. John O'Gorman Francisco Bueno Carrillo wrote: > > The mycin library is a package (also assertions). The use of packages > can be declared in three ways: > > :- use_package(mycin). > :- module(MyModule,[...],[mycin]). > :- mycin(MyModule). > > As mentioned, :- mycin(MyModule). is just a way of declaring the use > of the mycin package within a file. It is not a call to a > predicate. So, in fact, there is no much use in calling it from the > top level. > > Paco Bueno ============================================================================== Message: Address: Action: help majordomo(a)clip.dia.fi.upm.es Info. on useful commands subscribe ciao-users-request(a)clip.dia.fi.upm.es Subscribe to this list unsubscribe ciao-users-request(a)clip.dia.fi.upm.es Unsubscribe from this list <whatever> ciao-users(a)clip.dia.fi.upm.es Send message to list ----------------------------------------------------------------------------- Archived messages: http://www.clip.dia.fi.upm.es/Mail/ciao-users/ -----------------------------------------------------------------------------

1 0

Re: horse_riding and MYCIN -> the file
by John O'Gorman 25 Sep '02

25 Sep '02

Paco Thanks. 1 step forward - but still no go! when I try the sequence (in ...mycin/examples): ciao ?-use_module(horse_riding). yes ?-horse_riding(mycin(winner(C),P). {ERROR: horse_riding:weather/1 - undefined predicate} The above line is repeated 9 times. Similarly if I try: ?-use_module(age). I get a message {compiling /opt/ciao/...../examples/age.pl WARNING: (lns 11-12_ Predicate age/2 undefined in source --- above line 4 times ---- } When I try: ?-use_module(library(mycin)). {Reading /usr/lib/.....mycin/mycin.itf ERROR: Source used as module without module declaration } Should I do some sort of recompile? (perhaps rm the .po .itf files, and run make from high up in the tree). The documentation comments in file mycin_doc.pl says to use as first statememt in a file: :-mycin(MyDatabaseName). Is this right? If is the first statement (without a use_module(library(mycin))) or use_module(library(assertion)), then how can the :-mycin statement work? If running at the top level, do you call the :-mycin(X) as a predicate? ?-mycin(C). Also what does the mycin directive do? Is it meant to called with an atom :-mycin(john). or a var :-mycin(John). ??? John O'Gorman Francisco Bueno Carrillo wrote: > > John, > > Here is the promised update. You have to replace the file > contrib/mycin/mycin_tr.pl in your Ciao distribution by the following > code. > > Best, > > Paco Bueno > ============================================================================== Message: Address: Action: help majordomo(a)clip.dia.fi.upm.es Info. on useful commands subscribe ciao-users-request(a)clip.dia.fi.upm.es Subscribe to this list unsubscribe ciao-users-request(a)clip.dia.fi.upm.es Unsubscribe from this list <whatever> ciao-users(a)clip.dia.fi.upm.es Send message to list ----------------------------------------------------------------------------- Archived messages: http://www.clip.dia.fi.upm.es/Mail/ciao-users/ -----------------------------------------------------------------------------

2 1

horse_riding and MYCIN
by John O'Gorman 24 Sep '02

24 Sep '02

Hi everyone. I am brand new to ciao. I am trying to get started with a MYCIN style program. The horse_riding example in contrib/mycin/examples looks good. But I get errors relating to modules, exports, etc. I programmed in prolog years ago, but I am unfamiliar with the ciao module and compilation processes. Would some kind soul please give me the minimum needed to call horse_riding:winner(X,P) from a. ciao shell b. compiling with ciaoc Also, I am puzzled that horse_riding seems to be defined as winner/1 bvt the author suggests calling it as winner/2. Is this an error or is the Certainty Factor variabel an implicit 2nd arg? John O'Gorman ============================================================================== Message: Address: Action: help majordomo(a)clip.dia.fi.upm.es Info. on useful commands subscribe ciao-users-request(a)clip.dia.fi.upm.es Subscribe to this list unsubscribe ciao-users-request(a)clip.dia.fi.upm.es Unsubscribe from this list <whatever> ciao-users(a)clip.dia.fi.upm.es Send message to list ----------------------------------------------------------------------------- Archived messages: http://www.clip.dia.fi.upm.es/Mail/ciao-users/ -----------------------------------------------------------------------------

2 1

Re: Solaris environment, compile problems, setenv() and friends
by Richard A. O'Keefe 24 Sep '02

24 Sep '02

Vasilios Hoffman <vhoffman01(a)wesleyan.edu> wrote: that building ciao-1.8p2 on SPARC Solaris 5.7 ran into a problem with setenv(). Is setenv() a Linuxism? There is certainly no such function in Solaris or OSF/1, although they do have a putenv("VAR=VALUE") function. I note that there are warnings in the documentation against using putenv() in multithreaded code, for the obvious reasons, and that the OSF/1 manual page is flat-out WRONG in what it says putenv() does. ============================================================================== Message: Address: Action: help majordomo(a)clip.dia.fi.upm.es Info. on useful commands subscribe ciao-users-request(a)clip.dia.fi.upm.es Subscribe to this list unsubscribe ciao-users-request(a)clip.dia.fi.upm.es Unsubscribe from this list <whatever> ciao-users(a)clip.dia.fi.upm.es Send message to list ----------------------------------------------------------------------------- Archived messages: http://www.clip.dia.fi.upm.es/Mail/ciao-users/ -----------------------------------------------------------------------------

1 0

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

Ciao-users September 2002