[developers] PET + 64-bit architecture = int bugs?

Eric Nichols eric-n at is.naist.jp
Thu Feb 16 19:46:51 CET 2006


Greetings,

Excellent work on this release, Bernd! I was able to install without a 
hitch on my 32-bit Ubuntu machine after making a small tweak to 
configure.ac to find my itsdb libraries.

However, I get strange error messages when trying to compile on a 64-bit 
machine of the same configuration. I have attached a log to this 
message. I can get around these errors by replacing all of the pointer 
type-casts to int with type-casts to long. I am assuming the reason I 
have this problem on 64-bit and not 32-bit is due to differences in the 
size in pointers. Does anyone else experience this problem? I also 
encounter it when compiling PET 0.99.7 on 64-bit machines using the ECL 
fix mentioned the other day.

Eric

Ben Waldron wrote:
> Thanks for the new release.
>
> I was able to install pet-0.99.10 relatively painlessly on my SuSE 
> laptop (modulo an mmap problem: I had to use the KERNEL_2_4 settings 
> although I am running a 2.6.11 kernel). I had some difficulty 
> installing it on our 32- and 64-bit FC3 machines, until I realised 
> that I had omitted the '--with-cmuformat' to the ECL configure script 
> (in case anyone else does the same, you can see below what happens 
> when you try to 'make' PET in such circumstances). Would it be 
> possible to have the PET install procedure check at configure time 
> that '--cmu-format' has been compiled into the ECL used (rather than 
> waiting for the 'make' to fail)?
>
> Thanks,
> - Ben
>
> ...
> ( \
>  echo "(pushnew :i686 *features*)"; \
>  echo ";;(pushnew :debug *features*)"; \
>  echo "(load \"/home/bmw20/tmp/32/lkb//src/general/loadup\")"; \
>  echo ";;MUCH TOO DANGEROUS!!!  (si::system (format nil \"cd ~a; find 
> -type f -exec rm {} \;\" (get-binaries-dir \"mrs\")))"; \
>  echo "(setq lib-dir (make-pathname :directory '(:relative)))"; \
>  echo "(compile-system \"mrs\" :force t)"; \
> ) | /usr/groups/nltools/lingo/images/linux.x86.32//bin/ecl
> ECL (Embeddable Common-Lisp) 0.9h
> Copyright (C) 1984 Taiichi Yuasa and Masami Hagiya
> Copyright (C) 1993 Giuseppe Attardi
> Copyright (C) 2000 Juan J. Garcia-Ripoll
>        ECL is free software, and you are welcome to redistribute it
> under certain conditions; see file 'Copyright' for details.
> Type :h for Help.  Top level.
> > (:I686 :LINUX :IEEE-FLOATING-POINT :DFFI :CLOS-STREAMS :CMU-FORMAT 
> :UNIX
> :DLOPEN :CLOS :BOEHM-GC :ANSI-CL :COMMON-LISP :ECL :COMMON :ATHLON :FFI)
> > ;;; Loading #P"/home/bmw20/tmp/32/lkb/src/general/loadup.lisp"
> ;;; Loading 
> #P"/local/scratch/bmw20/tmp/32/lkb/src/general/ecl-patches.lsp"
> ;;; Loading 
> #P"/local/scratch/bmw20/tmp/32/lkb/src/general/defsystem.lisp"
> Filesystem error with pathname "SYS:cmp".
> Either
> 1) the file does not exist, or
> 2) we are not allow to access the file, or
> 3) the pathname points to a broken symbolic link.
> Broken at LAMBDA.
> MAKE>>
> g++ -g -Wall -O2 -DHAVE_CONFIG_H -DDYNAMIC_SYMBOLS -DDAG_TOMABECHI   
> -L/usr/gro ps/nltools/lingo/images/linux.x86.32//lib/ecl 
> -Wl,-rpath,/usr/groups/nltools/li go/images/linux.x86.32//lib/ecl 
> -L/usr/groups/nltools/lingo/images/linux.x86.32 /lib 
> -Wl,-rpath,/usr/groups/nltools/lingo/images/linux.x86.32//lib 
> -L/usr/group /nltools/lingo/images/linux.x86.32/lib 
> -Wl,-rpath,/usr/groups/nltools/lingo/ima es/linux.x86.32/lib  -o 
> cheap  chart.o cheap.o fs.o grammar.o item-printer.o it m.o lexicon.o 
> lexparser.o lingo-tokenizer.o morph.o options.o parse.o paths.o p 
> stags.o sm.o task.o tsdb++.o yy-tokenizer.o dag-tomabechi.o failure.o 
> qc.o cppb idge.o petecl.o petmrs.o pic-handler.o xmlparser.o  
> unicode.o psqllex.o yy.o bi code.o chunk-alloc.o dag-alloc.o 
> dag-arced.o dag-common.o dag-io.o dumper.o gra mar-dump.o hash.o 
> lex-io.o lex-tdl.o mfile.o settings.o types.o utility.o   lib rs.a 
> -lpq -lstdc++  -lpq -lxerces-c -lpthread -licuuc -licui18n -licudata 
> -lecl
> g++: libmrs.a: No such file or directory
> make[3]: *** [cheap] Error 1
> make[3]: Leaving directory 
> `/local/scratch/bmw20/tmp/x32/pet-0.99.10/cheap'
> make[2]: *** [all] Error 2
> make[2]: Leaving directory 
> `/local/scratch/bmw20/tmp/x32/pet-0.99.10/cheap'
> make[1]: *** [all-recursive] Error 1
> make[1]: Leaving directory `/local/scratch/bmw20/tmp/x32/pet-0.99.10'
> make: *** [all] Error 2
>
>
>
> Bernd Kiefer wrote:
>> Sorry about that.
>>
>> I should have run a make distcheck. I've put a new one up:
>>
>> http://www.dfki.de/~kiefer/pet-0.99.10.tar.gz
>>
>> Whew! Two releases in one day ;-)
>>
>> I've used the opportunity to change the name of config.h(.in)
>> to pet-config.h(.in) (do you know why?) and fix some messy things in
>> the configure.ac
>>
>> Best,
>>        Bernd
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: errors.txt
URL: <http://lists.delph-in.net/archives/developers/attachments/20060217/ff5cc27f/attachment.txt>


More information about the developers mailing list