[M3devel] compile problem on LINUXLIBC6

Daniel Alejandro Benavides D. dabenavidesd at yahoo.es
Sun Jan 23 08:08:29 CET 2011


hi all:
yes I think you are correct in that too, ok. 
About your time/effort paradigm.
Let me tell you something if you allow me: here in my country I was talking with another co partner, here, we have this kind of thinking : "you don't need to be talented, you just need to work hard", and certainly I have read lot of times perhaps that the harder things are, more worth they are.
Certainly both of them are surely right, but let me tell you what a "neuro linguist" programmer says about that (I don't remember his name though he gives international conferences): "most of people think that the harder you work the more work you get done, but I tell you something is better to work intelligently than harder, if you are intelligent worker you get things done and are more worth the value, this is a paradigm you need to work your entire life to get something done and you realize how much trouble you can get solve if you are smarter than hard worker".

I think also is true, the last point I mean, why we should spend the all night reading other people's code, even if they are worth, we should be better than that. My point is this, ok let's do some hard work in an intelligent way, we will be smarter than others by following the smarter things they had on harder things, but doing hard work on it if you don't need that last thing then you have a point.

Then now, in this thing, let's be clear IA prominent guys are working towards making and u turn (maybe just and an "ear" turn here we say too). If I interpret this guys well, without too much knowledge of the ideas I do now something; at some point they should be going a little backwards to start over from where they went to work in fields that are not worth today, you see, a lot of work doesn't imply worth results even if finished, but intelligent work even if not absolutely complete is more worth the value than those they did.

So I was reading a Computer magazine, March 1982, a local copy courtesy of somebody here. And what where doing Japanese Fifth Computer Generation Computer Systems plans at those days, planning the according following next decade of computing, the computers fifth generation FGCS.

Basically what they were trying to do in order to do the major tasks, to build a complete framework up to a complete stack of technology in some of its subtasks (I know, but why should we worry if others are working nicely on the other parts, tell me if so): technology systematization, which compromised the following points:

"Specifically this involve:
(1) Modular programming and verification technology
(2) A theory to specify description and program synthesis
(3) A system for program synthesis, and program verfication and a program base
(4) A system to maintain, improve and manage programs
(5) A consultant system for program design 
"
____________SYSTEMATIZATION TECHNOLOGY_________________
Intelligent          | Knowledge |  Systematization | Database
programming   | base design | for computer      | and distributed
system               |  system       |   architecture      | database system
________________________________________________________________
in 
[1] Treleaven and Lima, “Japan's Fifth Generation Computer Systems,” Computer, vol. 15, no. 8, pp. 79-88, 1982.

Basically they wanted to build automatic language translators for languages like japanese to english, etc, a system to receive speech, and system to retrieve images. They main problem was that when arrived in the next decade they didn't get the expected rate or objective of efficiency was:

1-10 BIPS with a computer of 1G-10G of memory =  (though first goal was 1200'000.000 logical inferences per second, LIPS. 1 LIPS = [100..300] MIPS) = 400 000 MIPS and in full goal = 333'333.333 LIPS 

however they got was for first goal:
30.000 LIPS = 100 MIPS.
The long term goal was not achieved but because of technical problem in the theory and semantic of concurrent languages.
You can read about this on 
http://en.wikipedia.org/wiki/Fifth_generation_computer
and:
http://books.google.com/books?id=BlRrHHpoxLgC&lpg=PA14&ots=ArxnlvBiyx&dq=inference%20machine%20-learning&pg=PA15#v=onepage&q&f=false
But for the long term project was getting just 200 000 LIPS < 700 MIPS
 
Obviously this is not god like 100 times slower than thought, so they stopped funding the project and continue a sequel in another one.

And NEC was one of the participants, precisely they wrote back in 1995 a bayesian algorithm in  Modula-3, looking for a chess contest (see 
http://www.nzdl.org/gsdlmod?e=d-00000-00---off-0tcc--00-0----0-10-0---0---0direct-10---4-------0-1l--11-en-50---20-about---00-0-1-00-0-0-11-1-0utfZz-8-10-0-0-11-1-0utfZz-8-00&a=d&c=tcc&cl=CL2.5&d=HASH01ddef928d9edae8cb47030b.5)

Surely you are worried about this kind of issues, but I tell you I think we have a point if so, there are good opportunities they are just waiting for us and good looking for good chance of making it.

Thanks in advance









--- El sáb, 22/1/11, Jay K <jay.krell at cornell.edu> escribió:

De: Jay K <jay.krell at cornell.edu>
Asunto: RE: [M3devel] compile problem on LINUXLIBC6
Para: "Mika Nystrom" <mika at async.caltech.edu>, dabenavidesd at yahoo.es
CC: "m3devel" <m3devel at elegosoft.com>
Fecha: sábado, 22 de enero, 2011 14:44




libcpp is part of any gcc frontend/backend. I have reduced it significantly, but it is still there.


Maybe it was caused by this:

2010-12-26 18:50  jkrell

  * m3-sys/m3cc/src/clean_marker.txt:

  system.h: all systems have:
   stddef.h
   limits.h
   abort
   realloc
...

anyway, I made a change that should resolve this. Let me know.
I'll try to keep up with Hudson, but things have changed
semi-permanently very significantly for me..less time on Modula-3 probably forever now...
(I'll still try to fix the jumpbuf over-use and maybe stack-walker absence...)


 - Jay

> To: dabenavidesd at yahoo.es
> Date: Sat, 22 Jan 2011 08:31:29 -0800
> From: mika at async.caltech.edu
> CC: m3devel at elegosoft.com
> Subject: Re: [M3devel] compile problem on LINUXLIBC6
> 
> In order to build m3cc, all I had to do was remove the declaration for
> basename() from /usr/cvs/cm3/m3-sys/m3cc/gcc-4.5/include/libiberty.h
> 
> I find it a bit odd since my system uses gcc 4.5.0 but am not quite sure
> what should be done about it.
> 
>     Mika
> 
> "Daniel Alejandro Benavides D." writes:
> >Hi all:
> >I wonder why it's making libcpp, does it some tool use it? just for c this =
> >days? anyway it would be good if you can test it just with another compiler=
> > version, perhaps the clash is product of incompability of version 4.50, 4.=
> >5.1 -><- perhaps you can try to build the gcc trunk to see how it builds, i=
> >f it builds then you have a problem of compability in your gcc headers, Thi=
> >s is why is this C code is sort of not portable, you don't know when you ge=
> >t a compiler error after thinking it is OK.
> >Then if so, try to use gcc trunk instead to build in your cm3 building.
> >Thanks in advance
> >
> >--- El vie, 21/1/11, Mika Nystrom <mika at async.caltech.edu> escribi=F3:
> >
> >> De: Mika Nystrom <mika at async.caltech.edu>
> >> Asunto: [M3devel] compile problem on LINUXLIBC6
> >> Para: m3devel at elegosoft.com
> >> Fecha: viernes, 21 de enero, 2011 21:15
> >>=20
> >> On the following system:
> >>=20
> >> Linux ballista 2.6.34.7-0.5-desktop #1 SMP PREEMPT
> >> 2010-10-25 08:40:12 +0200 i686 i686 i386 GNU/Linux
> >>=20
> >>  gcc -v
> >> Using built-in specs.
> >> COLLECT_GCC=3Dgcc
> >> COLLECT_LTO_WRAPPER=3D/usr/lib/gcc/i586-suse-linux/4.5/lto-wrapper
> >> Target: i586-suse-linux
> >> Configured with: ../configure --prefix=3D/usr
> >> --infodir=3D/usr/share/info --mandir=3D/usr/share/man
> >> --libdir=3D/usr/lib --libexecdir=3D/usr/lib
> >> --enable-languages=3Dc,c++,objc,fortran,obj-c++,java,ada
> >> --enable-checking=3Drelease
> >> --with-gxx-include-dir=3D/usr/include/c++/4.5 --enable-ssp
> >> --disable-libssp --disable-plugin --with-bugurl=3Dhttp://bugs.opensuse.or=
> >g/ --with-pkgversion=3D'SUSE
> >> Linux' --disable-libgcj --disable-libmudflap
> >> --with-slibdir=3D/lib --with-system-zlib --enable-__cxa_atexit
> >> --enable-libstdcxx-allocator=3Dnew --disable-libstdcxx-pch
> >> --enable-version-specific-runtime-libs --program-suffix=3D-4.5
> >> --enable-linux-futex --without-system-libunwind
> >> --enable-gold --with-plugin-ld=3D/usr/bin/gold
> >> --with-arch-32=3Di586 --with-tune=3Dgeneric
> >> --build=3Di586-suse-linux
> >> Thread model: posix
> >> gcc version 4.5.0 20100604 [gcc-4_5-branch revision 160292]
> >> (SUSE Linux)
> >>=20
> >> Installing 5.8.6 from the web and then attempting to
> >> compile the CVS head using "upgrade.py" leads
> >> to the following:
> >>=20
> >> echo ./regex.o ./cplus-dem.o ./cp-demangle.o ./md5.o
> >> ./sha1.o ./alloca.o ./argv.o ./choose-temp.o ./concat.o
> >> ./cp-demint.o ./crc32.o ./dyn-string.o ./fdmatch.o
> >> ./fibheap.o ./filename_cmp.o ./floatformat.o ./fnmatch.o
> >> ./fopen_unlocked.o ./getopt.o ./getopt1.o ./getpwd.o
> >> ./getruntime.o ./hashtab.o ./hex.o ./lbasename.o
> >> ./lrealpath.o ./make-relative-prefix.o ./make-temp-file.o
> >> ./objalloc.o ./obstack.o ./partition.o ./pexecute.o
> >> ./physmem.o ./pex-common.o ./pex-one.o ./pex-unix.o
> >> ./safe-ctype.o ./sort.o ./spaces.o ./splay-tree.o
> >> ./strerror.o ./strsignal.o ./unlink-if-ordinary.o
> >> ./xatexit.o ./xexit.o ./xmalloc.o ./xmemdup.o ./xstrdup.o
> >> ./xstrerror.o ./xstrndup.o > required-list
> >> make[2]: Entering directory
> >> `/home/mika/cm3-cvs/cm3/m3-sys/m3cc/LINUXLIBC6/libiberty/testsuite'
> >> make[2]: Nothing to be done for `all'.
> >> make[2]: Leaving directory
> >> `/home/mika/cm3-cvs/cm3/m3-sys/m3cc/LINUXLIBC6/libiberty/testsuite'
> >> make[1]: Leaving directory
> >> `/home/mika/cm3-cvs/cm3/m3-sys/m3cc/LINUXLIBC6/libiberty'
> >> cd . && cd libcpp && make=20
> >> MAKE=3Dmake  AUTOCONF=3D: AUTOMAKE=3D: LEX=3D'touch lex.yy.c'
> >> MAKEINFO=3D: libcpp.a
> >> cd . && cd libcpp && make=20
> >> MAKE=3Dmake  AUTOCONF=3D: AUTOMAKE=3D: LEX=3D'touch lex.yy.c'
> >> MAKEINFO=3D: libcpp.a
> >> g++  -I../../gcc-4.5/libcpp -I.
> >> -I../../gcc-4.5/libcpp/../include
> >> -I../../gcc-4.5/libcpp/include  -g -O2 -W -Wall
> >> -Wwrite-strings -Wmissing-format-attribute -pedantic
> >> -Wno-long-long  -I../../gcc-4.5/libcpp -I.
> >> -I../../gcc-4.5/libcpp/../include
> >> -I../../gcc-4.5/libcpp/include  -c -o charset.o -MT
> >> charset.o -MMD -MP -MF .deps/charset.Tpo
> >> ../../gcc-4.5/libcpp/charset.c
> >> In file included from ../../gcc-4.5/libcpp/system.h:335:0,
> >>             =20
> >>    from ../../gcc-4.5/libcpp/charset.c:24:
> >> ../../gcc-4.5/libcpp/../include/libiberty.h:106:36: error:
> >> new declaration =E2
> >> /usr/include/string.h:601:28: error: ambiguates old
> >> declaration =E2
> >> make: *** [charset.o] Error 1
> >> "/home/mika/cm3-cvs/cm3/m3-sys/m3cc/src/m3makefile", line
> >> 273: quake runtime error: exit 2: cd . && cd libcpp
> >> && make  MAKE=3Dmake  AUTOCONF=3D: AUTOMAKE=3D:
> >> LEX=3D'touch lex.yy.c' MAKEINFO=3D: libcpp.a
> >>=20
> >> --procedure--  -line-  -file---
> >> exec          =20
> >>    --  <builtin>
> >> m3cc_Run          273=20
> >> /home/mika/cm3-cvs/cm3/m3-sys/m3cc/src/m3makefile
> >> include_dir       518=20
> >> /home/mika/cm3-cvs/cm3/m3-sys/m3cc/src/m3makefile
> >>               =20
> >>     5=20
> >> /home/mika/cm3-cvs/cm3/m3-sys/m3cc/LINUXLIBC6/m3make.args
> >>=20
> >> Fatal Error: package build failed
> >>  *** execution of [<function _BuildGlobalFunction at
> >> 0x814db54>, <function _ShipFunction at 0x814db8c>]
> >> failed ***
> >>=20
> >> It looks like a clash between a declaration in libiberty.h
> >> and one in string.h
> >>=20
> >>      Mika
> >> =0A=0A=0A      
 		 	   		  



      
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://m3lists.elegosoft.com/pipermail/m3devel/attachments/20110123/321394a0/attachment-0002.html>


More information about the M3devel mailing list