Re: [Gcl-devel] Compile failure

[Camm Maguire]

Greetings!  OK, it would now appear that we're ready on Linux x86,
though I'd always appreciate more reports from people with different
distributions.  

Given this, do you think you might have some time to analyze the gcl
memory usage, and make some proposals, Vadim?  As for myself, my plan
is to incorporate libbfd in 2.5.0 with as many working non-x86 ports
as I can, then try gcl with the Boehm conservative garbage collector,
hopefully eliminating thereby the DBEGIN/NULL_OR_ON_CSTACK stuff.

"Vadim V. Zhytnikov" <address@hidden> writes:

> Now it works fine!
> 
> Vadim
> 
> Camm Maguire wrote:
> 
> > Greetings, and thanks for the report!  OK, my mistake.  the original
> > setjmp/sigsetjmp problem I was having was due to a bug in maxima's
> > make clean.  There were some old objects in there which expected
> > __sigsetjmp when I was compiling on the new system.  I've reverted the
> > change in rsym_elf.c, so all should work now on Linux systems of all
> > flavors.  Please confirm!
> >
> > Found one other small bug in unexelf.c which was giving me erratic
> > segfaults on maxima compilation.  For any elf gurus out there, please
> > look into it and comment.  It would have seemed that no .rel or .rela
> > section header could have a sh_info of 0, but this did turn up, so I
> > guarded against it.  This field is supposed to specify the section
> > number containing the items described by the following relocations,
> > and all section numbers start with 1, to my understanding.
> >
> > Take care,
> >
> > "Vadim V. Zhytnikov" <address@hidden> writes:
> >
> > > Camm Maguire wrote:
> > >
> > > > Greetings, and thanks for your report!  Is this after my commits of
> > > > this afternoon?  What is your version of rsym_elf.c?  What is your
> > > > system and libc version?
> > > >
> > > > Take care,
> > > >
> > > > C Y <address@hidden> writes:
> > > >
> > > > > I took a shot at compiling the latest CVS, and here's what happened.
> > > > > If you want the full output I can send it, but it's rather long.  
> > > > > Sorry
> > > > > if this duplicates a problem already mentioned.
> > > > >
> > > > > CY
> > > > >
> > > > > [snip]
> > > > >
> > > > > Make of GCL 2.4.0 completed.
> > > > > make[1]: Entering directory `/home/cliff/gcl/gcl-tk'
> > > > > (cd demos ; \
> > > > > echo '(load "../tkl.o")(TK::GET-AUTOLOADS (directory "*.lisp"))' |
> > > > > ../../unixport/saved_gcl)
> > > > > GCL (GNU Common Lisp)  Version(2.4.0) Fri Jan  4 15:45:09 CST 2002
> > > > > Licensed under GNU Library General Public License
> > > > > Contains Enhancements by W. Schelter
> > > > >
> > > > > >
> > > > > Loading ../tkl.o
> > > > > symbol "_setjmp" is not in base imagestart address -T 82ac000 Finished
> > > > > loading ../tkl.o
> > > > > 46432
> > > > >
> > > > > >
> > > > > Error: Caught fatal error [memory may be damaged]
> > > > > Fast links are on: do (si::use-fast-links nil) for debugging
> > > > > Error signalled by EVAL.
> > > > > Broken at TK::GET-AUTOLOADS.  Type :H for Help.
> > > > > >>make[1]: Leaving directory `/home/cliff/gcl/gcl-tk'
> > > > > (cd info ; make)
> > > > > make[1]: Entering directory `/home/cliff/gcl/info'
> > > > > make[1]: Nothing to be done for `all'.
> > > > > make[1]: Leaving directory `/home/cliff/gcl/info'
> > > > >
> > >
> > > I've similar but a bit different problem with GCL
> > > with latest fixes. Now it builds on both my systems
> > > (glibc 2.2.4 and 2.2.3). Maxima also builds on it
> > > without any visible trouble but on both systems
> > > Maxima tests dump core with the following message
> > >
> > > %Input is
> > > NUSUM(N N!, N, 0, N)
> > > symbol "_setjmp" is not in base image
> > > Unrecoverable error: Segmentation violation..
> > >
> > > It seems that something is wrong with _setjmp.
> > > Notice, that before you did this _setjmp modification
> > > I had not any errors or warnings about _sigsetjmp.
> > > But as far as I understand you experienced problems
> > > with _sigsetjmp on your 2.2.4 machine?
> > >
> > > Vadim
> > >
> > > --
> > >
> > > [ Vadim V. Zhytnikov  <address@hidden>  <address@hidden> ]
> > >
> > >
> > >
> >
> > --
> > Camm Maguire                                            address@hidden
> > ==========================================================================
> > "The earth is but one country, and mankind its citizens."  --  Baha'u'llah
> >
> > _______________________________________________
> > Gcl-devel mailing list
> > address@hidden
> > http://mail.gnu.org/mailman/listinfo/gcl-devel
> 
> --
> 
> [ Vadim V. Zhytnikov  <address@hidden>  <address@hidden> ]
> 
> 
> 
> 
> 

-- 
Camm Maguire                                            address@hidden
==========================================================================
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah