Re: [Gcl-devel] Re: [Axiom-developer] Failure in Solaris 9 build, patch 23

[Camm Maguire]

Greetings!

"Kostas Oikonomou" <address@hidden> writes:

> Hi Camm,
> 
> Here is the status:
> 
> 1) I built GCL with --disable-statsysbfd --enable-locbfd.
> 
> =================================================================================
> 
> 2) The error remains:
> 
> ...
> Loading /home/build/axiom--main--1--patch-23/obj/sol9gcc/interp/interop.o
> start address -T cbe000 Finished loading 
> /home/build/axiom--main--1--patch-23/obj/sol9gcc/interp/interop.o
> Loading /home/build/axiom--main--1--patch-23/obj/sol9gcc/interp/patches.o
> start address -T 63aaa0 Finished loading 
> /home/build/axiom--main--1--patch-23/obj/sol9gcc/interp/patches.o
> start address -T 539ec8    Using local database 
> /home/build/axiom--main--1--patch-23/src/share/algebra/compress.daase..   
> Re-reading compress.daase   Using local database 
> /home/build/axiom--main--1--patch-23/src/share/algebra/interp.daase..   
> Re-reading interp.daase
>     AXIOM initialization: interpreter
> start address -T b45948    AXIOM initialization: database
>     AXIOM initialization: constructors
>     AXIOM initialization: history
>                          AXIOM Computer Algebra System
>                     Version: Axiom 3.0 Beta (January 2005)
>                Timestamp: Wednesday January 5, 2005 at 09:29:42
> -----------------------------------------------------------------------------
>     Issue )copyright to view copyright notices.
>     Issue )summary for a summary of useful system commands.
>     Issue )quit to leave AXIOM and return to shell.
> -----------------------------------------------------------------------------
> 
> 
> Error:  Lisps arglist maximum surpassed
> Fast links are on: do (si::use-fast-links nil) for debugging
> Error signalled by APPLY.
> Broken at APPLY.  Type :H for Help.
> BOOT>>make[3]: *** 
> [/home/build/axiom--main--1--patch-23/obj/sol9gcc/bin/interpsys] Error 255

Good! I think we can be confident in your system bfd setup then.  My
experiences with solaris to the contrary must therefore be a local
situation.

Take care,

> make[3]: Leaving directory `/home/build/axiom--main--1--patch-23/src/interp'
> make[2]: *** [interpdir] Error 2
> make[2]: Leaving directory `/home/build/axiom--main--1--patch-23/src'
> make[1]: *** [srcdir] Error 2
> make[1]: Leaving directory `/home/build/axiom--main--1--patch-23'
> make: *** [all] Error 2
> bash-2.05$
> bash-2.05$
> 
> =================================================================================
> 
> 3)  The foo/bar test you suggested:
> bash-2.05$ obj/sol9gcc/bin/lisp
> 
> > (defun foo (x) (* x x))
> FOO
> > (compile 'foo)
> Compiling gazonk0.lsp.
> End of Pass 1.
> End of Pass 2.
> OPTIMIZE levels: Safety=0 (No runtime error checking), Space=0, Speed=3
> Finished compiling gazonk0.lsp.
> Loading gazonk0.o
> start address -T 7c9f10 Finished loading gazonk0.o
> #<compiled-function FOO>
> > (si::save-system "/tmp/bar")
> bash-2.05$ /tmp/bar
> > (foo 2)
> 4
> > (quit)
> bash-2.05$
> 
> =================================================================================
> 
> 4)  I can't seem to find these problems in the make output.
> But I don't totally know what to look for.
> 
> > Also noticed that your util.lisp is loaded into the interpreter
> > (i.e. non-compiled), the only difference I can see with my last
> > successful build log.  Also, there was a failure to find a package
> > file earlier on, the resolution of which is unclear from the output.

The long sequence of loads before the error shows util.lisp, not
util.o as it does for me.  Matter of fact, this is probably the cause
of the discrepancy between linux and solaris -- GCL's compiler will
inline the apply and circumvent the arglist maximum thereby.


> 
> =================================================================================
> 
> 5) I don't know how to do what you are describing here.
> In obj/sol9gcc/bin I have bootsys,  depsys, and  lisp.  As a matter of
> fact, there is no "interpsys" anywhere, despite this line from make
> 
> 617 the database was updated...remaking interpsys
> 
> which might be taken to mean that there is or was an "interpsys" somewhere.
> 
> > Barring bfd, the next step is to extract the input being fed to
> > (interpsys?) in the error below, and run it by hand thus:
> >
> > .....interpsys
> > )lisp (si::use-fast-links nil)
> > )lisp (load "input file.lisp")
> >>> :bt  (backtrace)
> 
>                                       Kostas
> 
> 
> 

Take care,

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