Thank you Juho for your help :-)
I do not succeed in installing a SBCL binary release (found at
rpmfind.net) since they all recquiert libc.so.6
when, curiously, i try rpm -ivh --test sbcl-0.X.x-i386 I get the error
Failed dependencies: libc.so.6 is needed, even if I can found (find /
-name libc) it in :
/lib/libc.so.6
/lib/i686/libc.so.6
/lib/tls/libc.so.6
/lib dir of course is in $PATH both user and root...
As you said, something seems wrong in the rh9 environnement, just
freshly installed from redhat iso cd, same i should use with the cluster
i can not update... what may be wrong ? (I tried sbcl-0.X.x-i386.rpm=20
for X from 7 to 9 and x from 5 to 11-1.fc4.i386...)
So I compiled SBCL first time again (with clean.sh) with CMUCL 19c
binary, sb-thread disabled and :sb-show enabled : it's OK, successfull.
Then I compiled again SBCL from this last built, this time with
sb-threads and :sb-show enabled : it's more verbose, but same loop for
hundreds lines :
Argh ! error on cold init, halting]
/entering ERROR, argument list=3D..
0x0500000b
/cold-printing ERROR arguments one by one..
/entering INFINITE-ERROR-PROTECTOR, *CURRENT-ERROR-DEPTH*=3D..
0x00000000
fatal error...
The system is too badly corrupted...
and this is the sb-show part (interrupted when in the loop Argh!) :
/setting the symbol-function of ~A
/setting fdefinition of ~A
/interning ~A
/interning ~A
/setting fdefinition of ~A
/interning ~A
/setting fdefinition of ~A
/interning ~A
/interning ~A
/setting fdefinition of ~A
/setting fdefinition of ~A
/interning ~A
/setting fdefinition of ~A
/interning ~A
/setting fdefinition of ~A
/interning ~A
/setting fdefinition of ~A
/entering INFINITE-ERROR-PROTECTOR, *CURRENT-ERROR-DEPTH*=3D..
0x00000000
/returning normally from INFINITE-ERROR-PROTECTOR
/back from INFINITE-ERROR-PROTECTOR
/in INFINITE-ERROR-PROTECT, incremented error depth
0: (SB-DEBUG:BACKTRACE 8 #<SYNONYM-STREAM :SYMBOL *TERMINAL-IO* {9129539}=
>)
1: (SB-INT:%BREAK SB-UNIX:SIGINT "interrupted at #X~X" 152550248)
2: (SB-UNIX::SIGINT-HANDLER
#<unused argument>
#<unused argument>
#.(SB-SYS:INT-SAP #X40249498))
3: (SB-SYS:INVOKE-INTERRUPTION #<CLOSURE (LAMBDA #) {BC2A785}>)
4: ("foreign function: call_into_lisp")
5: ("foreign function: funcall3")
6: ("foreign function: interrupt_handle_now")
7: ("foreign function: #x80525CB")
debugger invoked on a SIMPLE-CONDITION: interrupted at #X917BB68
Type HELP for debugger help, or (SB-EXT:QUIT) to exit from SBCL.
restarts (invokable by number or by possibly-abbreviated name):
0: [CONTINUE] Return from SB-UNIX:SIGINT.
1: [ABORT ] Exit debugger, returning to top level.
(SB-UNIX::SIGINT-HANDLER
#<unused argument>
#<unused argument>
#.(SB-SYS:INT-SAP #X40249498))
//getting default external format
source:
; file: SYS:SRC;CODE;TARGET-SIGNAL.LISP
/interning ~A
/interning ~A
/interning ~A
(DEFINE-SIGNAL-HANDLER SIGINT-HANDLER "interrupted" SIGINT-%BREAK)
0]
Is this usefull ?
Many thank's for your help :-)
fv
Juho Snellman a =C3=A9crit :
>fred voisin <fred@...> writes:
> =20
>
>>Thank you Juho,
>>
>>I'm using SBCL from CVS (0.9.11.8) and CMUCL 19c (took from rpm binary=
)
>>as the host compiler.
>>The first built (make.sh + test + install) of SBCL , without threads,
>>from CMUCL was fine - no problem at all.
>> =20
>>
>
>Ok, thanks for the clarifications. However, I think you missed my
>question of whether an SBCL binary release works on your
>machine. Could you check that? It's possible that something is just
>broken in your build environment. It could also be helpful to build
>with the :SB-SHOW build feature, to make it print out some debug
>output during the build process.
>
>I don't see any reason to assume that this is a problem with missing
>NPTL support. SBCL should detect that at startup, and quit gracefully.
>
> =20
>
|