Re: build segfault (2.31)

[<don...@is...>]

Sam Steingold writes:
 > zbacktrace will give you the CLISP backtrace.
Does that help me find the segfault?

 > what are k1, k2, n?
Below.  What do they mean?  Again, what's the relation to segfault?

(gdb) zbacktrace
Undefined command: "zbacktrace".  Try "help".
(gdb) p back_trace_out(0,0)
[0/0xBFFFAC4C]> #<COMPILED-CLOSURE COMMON-LISP::COMPILE-FILE> delta: STACK=11; SP=24
[1/0xBFFFADD8]> #<SYSTEM-FUNCTION SYSTEM::READ-EVAL-PRINT> delta: STACK=5; SP=24
[2/0xBFFFAF60]> #<SPECIAL-OPERATOR COMMON-LISP::WHEN> delta: STACK=7; SP=33
[3/0xBFFFB178]> #<SPECIAL-OPERATOR COMMON-LISP::CATCH> delta: STACK=18; SP=36
[4/0xBFFFB3B8]> #<CLOSURE :LAMBDA> 0 args delta: STACK=2; SP=22
[5/0xBFFFB520]> #<SYSTEM-FUNCTION SYSTEM::DRIVER> delta: STACK=4; SP=24
[6/0xBFFFB6A8]> #<SPECIAL-OPERATOR COMMON-LISP::PROGN> delta: STACK=17; SP=36
[7/0xBFFFB8E8]> #<CLOSURE SYSTEM::MAIN-LOOP> 0 args delta: STACK=0; SP=20
[8/0xBFFFBA30]> #<SYSTEM-FUNCTION SYSTEM::DRIVER>
$1 = 9
(gdb) p k1
$2 = 2
(gdb) p k2
$3 = 1
(gdb) p n
$4 = 47
(gdb) 

 > can you reproduce this when you build with g++ and DEBUG_GCSAFETY?
 > this might catch the original problem early on.
First I've heard of it.  Is there some configure option I should try
or you mean export CC="gcc -DDEBUG_GCSAFETY" ?
And what does this do?

 > (I remember you had problems with g++, did you manage to build with g++?)
Not even close.  I thought you had the same problems.