Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Linux/PPC Segfault in Build

4 views
Skip to first unread message

Chromatic

unread,
Oct 5, 2006, 1:33:07 AM10/5/06
to parrot-...@perl.org
It looks like the latest coroutine changes have uncovered a bug; now I can't
get Parrot to build. Here's the relevant message:

gmake[1]: Entering directory `/home/chromatic/dev/parrot/compilers/pge'
/usr/bin/perl5.8.8 -MExtUtils::Command -e rm_rf
PGE.pbc ../../runtime/parrot/library/PGE.pbc
No root path(s) specified at /usr/lib/perl5/5.8.8/ExtUtils/Command.pm line 105
/usr/bin/perl5.8.8 -e "" >PGE/builtins_gen.pir
../../parrot -o PGE.pbc --output-pbc PGE.pir
../../parrot pgc.pir --output=PGE/builtins_gen.pir PGE/builtins.pg
gmake[1]: *** [PGE.pbc] Segmentation fault

Going into compilers/pge and massaging the Makefile not to delete the
appropriate files gets me further. Now TGE has similar problems, and I get
back to:

$ ../../parrot ../../compilers/pge/pgc.pir --output=TGE/Parser.pir
TGE/Parser.pg
Segmentation fault
$ gdb ../../parrot
(gdb) run ../../compilers/pge/pgc.pir --output=TGE/Parser.pir TGE/Parser.pg
Starting
program: /home/chromatic/bin/parrot ../../compilers/pge/pgc.pir --outpu
t=TGE/Parser.pir TGE/Parser.pg
[Thread debugging using libthread_db enabled]
[New Thread 805434208 (LWP 14756)]
[New Thread 814015728 (LWP 14759)]
[New Thread 822404336 (LWP 14760)]

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 805434208 (LWP 14756)]
0x0fd5c15c in mark_register_stack (interpreter=0x1001a008, chunk=0x101b5808)
at register.c:532
532 PObj * const obj = (PObj *)CTX_REG_STR(&ctx, i);
(gdb) bt
#0 0x0fd5c15c in mark_register_stack (interpreter=0x1001a008,
chunk=0x101b5808) at register.c:532
#1 0x0fe62ea0 in Parrot_Continuation_mark (interpreter=0x1001a008,
pmc=0x10284510) at continuation.pmc:91
#2 0x0fddec44 in Parrot_dod_trace_children (interpreter=0x1001a008,
how_many=4294966872) at dod.c:396
#3 0x0fdde92c in trace_active_PMCs (interpreter=0x1001a008, trace_stack=1)
at dod.c:317
#4 0x0fddf9cc in Parrot_dod_ms_run (interpreter=0x1001a008, flags=1)
at dod.c:959
#5 0x0fddfafc in Parrot_do_dod_run (interpreter=0x1001a008, flags=1)
at dod.c:998
#6 0x0fe1d0f4 in mem_allocate (interpreter=0x1001a008, size=4292,
pool=0x1001a9d0) at resources.c:146
#7 0x0fe1df60 in Parrot_reallocate_string (interpreter=0x1001a008,
str=0x10203418, tosize=4288) at resources.c:634
#8 0x0fd69548 in string_append (interpreter=0x1001a008, a=0x10203418,
b=0x102035f8, Uflags=1) at string.c:456
#9 0x0fd6a36c in string_concat (interpreter=0x1001a008, a=0x102034b8,
b=0x102035f8, Uflags=1) at string.c:876
#10 0x0fd69420 in string_append (interpreter=0x1001a008, a=0x102034b8,
b=0x102035f8, Uflags=1) at string.c:429
#11 0x0fef65bc in Parrot_scalar_i_concatenate_str (interpreter=0x1001a008,
#12 0x0fe0c47c in mmd_dispatch_v_ps (interpreter=0x1001a008, left=0x1020da48,
right=0x102035f8, func_nr=44) at mmd.c:476
#13 0x0fc2f2c0 in Parrot_infix_ic_p_s (cur_opcode=0x3104f30c,
interpreter=0x1001a008) at math.ops:74
#14 0x0fd70c8c in runops_slow_core (interpreter=0x1001a008, pc=0x3104f30c)
at runops_cores.c:184
#15 0x0fd53908 in runops_int (interpreter=0x1001a008, offset=3)
at interpreter.c:775
#16 0x0fd5a170 in runops (interpreter=0x1001a008, offs=3) at inter_run.c:87
#17 0x0fd5a494 in runops_args (interpreter=0x1001a008, sub=0x101a5e38,
obj=0x1005ce98, meth=0x0, sig=0xff534d8 "vP", ap=0x7ff5ed5c)
at inter_run.c:193
#18 0x0fd5a688 in Parrot_runops_fromc_args (interpreter=0x1001a008,
sub=0x101a5e38, sig=0xff534d8 "vP") at inter_run.c:295
#19 0x0fdd2cf8 in Parrot_runcode (interpreter=0x1001a008, argc=3,
argv=0x7ff5f108) at embed.c:812
#20 0x10003e28 in main (argc=3, argv=0x7ff5f108) at main.c:731
(gdb)

-- c

Bob Rogers

unread,
Oct 5, 2006, 7:31:35 AM10/5/06
to chromatic, parrot-...@perl.org
From: chromatic <chro...@wgz.org>
Date: Wed, 4 Oct 2006 22:33:07 -0700

It looks like the latest coroutine changes have uncovered a bug; now I can't
get Parrot to build. Here's the relevant message:

It's actually something I introduced (in continuation changes; coroutine
changes are still yet to come). r14845 should correct it. Sorry for
the hassle.

-- Bob

Chromatic

unread,
Oct 5, 2006, 2:08:37 PM10/5/06
to Bob Rogers, parrot-...@perl.org
On Thursday 05 October 2006 04:31, Bob Rogers wrote:

> It's actually something I introduced (in continuation changes; coroutine
> changes are still yet to come). r14845 should correct it. Sorry for
> the hassle.

Confirmed fixed! Thanks.

-- c

0 new messages