Unusual gmake bug

[nigel]

        Hi. What I think is an interesting error with a dying child:

% uname -a    
OSF1 goanna.ind.tansu.com.au V4.0 1530 alpha

% gmake -v
GNU Make version 3.79.1, by Richard Stallman and Roland McGrath.
Built for alphaev56-dec-osf4.0g
Copyright (C) 1988, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 2000
        Free Software Foundation, Inc.
This is free software; see the source for copying conditions.
There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A
PARTICULAR PURPOSE.

Report bugs to <address@hidden>.


% gmake 
/proj/rscl/tools/ose/4.3pl21-ind2/include/makeit/C.mk:485: warning: overriding 
commands for target `ALPHA_OSF.DEC6.3_dbg/tcaptest.o'
/proj/rscl/tools/ose/4.3pl21-ind2/include/makeit/cc.mk:520: warning: ignoring 
old commands for target `ALPHA_OSF.DEC6.3_dbg/tcaptest.o'
/proj/rscl/tools/ose/4.3pl21-ind2/include/makeit/C.mk:788: warning: overriding 
commands for target `ALPHA_OSF.DEC6.3_dbg/tcaptest.d'
/proj/rscl/tools/ose/4.3pl21-ind2/include/makeit/cc.mk:803: warning: ignoring 
old commands for target `ALPHA_OSF.DEC6.3_dbg/tcaptest.d'
/proj/be/release/V0_1_0/makeit/beModules.mk:978: warning: overriding commands 
for target 
`/proj/users/nigel/telain/telainInfras/V0_7_0/ALPHA_OSF/include/tcaptest.h'
/proj/telain/telainInfras/release/V0_7_0/src/id_InapDec/snacc.mk:189: warning: 
ignoring old commands for target 
`/proj/users/nigel/telain/telainInfras/V0_7_0/ALPHA_OSF/include/tcaptest.h'
==================================================
=== Starting export.include.target: .
=== Started at: Fri Jul 20 13:39:06 EST 2001 (NO PATCHES)
======================
/proj/telain/telainInfras/release/V0_7_0/ALPHA_OSF/DEC6.3/bin_dbg/snacc -C -u 
/proj/telain/telainInfras/release/V0_7_0/share/cfg/useful.asn1  tcaptest.asn1
gmake: zsh: 27404 segmentation fault (core dumped)  gmake

%


        Now, if I poke around in dbx, the culprit seems to be strlen:

% dbx `which gmake` core
dbx version 3.11.10
Type 'help' for help.
Core file created by program "gmake"

signal Segmentation fault at >*[strlen, 0x3ff8019f9f0]  ldq_u   r1, 0(r16)
(dbx) where
>  0 strlen(0x0, 0x14008e4d0, 0x0, 0x14000f4c0, 0x11fffca80) [0x3ff8019f9f0]
   1 _doprnt_dis(0xb, 0x100000007, 0x11fffd130, 0x100000010, 
0x7ff0000000000000) [0x3ff801a2af8]
   2 _doprnt(0x7ff0000000000000, 0xe000, 0x3ff8014c350, 0x11fffd1c0, 
0x500000010) [0x3ff800dbeb8]
   3 vfprintf(0x12001b14c, 0x0, 0x11fffd1c0, 0x500000010, 0x1400013f0) 
[0x3ff8014c34c]
   4 error(flocp = (nil), fmt = 0x1400013f0 = "*** [%s] %s%s", va_alist = 
5368855952) ["misc.c":272, 0x12001b148]
   5 child_error(target_name = 0x140023d90 = 
"ALPHA_OSF.DEC6.3_dbg/tcaptest.cc", exit_code = -1, exit_sig = 11, coredump = 
128, ignored = 0) ["job.c":262, 0x120013164]
   6 reap_children(block = 1, err = 0) ["job.c":667, 0x12001384c]
   7 new_job(file = 0x140029f20) ["job.c":1525, 0x120014d60]
   8 execute_file_commands(file = 0x140029f20) ["commands.c":362, 0x120007c24]
   9 remake_file(file = 0x140029f20) ["remake.c":1008, 0x120023a5c]
...



        Yet, if I run the crash-generating command on its own:

% rm -f core            

% /proj/telain/telainInfras/release/V0_7_0/ALPHA_OSF/DEC6.3/bin_dbg/snacc -C -u 
/proj/telain/telainInfras/release/V0_7_0/share/cfg/useful.asn1  tcaptest.as
n1
zsh: 31057 segmentation fault (core dumped)  
/proj/telain/telainInfras/release/V0_7_0/ALPHA_OSF/DEC6.3/bin_dbg/snacc -C -u

% dbx /proj/telain/telainInfras/release/V0_7_0/ALPHA_OSF/DEC6.3/bin_dbg/snacc 
core
dbx version 3.11.10
Type 'help' for help.
Core file created by program "snacc"

thread 0x3 signal Segmentation fault at >*[__nxm_thread_kill, 0x3ff805c6ca8]   
ret      r31, (r26), 1
(dbx) where
>  0 __nxm_thread_kill(0x140029c10, 0x0, 0x3ff8059e790, 0x3ffc0089f08, 
> 0x3ffc01b2e60) [0x3ff805c6ca8]
   1 pthread_kill(0x0, 0x1, 0x11fffe420, 0x120008960, 0x0) [0x3ff805acdac]
   2 (unknown)() [0x3ff80596e5c]
   3 (unknown)() [0x3ff807e3774]
   4 exc_unwind(0x11fffcde8, 0xabadabad00beed00, 0x11fffd080, 0x140029c98, 
0x3ff807e3b3c) [0x3ff807e386c]
   5 exc_raise_signal_exception(0x86, 0x0, 0x3ff8019f9f0, 0x1, 0x1) 
[0x3ff807e3b38]
   6 (unknown)() [0x3ff805ae8e4]
   7 strlen(0x3ff8059e90c, 0x14007abd0, 0x3ff8016ff6c, 0xa, 0x3ff801a46a8) 
[0x3ff8019f9ec]
   8 _doprnt_dis(0x0, 0x140014758, 0x1400206e0, 0x0, 0x3ff800db058) 
[0x3ff801a2af8]
   9 _doprnt(0x3ff800db058, 0x3ffc0080380, 0x3ff800db024, 0x11fffe170, 0x10) 
[0x3ff800dbeb8]
  10 fprintf(0x3ffc0080348, 0x1400149c8, 0x14007abc0, 0x1, 0x25) [0x3ff800db020]
...


        Note the similar stack trace up until strlen.

        So, GNU make bug gurus, is this likely to be a gdb problem,
or a problem with GNU make's child process reaping ?

--
| Nigel Pearson, address@hidden | "Reality is that which,   |
|   Telstra NW-D, Sydney, Australia.    |  when you stop believing  |
| Office: 9206 3468    Fax:  9212 6329  |  in it, doesn't go away." |
| Mobile: 0408 664435  Home: 9792 6998  | Philip K. Dick - 'Valis.' |