I'm struggling with finding a memory management problem in my latest array container (non-RC just C++-style) at

https://github.com/nordlow/phobos-next/blob/master/src/array_ex.d

When I run the program separately my terminal spits out


*** Error in `./array_ex': corrupted double-linked list: 0x0000000001c3a2b0 ***
======= Backtrace: =========
/lib/x86_64-linux-gnu/libc.so.6(+0x77725)[0x7fc0bb3be725]
...
./array_ex[0x4a3b7f]
...
/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf0)[0x7fc0bb367830]
./array_ex[0x401ef9]
======= Memory map: ========
00400000-004fe000 r-xp 00000000 08:05 5379197                            /home/per/Work/justd/phobos-next/src/array_ex
004fe000-004ff000 r--p 000fd000 08:05 5379197                            /home/per/Work/justd/phobos-next/src/array_ex
004ff000-00515000 rw-p 000fe000 08:05 5379197                            /home/per/Work/justd/phobos-next/src/array_ex
00515000-00516000 rw-p 00000000 00:00 0
01c2f000-02001000 rw-p 00000000 00:00 0                                  [heap]
7fc0b4000000-7fc0b4021000 rw-p 00000000 00:00 0
7fc0b4021000-7fc0b8000000 ---p 00000000 00:00 0
7fc0bb347000-7fc0bb507000 r-xp 00000000 08:05 2410363                    /lib/x86_64-linux-gnu/libc-2.23.so
7fc0bb507000-7fc0bb706000 ---p 001c0000 08:05 2410363                    /lib/x86_64-linux-gnu/libc-2.23.so
7fc0bb706000-7fc0bb70a000 r--p 001bf000 08:05 2410363                    /lib/x86_64-linux-gnu/libc-2.23.so
7fc0bb70a000-7fc0bb70c000 rw-p 001c3000 08:05 2410363                    /lib/x86_64-linux-gnu/libc-2.23.so
7fc0bb70c000-7fc0bb710000 rw-p 00000000 00:00 0
7fc0bb710000-7fc0bb726000 r-xp 00000000 08:05 2369742                    /lib/x86_64-linux-gnu/libgcc_s.so.1
7fc0bb726000-7fc0bb925000 ---p 00016000 08:05 2369742                    /lib/x86_64-linux-gnu/libgcc_s.so.1
7fc0bb925000-7fc0bb926000 r--p 00015000 08:05 2369742                    /lib/x86_64-linux-gnu/libgcc_s.so.1
7fc0bb926000-7fc0bb927000 rw-p 00016000 08:05 2369742                    /lib/x86_64-linux-gnu/libgcc_s.so.1
7fc0bb927000-7fc0bb92e000 r-xp 00000000 08:05 2410358                    /lib/x86_64-linux-gnu/librt-2.23.so
7fc0bb92e000-7fc0bbb2d000 ---p 00007000 08:05 2410358                    /lib/x86_64-linux-gnu/librt-2.23.so
7fc0bbb2d000-7fc0bbb2e000 r--p 00006000 08:05 2410358                    /lib/x86_64-linux-gnu/librt-2.23.so
7fc0bbb2e000-7fc0bbb2f000 rw-p 00007000 08:05 2410358                    /lib/x86_64-linux-gnu/librt-2.23.so
7fc0bbb2f000-7fc0bbb47000 r-xp 00000000 08:05 2410362                    /lib/x86_64-linux-gnu/libpthread-2.23.so
7fc0bbb47000-7fc0bbd46000 ---p 00018000 08:05 2410362                    /lib/x86_64-linux-gnu/libpthread-2.23.so
7fc0bbd46000-7fc0bbd47000 r--p 00017000 08:05 2410362                    /lib/x86_64-linux-gnu/libpthread-2.23.so
7fc0bbd47000-7fc0bbd48000 rw-p 00018000 08:05 2410362                    /lib/x86_64-linux-gnu/libpthread-2.23.so
7fc0bbd48000-7fc0bbd4c000 rw-p 00000000 00:00 0
7fc0bbd4c000-7fc0bbd72000 r-xp 00000000 08:05 2410359                    /lib/x86_64-linux-gnu/ld-2.23.so
7fc0bbe42000-7fc0bbf46000 rw-p 00000000 00:00 0
7fc0bbf6e000-7fc0bbf71000 rw-p 00000000 00:00 0
7fc0bbf71000-7fc0bbf72000 r--p 00025000 08:05 2410359                    /lib/x86_64-linux-gnu/ld-2.23.so
7fc0bbf72000-7fc0bbf73000 rw-p 00026000 08:05 2410359                    /lib/x86_64-linux-gnu/ld-2.23.so
7fc0bbf73000-7fc0bbf74000 rw-p 00000000 00:00 0
7ffdbd0cc000-7ffdbd0ee000 rw-p 00000000 00:00 0                          [stack]
7ffdbd1f9000-7ffdbd1fb000 r--p 00000000 00:00 0                          [vvar]
7ffdbd1fb000-7ffdbd1fd000 r-xp 00000000 00:00 0                          [vdso]
ffffffffff600000-ffffffffff601000 r-xp 00000000 00:00 0                  [vsyscall]
Interrupted (SIGABRT


The problem here is that I can't even pipe the call to a file; neither

    ./array_ex > dump

nor

    ./array_ex 2> dump

succeeds in creating any file. This might be the reason why my Dustmite call


dustmite --no-redirect src "dmd -main -unittest -g -debug \
    array_ex container_traits searching_ex && \
    ./array_ex 2>&1 | grep -qF double-linked"


stops and spits out


Aborted (core dumped)
No
object.Exception@dustmite.d(270): Initial test fails
----------------
??:? _Dmain [0x5c6be8]
??:? _D2rt6dmain211_d_run_mainUiPPaPUAAaZiZ6runAllMFZ9__lambda1MFZv [0x62544e]
??:? void rt.dmain2._d_run_main(int, char**, extern (C) int function(char[][])*).tryExec(scope void delegate()) [0x625398]
??:? void rt.dmain2._d_run_main(int, char**, extern (C) int function(char[][])*).runAll() [0x62540a]
??:? void rt.dmain2._d_run_main(int, char**, extern (C) int function(char[][])*).tryExec(scope void delegate()) [0x625398]
??:? _d_run_main [0x625309]
??:? main [0x61732d]
??:? __libc_start_main [0xaf26b82f]
[


It seems there is some special behaviour of not being able to redirect the ouputs of stdout and stderr here.
What to do?

I need your help here, I'm complete stuck. Vladimir?


Note: The problem should be reproducable by cloning

    https://github.com/nordlow/phobos-next.git

and doing the call to DustMite at the root of this checkout.

On Thursday, 20 October 2016 at 21:33:59 UTC, Nordlöw wrote:
> I need your help here, I'm complete stuck. Vladimir?

If you see output even after redirecting stdout and stderr, the program is probably writing to /dev/tty or similar.

I'm not sure what the "proper" way is to run a program without a TTY... but what happens if you run the program via ssh, without allocating a PTY? I.e. with -T, or with a command and without -t.

On Thursday, 20 October 2016 at 22:18:20 UTC, Vladimir Panteleev wrote:
> On Thursday, 20 October 2016 at 21:33:59 UTC, Nordlöw wrote:
>> I need your help here, I'm complete stuck. Vladimir?

The exit status of the `./array_ex` call (as seen from bash) is 134. How do I, in Bash, map that exit status to zero, and all other exit statuses to non-zero?

Next try:

dustmite --no-redirect src "dmd -main -unittest -g -debug array_ex container_traits searching_ex && { ./array_ex; } >&log; grep SIGABRT log"

but that fails too as

Loading src/w3c.html
None => /bin/sh: 1: Syntax error: Bad fd number
No
object.Exception@dustmite.d(270): Initial test fails
----------------
??:? _Dmain [0x5c6be8]
??:? _D2rt6dmain211_d_run_mainUiPPaPUAAaZiZ6runAllMFZ9__lambda1MFZv [0x62544e]
??:? void rt.dmain2._d_run_main(int, char**, extern (C) int function(char[][])*).tryExec(scope void delegate()) [0x625398]
??:? void rt.dmain2._d_run_main(int, char**, extern (C) int function(char[][])*).runAll() [0x62540a]
??:? void rt.dmain2._d_run_main(int, char**, extern (C) int function(char[][])*).tryExec(scope void delegate()) [0x625398]
??:? _d_run_main [0x625309]
??:? main [0x61732d]
??:? __libc_start_main [0x8b69a82f]


I can't get ! operator to work either.

On Friday, 21 October 2016 at 06:44:54 UTC, Nordlöw wrote:
> On Thursday, 20 October 2016 at 22:18:20 UTC, Vladimir Panteleev wrote:
>> On Thursday, 20 October 2016 at 21:33:59 UTC, Nordlöw wrote:
>>> I need your help here, I'm complete stuck. Vladimir?
>
> The exit status of the `./array_ex` call (as seen from bash) is 134. How do I, in Bash, map that exit status to zero, and all other exit statuses to non-zero?

Solution construct a Bash script named, say show-segfault:


#!/usr/bin/env python3

import sys
import pty

stat = pty.spawn(sys.argv[1:])
if stat == 256:
    exit(42)                    # remap to 42
else:
    exit(stat)


and wrap call to rdmd as


show-segfault rdmd ...

On Friday, 21 October 2016 at 10:13:23 UTC, Nordlöw wrote:
> #!/usr/bin/env python3
>
> import sys
> import pty
>
> stat = pty.spawn(sys.argv[1:])
> if stat == 256:
>     exit(42)                    # remap to 42
> else:
>     exit(stat)
>

Assuming you want to remap 134 to 0 (success):

#!/bin/bash

cmd="$1"
shift 2>/dev/null

$cmd "$@"
[ $? = 134 ] && exit 0 || exit 1