• =?UTF-8?B?Tm9yZGzDtnc=?= (23/23) Sep 21 2016 Doing a null deref such as
• ketmar (5/6) Sep 21 2016 { import etc.linux.memoryerror; registerMemoryErrorHandler(); }
• =?UTF-8?B?Tm9yZGzDtnc=?= (23/24) Sep 22 2016 Thx! That at least triggered an exception. However the line
• ketmar (4/5) Sep 22 2016 prolly. segfault handler is highly non-standard hack, it may miss
• ketmar (1/1) Sep 22 2016 p.s. that is the reason it is not turned on by default, btw.
• Adam D. Ruppe (3/4) Sep 22 2016 Yeah, it shows the line before instead of the line of. But it
• Jonathan Marler (3/7) Sep 21 2016 Can you include compiler command line? I use -g -gs -debug to

=?UTF-8?B?Tm9yZGzDtnc=?= <per.nordlow gmail.com> writes:

Doing a null deref such as

     int* y = null;
     *y = 42;                    // boom

currently outputs

/home/per/.emacs.d/auto-builds/dmd/Debug-Unittest-Require-Full-Coverage/home/per/foo[0x406474]
/lib/x86_64-linux-gnu/libpthread.so.0(+0x113d0)[0x7f08cf25e3d0]
/home/per/.emacs.d/auto-builds/dmd/Debug-Unittest-Require-Full-Coverage/home/per/foo[0x40233a]
/home/per/.emacs.d/auto-builds/dmd/Debug-Unittest-Require-Full-Coverage/home/per/foo[0x4023b9]
/home/per/.emacs.d/auto-builds/dmd/Debug-Unittest-Require-Full-Coverage/home/per/foo[0x4064c1]
/home/per/.emacs.d/auto-builds/dmd/Debug-Unittest-Require-Full-Coverage/home/per/foo[0x405704]
/home/per/.emacs.d/auto-builds/dmd/Debug-Unittest-Require-Full-Coverage/home/per/foo[0x40be2b]
/home/per/.emacs.d/auto-builds/dmd/Debug-Unittest-Require-Full-Coverage/home/per/foo[0x40c2d2]
/home/per/.emacs.d/auto-builds/dmd/Debug-Unittest-Require-Full-Coverage/home/per/foo[0x40bdbc]
/home/per/.emacs.d/auto-builds/dmd/Debug-Unittest-Require-Full-Coverage/home/per/foo[0x4056e0]
/home/per/.emacs.d/auto-builds/dmd/Debug-Unittest-Require-Full-Coverage/home/per/foo[0x4063b3]
/home/per/.emacs.d/auto-builds/dmd/Debug-Unittest-Require-Full-Coverage/home/per/foo[0x402a97]
/home/per/.emacs.d/auto-builds/dmd/Debug-Unittest-Require-Full-Coverage/home/per/foo[0x402a39]
/home/per/.emacs.d/auto-builds/dmd/Debug-Unittest-Require-Full-Coverage/home/per/foo[0x4029aa]
/home/per/.emacs.d/auto-builds/dmd/Debug-Unittest-Require-Full-Coverage/home/per/foo[0x402540]
/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf0)[0x7f08ce77c830]
/home/per/.emacs.d/auto-builds/dmd/Debug-Unittest-Require-Full-Coverage/home/per/foo[0x402219]

Is there a way to get a stacktrace instead?

I'm using dmd git master.

ketmar <ketmar ketmar.no-ip.org> writes:

On Wednesday, 21 September 2016 at 23:36:08 UTC, Nordlöw wrote:

 Is there a way to get a stacktrace instead?

{ import etc.linux.memoryerror; registerMemoryErrorHandler(); }

warning: this is hightly system-specific, and may work or not 
work depending of the moon phase, and may broke on any system 
upgrade.

=?UTF-8?B?Tm9yZGzDtnc=?= <per.nordlow gmail.com> writes:

On Thursday, 22 September 2016 at 00:46:19 UTC, ketmar wrote:
 { import etc.linux.memoryerror; registerMemoryErrorHandler(); }

Thx! That at least triggered an exception. However the line 
number for the innermost function is wrong. For instance

1   void boom()
2   {
3       int* y = null;
4       *y = 42;                    // boom
5   }

gives stacktrace

etc.linux.memoryerror.NullPointerError src/etc/linux/memoryerror.d(325)
----------------
??:? [0x404e6d]
??:? [0x404dba]
t_segFaulException.d:3 [0x402301]
t_segFaulException.d:22 [0x402370]
??:? [0x402a2e]
??:? [0x402978]
??:? [0x4029ea]
??:? [0x402978]
??:? [0x4028e9]
??:? [0x40247f]
??:? __libc_start_main [0x2867d82f]

A known bug?

ketmar <ketmar ketmar.no-ip.org> writes:

On Thursday, 22 September 2016 at 19:51:31 UTC, Nordlöw wrote:
 A known bug?

prolly. segfault handler is highly non-standard hack, it may miss 
exact position or something. as is "it is not guaranteed to work, 
and if it will work, it is not guaranteed to work correctly".

ketmar <ketmar ketmar.no-ip.org> writes:

p.s. that is the reason it is not turned on by default, btw.

Adam D. Ruppe <destructionator gmail.com> writes:

On Thursday, 22 September 2016 at 19:51:31 UTC, Nordlöw wrote:
 A known bug?

Yeah, it shows the line before instead of the line of. But it 
still shows basically where you are.

Jonathan Marler <johnnymarler gmail.com> writes:

On Wednesday, 21 September 2016 at 23:36:08 UTC, Nordlöw wrote:
 Doing a null deref such as

     int* y = null;
     *y = 42;                    // boom

 [...]

Can you include compiler command line?  I use -g -gs -debug to 
get stack traces on windows.