[GRASS-user] v.net.iso - segmentation fault error
Eduardo Corbelle Rico
eduardo.corbelle at gmx.net
Fri Nov 18 07:48:12 PST 2016
O Xov, 17-11-2016 ás 08:52 +0100, Markus Metz escribiu:
>
> On Wed, Nov 16, 2016 at 7:14 PM, Eduardo Corbelle Rico <eduardo.corbe
> lle at gmx.net> wrote:
> >
> > O Mér, 16-11-2016 ás 15:22 +0100, Markus Metz escribiu:
> > >
> > > On Wed, Nov 16, 2016 at 11:44 AM, Eduardo Corbelle Rico
> <eduardo.corb
> > > elle at gmx.net> wrote:
> > > >
> > > > O Xov, 27-10-2016 ás 14:56 +0200, Markus Metz escribiu:
> > > > > On Thu, Oct 27, 2016 at 11:16 AM, Eduardo Corbelle Rico
> > > > > <eduardo.corbelle at gmx.net> wrote:
> > > > > >
> > > > > > O Mér, 26-10-2016 ás 23:18 +0200, Markus Metz escribiu:
> > > > > > > On Thu, Oct 13, 2016 at 9:55 AM, Eduardo Corbelle Rico
> > > > > > > <eduardo.corbelle at gmx.net> wrote:
> > > > > > > >
> > > > > > > > Dear all,
> > > > > > > >
> > > > > > > > I am unable to use v.net.iso because of a "segmentation
> > > fault"
> > > > > > > > error. I
> > > > > > > > have a script that used to work correctly in GRASS 6.4
> (32
> > > > > > > > bit)
> > > > > > > > but
> > > > > > > > produces this error after I switched to GRASS 7.0.4 (64
> > > bit).
> > > > > > > >
> > > > > > > > If I try to follow both examples shown in https://grass
> .osg
> > > eo.o
> > > > > > > > rg/g
> > > > > > > > rass
> > > > > > > > 70/manuals/v.net.iso.html (using the Spearfish dataset
> for
> > > > > > > > GRASS 7)
> > > > > > > > the
> > > > > > > > same error appears:
> > > > > > > >
> > > > > > > > > Building graph...
> > > > > > > > > Registering arcs...
> > > > > > > > > Segmentation fault
> > > > > > >
> > > > > > > The segmentation fault must happen in the vector
> libraries. I
> > > > > > > tested
> > > > > > > on Linux and can not reproduce the segmentation fault.
> > > Valgrind
> > > > > > > also
> > > > > > > does not show anything that could cause a segmentation
> fault.
> > > Can
> > > > > > > you
> > > > > > > provide a gdb backtrace?
> > > > > > >
> > > > > > > Markus M
> > > > > >
> > > > > > Thank you Markus,
> > > > > >
> > > > > > My system is debian Stretch, and the following lines
> describe
> > > my
> > > > > > attempt at generating a gdb backtrace. Please let me know
> if I
> > > > > > should
> > > > > > do otherwise.
> > > > >
> > > > > So far so good. The module crashes in a library function that
> > > comes
> > > > > from libavl without modifications, but I can still not see
> why it
> > > > > could cause a segmentation fault. Can you recompile with the
> > > compiler
> > > > > debugging option -g? gdb will then be able to provide more
> > > > > information:
> > > > >
> > > > > make distclean
> > > > > CFLAGS="-g" ./configure <configure options>
> > > > > make
> > > > >
> > > > > optionally make install
> > > > >
> > > > > Then start GRASS and use gdb for v.net.iso as before.
> > > > >
> > > > > Markus M
> > > > >
> > > >
> > > > Dear Markus,
> > > >
> > > > Following those instructions does not seem to have changed the
> > > > resulting backtrace info (copied below). Maybe I missed
> something
> > > along
> > > > the way?
> > > >
> > > > ---------------------
> > >
> > >
> > > Is this something Debian specific that debugging symbols must be
> > > explicitly activated? Are you using GCC or another compiler?
> > >
> > > Alternatively, you can also try valgrind:
> > >
> > > CMD="v.net.iso input=myroads_net output=myroads_net_iso
> > > center_cats=1-100000 costs=1000,2000,5000 --o"
> > > valgrind --tool=memcheck --log-file=v.net.iso.vg.log $CMD
> > >
> > > Markus M
> > >
> >
> > You were right: it was my fault not having grass-core-dbgsym and
> grass-
> > dev-dbgsym packages installed.
>
> You do not need these packages when you compile from source.
>
> > I copied below the results of the
> > debugging commands:
> >
> > ================
> > GRASS 7.0.5 (spearfish60_grass7):~ > gdb `which v.net.iso`
> > GNU gdb (Debian 7.11.1-2+b1) 7.11.1
> > Copyright (C) 2016 Free Software Foundation, Inc.
> > License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses
> /gpl
> > .html>
> > This is free software: you are free to change and redistribute it.
> > There is NO WARRANTY, to the extent permitted by law. Type "show
> > copying"
> > and "show warranty" for details.
> > This GDB was configured as "x86_64-linux-gnu".
> > Type "show configuration" for configuration details.
> > For bug reporting instructions, please see:
> > <http://www.gnu.org/software/gdb/bugs/>.
> > Find the GDB manual and other documentation resources online at:
> > <http://www.gnu.org/software/gdb/documentation/>.
> > For help, type "help".
> > Type "apropos word" to search for commands related to "word"...
> > Reading symbols from /usr/lib/grass70/bin/v.net.iso...Reading
> symbols
> > from /usr/lib/debug/.build-
> > id/5a/060fc6c3da41153d82601e3fa37ba6d88e9c6c.debug...done.
> > done.
> >
> >
> > (gdb) run input=myroads_net output=myroads_net_iso center_cats=1-
> 100000
> > costs=1000,2000,5000
>
> This seems to be the GRASS version installed with your package
> manager because make install installs by default to /usr/local/:
> > Starting program: /usr/lib/grass70/bin/v.net.iso input=myroads_net
> > output=myroads_net_iso center_cats=1-100000 costs=1000,2000,5000
> > [Thread debugging using libthread_db enabled]
> > Using host libthread_db library "/lib/x86_64-linux-
> > gnu/libthread_db.so.1".
> > Building graph...
> > Registering arcs...
> >
> > Program received signal SIGSEGV, Segmentation fault.
> > tavl_probe (tree=tree at entry=0x55555599b6c0, item=item at entry=0x55555
> 599f
> > 7b0)
> > at tavl.c:146
> > 146 tavl.c: Non hai tal ficheiro ou directorio.
> > (gdb) bt full
> > #0 tavl_probe (tree=tree at entry=0x55555599b6c0, item=item at entry=0x5
> 5555
> > 599f7b0)
> > at tavl.c:146
> > y = 0x0
> > z = 0x55555599b6c0
> > p = 0x0
>
> At this stage (tavl.c:146), p should be equal to z, and p has been
> accessed previously, i.e. the segmentation fault should occur
> earlier. Therefore I suspect changes in tavl.c introduced when
> creating the package. You would need to remove all GRASS 7 packages
> from your system, then compile from source again.
>
> Markus M
Ok. It seems that I was still running the version installed with the
package manager. When using the version compiled from source, the
segmentation error does not appear at all.
Thank you very much.
Eduardo
More information about the grass-user
mailing list