Bug#868759: gdal-bin: Segmentation fault with SQLite driver.

[Andy G Wood]

Hi Bas,

> You may have some packages from unstable or not fully upgraded all
> packages in testing/buster.
> 
> Please make sure that all your installed packages are from testing/buster.

Both of
$ dpkg-query -l | grep -v '^ii'
$ apt-show-versions | grep 'unstable '
show nothing that would suggest I have anything other than all installed 
packages from testing/buster.

Apologies if I am telling you nothing that you didn't already know, but a 
Google search suggested that a SIGSEGV caused by malloc could be a heap 
corruption and recommended valgrind.  Nothing ventured, nothing gained, I 
tried valgrind:
---------------------
$ valgrind --leak-check=yes ogr2ogr -t_srs EPSG:3857 -f "SQLite" -dsco 
SPATIALITE=YES test.sqlite test.xml
Memcheck, a memory error detector
Copyright (C) 2002-2015, and GNU GPL'd, by Julian Seward et al.
Using Valgrind-3.12.0 and LibVEX; rerun with -h for copyright info
Command: ogr2ogr -t_srs EPSG:3857 -f SQLite -dsco SPATIALITE=YES test.sqlite 
test.xml

Invalid free() / delete / delete[] / realloc()
   at 0x4C2D2DB: operator delete(void*) (in /usr/lib/valgrind/
vgpreload_memcheck-amd64-linux.so)
   by 0x6B60C7E: __cxa_finalize (cxa_finalize.c:56)
   by 0x1873C6C2: ??? (in /usr/lib/x86_64-linux-gnu/libdap.so.25.0.0)
   by 0x400FE79: _dl_fini (dl-fini.c:235)
   by 0x6B6090F: __run_exit_handlers (exit.c:83)
   by 0x6B60969: exit (exit.c:105)
   by 0x6B4B2B7: (below main) (libc-start.c:325)
 Address 0x1d043680 is 0 bytes inside a block of size 25 free'd
   at 0x4C2D2DB: operator delete(void*) (in /usr/lib/valgrind/
vgpreload_memcheck-amd64-linux.so)
   by 0x6B60C7E: __cxa_finalize (cxa_finalize.c:56)
   by 0xF8F08F2: ??? (in /usr/lib/x86_64-linux-gnu/libdap.so.23.1.1)
   by 0x400FE79: _dl_fini (dl-fini.c:235)
   by 0x6B6090F: __run_exit_handlers (exit.c:83)
   by 0x6B60969: exit (exit.c:105)
   by 0x6B4B2B7: (below main) (libc-start.c:325)
---------------------

Could this be a bug in libdap25 which also migrated into testing in recent 
days?  Although the test case gives a segfault, a valid sqlite database is 
created.

Andy.