> The issue here was that GDAL segfaulted when opening that database because of the null dereference of a corrupted record. Did SQLite segfault? If so that would be a bug that we need to look into. Or, did GDAL do an sqlite3_column_text() and get back a NULL pointer, then dereference the NULL pointer, thus causing a segfault. That would be a bug in GDAL. The sqlite3_column_text() might return NULL, even for a NOT NULL column, for example following an OOM error. That's part of the specification of how sqlite3_column_text() works. If you were not checking for a NULL return from sqlite3_column_text() and it always worked before, that is because you were lucky. It would be to same if you failed to check for a NULL return from malloc().