This looks mighty odd. Dee appears to segfault in a g_ptr_array_index(row, 3) - but we explicitly create all the GPtrArrays for the rows with lengths == n_columns+1 and we assert that 3 < n_columns just a few lines before the crash...
Anyway - I added some guards against weird internal state corruption (mem corruption?)
This looks mighty odd. Dee appears to segfault in a g_ptr_array_ index(row, 3) - but we explicitly create all the GPtrArrays for the rows with lengths == n_columns+1 and we assert that 3 < n_columns just a few lines before the crash...
Anyway - I added some guards against weird internal state corruption (mem corruption?)