matrix_rational_sparse_allocate_mpq_vector leak (from modular/ssmod/ssmod.py)

[mabshoff]

From Sage 2.8.3rc3:

==24845== 352 bytes in 44 blocks are definitely lost in loss record 1,423 of 2,587
==24845==    at 0x4A05809: malloc (vg_replace_malloc.c:149)
==24845==    by 0x946C2A6: __gmpq_init (in /tmp/Work2/sage-2.8.3.rc3/local/lib/libgmp.so.3.4.1)
==24845==    by 0x20EB34C3: __pyx_f_22matrix_rational_sparse_allocate_mpq_vector (matrix_rational_sparse.c:5358)
==24845==    by 0x20EB51D4: __pyx_f_22matrix_rational_sparse_mpq_vector_set_entry (matrix_rational_sparse.c:6394)
==24845==    by 0x20EB5367: __pyx_f_22matrix_rational_sparse_22Matrix_rational_sparse_set_unsafe (matrix_rational_sparse.c:8
463)
==24845==    by 0x1DF0A336: __pyx_f_7matrix0_6Matrix_add_multiple_of_column_c (matrix0.c:5631)
==24845==    by 0x1DA89C8B: __pyx_f_7matrix2_6Matrix_hessenbergize (matrix2.c:4699)
==24845==    by 0x4156A2: PyObject_Call (abstract.c:1860)
==24845==    by 0x47DB71: PyEval_CallObjectWithKeywords (ceval.c:3433)
==24845==    by 0x1DA86360: __pyx_f_7matrix2_6Matrix_hessenberg_form (matrix2.c:4233)
==24845==    by 0x4156A2: PyObject_Call (abstract.c:1860)
==24845==    by 0x47DB71: PyEval_CallObjectWithKeywords (ceval.c:3433)

Cheers,

Michael

[mabshoff]

With patches 6801 and 6802 this leaks has been but cut by more than half:

==6907== LEAK SUMMARY:
==6907==    definitely lost: 191,508 bytes in 16,482 blocks.
==6907==    indirectly lost: 1,128 bytes in 16 blocks.
==6907==      possibly lost: 327,526 bytes in 908 blocks.
==6907==    still reachable: 136,860,457 bytes in 22,940 blocks.
==6907==         suppressed: 0 bytes in 0 blocks.

vs.

==21394== LEAK SUMMARY:
==21394==    definitely lost: 77,340 bytes in 2,550 blocks.
==21394==    indirectly lost: 816 bytes in 3 blocks.
==21394==      possibly lost: 325,374 bytes in 903 blocks.
==21394==    still reachable: 136,860,401 bytes in 22,935 blocks.
==21394==         suppressed: 0 bytes in 0 blocks.

Cheers,

Michael

[mabshoff]

Fixed along the way.

Cheers,

Michael