An int may not be enough to store the amount of memory used

[lw]

A signed int on 32 bit allows to store the amount of bytes used by a program up to about 2.1GB. In some (very rare) cases this may not be enough. I've been able to trigger this behavior on compilation (where no memory limit is enforced) with the compile-timeout test. Here's the error I got:

DataError: (raised as a result of Query-invoked autoflush; consider using a session.no_autoflush block if this flush is occurring prematurely) (psycopg2.DataError) integer out of range
 [SQL: 'UPDATE submission_results SET compilation_outcome=%(compilation_outcome)s, compilation_text=%(compilation_text)s, compilation_stdout=%(compilation_stdout)s, compilation_stderr=%(compilation_stderr)s, compilation_time=%(compilation_time)s, compilation_wall_clock_time=%(compilation_wall_clock_time)s, compilation_memory=%(compilation_memory)s, compilation_shard=%(compilation_shard)s, compilation_sandbox=%(compilation_sandbox)s WHERE submission_results.submission_id = %(submission_results_submission_id)s AND submission_results.dataset_id = %(submission_results_dataset_id)s'] [parameters: {'compilation_text': '["Compilation timed out"]', 'submission_results_dataset_id': 17, 'compilation_shard': 0, 'compilation_stderr': u'In file included from batchfileio.cpp:1:0,\n                 from batchfileio.cpp:1,\n                 from batchfileio.cpp:1,\n                 from batchfileio.cpp:1,\n                 from batchfileio.cpp:1,\n                 from batchfileio.cpp:1,\n                 from batchfileio.cpp:1,\n                 from batchfileio.cpp:1,\n                 from batchfileio.cpp:1,\n                 from batchfileio.cpp:1,\n                 from batchfileio.cpp:1,\n                 from batchfileio.cpp:1,\n                 from batchfileio.cpp:1,\n                 from batchfileio.cpp:1,\n                 from batchfileio.cpp:1,\n                 from batchfileio.cpp:1,\n                 from batchfileio.cpp:1,\n                 from batchfileio.cpp:1,\n                 from batchfileio.cpp:1,\n                 from batchfileio.cpp:1,\n                 from batchfileio.cpp:1,\n                 from batchfileio.cpp:1,\n                 from batchfileio.cpp:1,\n                 from batchfileio.cpp:1,\n                 from batchfileio.cpp:', 'compilation_stdout': u'', 'compilation_sandbox': u'/tmp/tmp_KjOS5/tmp', 'compilation_outcome': u'fail', 'compilation_wall_clock_time': 21.069, 'compilation_time': 0.0, 'submission_results_submission_id': 59, 'compilation_memory': 2378412032}]

[stefano-maggiolo]

This should be only a matter of using BigInteger in the schema (handled as a usual int in Python). See http://docs.sqlalchemy.org/en/latest/core/type_basics.html

Since we're at it, it would also be nice to use decimal to store the score.

[wil93]

I think the milestone here should be changed from 1.3 to 1.4