search

uwescience / raco

Compilation and rule-based optimization framework for relational algebra. Raco is the language, optimization, and query translation layer for the Myria project.
Other
72 stars 19 forks source link

Fix failing master branch test? #530

Closed bmyerz closed 8 years ago

bmyerz commented 8 years ago

Some clang related tests in master failing, despite Travis reporting pass @jingjingwang "test_groupby_string_key" failing? I just ran it using the master branch and got this:

[9:45]
AssertionError: test: ({(4, 'S2SORBBZ9VO'): 1, (10, 'G59ZSNMHYPGDXK'): 1, (7, 'C83TT4Q'): 1, (2, '3AAFZFZYTH92SIX'): 1, (9, 'JRQX2DB4P1AQZI86BAT'): 1, (10, 'YXJTLICK9QXX7O'): 1, (1, 'N86KJRDXBA9KVQLC688EHW'): 1, (0, '2OGTZY'): 1, (10, 'HPBHPRIIHQ'): 1, (2, '47M5YR9I0DG6H1V4N'): 1, (0, 'SGJZ'): 1, (5, 'RANPGD6S'): 1, (9, 'W2NUICJ6U'): 1, (3, '5V85ETD'): 1, (2, 'I'): 1, (9, 3): 1, (1, '91T4IS8UQJT8A'): 1, (3, 'L46KMTUV'): 1, (4, 'LWCK85L4L70OJA'): 1, (5, '56AHL92M'): 1, (8, '503SUPC5UHSRMMTWWQBI'): 1, (0, 'CCWK2R5FS2C8G1'): 1, (10, 'J23AYDE5BI9'): 1, (1, '4223J4YCAA1IDWM'): 1, (4, 'GI0'): 1, (6, 'G94ELZZ7P3YKV54S'): 1, (0, 52): 1, (6, 'coffee'): 1, (10, 'LDS7K6F6BL62640YGPFZ'): 1, (3, 4): 1}, []) != expect:({(4, 'S2SORBBZ9VO'): 1, (10, 'G59ZSNMHYPGDXK'): 1, (3, ComparableFloat(4.0)): 1, (7, 'C83TT4Q'): 1, (2, '3AAFZFZYTH92SIX'): 1, (9, 'JRQX2DB4P1AQZI86BAT'): 1, (10, 'YXJTLICK9QXX7O'): 1, (1, 'N86KJRDXBA9KVQLC688EHW'): 1, (0, '2OGTZY'): 1, (9, ComparableFloat(3.0)): 1, (2, '47M5YR9I0DG6H1V4N'): 1, (0, 'SGJZ'): 1, (5, 'RANPGD6S'): 1, (1, '4223J4YCAA1IDWM'): 1, (0, ComparableFloat(52.0)): 1, (3, '5V85ETD'): 1, (2, 'I'): 1, (1, '91T4IS8UQJT8A'): 1, (10, 'HPBHPRIIHQ'): 1, (3, 'L46KMTUV'): 1, (4, 'LWCK85L4L70OJA'): 1, (5, '56AHL92M'): 1, (8, '503SUPC5UHSRMMTWWQBI'): 1, (0, 'CCWK2R5FS2C8G1'): 1, (10, 'J23AYDE5BI9'): 1, (9, 'W2NUICJ6U'): 1, (4, 'GI0'): 1, (6, 'G94ELZZ7P3YKV54S'): 1, (6, 'coffee'): 1, (10, 'LDS7K6F6BL62640YGPFZ'): 1}, [])

my guess is that it might have passed in Travis because it depends on a PRNG, so the bug can be hidden

coveralls commented 8 years ago

Coverage Status

Coverage remained the same at 91.877% when pulling fed6fb2743e46a8aab615fc01111738ffaecdffa on bmyerz/failing-floats into 04ad4953d1e6f01d2fd45cafc60568e34ad7854d on master.

bmyerz commented 8 years ago

I see the problem is that anyone with test inputs generated before commit 30a0e9e8 will have buggy test inputs. This problem can be avoided if we always regenerate the inputs when tests are run.

coveralls commented 8 years ago

Coverage Status

Coverage remained the same at 91.877% when pulling 1a8a116ad598c99d2d883d3b5d6a6a31560d3e1c on bmyerz/failing-floats into 04ad4953d1e6f01d2fd45cafc60568e34ad7854d on master.