recovery: windowed min and max filter implementation#398
Conversation
LPardue
left a comment
LPardue
left a comment
There was a problem hiding this comment.
I started making some comments but then as I worked through the code, it seems like its design is made really hard by MinmaxSample::time being an Option. No code seems to care about the None case, it just unwraps the Option which can result in a panic. All the test cases seem to require the reset method to be used which initializes the time. I think it would be easier to lose the option and make a Minmax::new() method that is used to initialize everything
lohith-bellad
force-pushed
the
windowed_minmax_filter
branch
3 times, most recently
from
f806c43 to
6895d5c
Compare
|
Addressed all the comments. Please review, thanks |
lohith-bellad
force-pushed
the
windowed_minmax_filter
branch
from
6895d5c to
a35728f
Compare
lohith-bellad
force-pushed
the
windowed_minmax_filter
branch
3 times, most recently
from
95356b6 to
d650b97
Compare
|
This will close #305 |
ghedo
left a comment
ghedo
left a comment
There was a problem hiding this comment.
Mostly looks good now, added a few more comments to try and make things a bit clearer.
lohith-bellad
force-pushed
the
windowed_minmax_filter
branch
2 times, most recently
from
9cceb4a to
627e2aa
Compare
Port google opensourced min-max filter code based on Kathleen Nichols' algorithm for tracking the minimum (or maximum) value of a data stream over some fixed time interval. Closes #305.
ghedo
force-pushed
the
windowed_minmax_filter
branch
from
627e2aa to
df9b6cd
Compare
|
For whatever reason rustc nightly seems to be crashing which causes the Travis CI builds to fail. Though that's probably unrelated to these changes (the crash happens when building |
|
FTR, I created rust-lang/rust#70041 to track the compiler crash. |
4 participants
Port google opensourced min-max filter code based on
Kathleen Nichols' algorithm for tracking the minimum
(or maximum) value of a data stream over some fixed
time interval.