O.C.Peras.Weight: add totalWeightForFragment/takeVolatileSuffix
#1621
Conversation
This is purely for concise QuickCheck counterexample output.
amesgen
force-pushed
the
amesgen/more-peras-weight-stuff
branch
2 times, most recently
from
0ae95b6 to
5cba290
Compare
ouroboros-consensus/src/ouroboros-consensus/Ouroboros/Consensus/Peras/Weight.hs
Outdated
Show resolved
Hide resolved
|
|
||
| prop_perasWeightSnapshot :: TestSetup -> Property | ||
| prop_perasWeightSnapshot testSetup = | ||
| tabulate "log₂ # of points" [show $ round @Double @Int $ logBase 2 (fromIntegral (length tsPoints))] |
There was a problem hiding this comment.
Why do we round to the nearest integer?
There was a problem hiding this comment.
To get a reasonable amount of labelling output; if we didn't round (or floor/ceiling, doesn't matter), then we would get one line of output for every amount of points that is present in some test case. This way, we essentially get "buckets" of exponentially increasing size.
The high-level motivation for this is to get an idea of how many points are present in this test, as the test is rather degenerate if there are only very few (and the test would be unnecessarily slow if there were, say, millions of points).
There was a problem hiding this comment.
I wasn't sure if the rounding was just for displaying, or also for producing the test cases.
There was a problem hiding this comment.
It is indeed purely affecting "metrics" of the property tests; generation/execution of test cases is completely unaffected 👍
ouroboros-consensus/test/consensus-test/Test/Consensus/Peras/WeightSnapshot.hs
Outdated
Show resolved
Hide resolved
ouroboros-consensus/test/consensus-test/Test/Consensus/Peras/WeightSnapshot.hs
Outdated
Show resolved
Hide resolved
ouroboros-consensus/test/consensus-test/Test/Consensus/Peras/WeightSnapshot.hs
Show resolved
Hide resolved
amesgen
force-pushed
the
amesgen/more-peras-weight-stuff
branch
from
5cba290 to
4c1afc8
Compare
Also add a test for
PerasWeightSnapshot. The tests forboostedWeightOf{Point,Fragment}are currently not that useful as the model implementation and the actual implementation are almost the same, but this will change with #1613. The test fortakeVolatileSuffixis already useful.