Ensure parameters.oustanding_kit does not drift far from the real value #156
Labels
testing
The issue relates to testing
Comments
This was referenced Jul 26, 2021
gkaracha
added a commit
that referenced
this issue
Aug 6, 2021
After some investigation I found out that underapproximations of the total outstanding kit are not all that uncommon, which justifies the need for #218. Here is a high-level example / regression test illustrating this. Other items: * Add `compute_outstanding_dissonance` to help identify underapproximations (useful for investigating #156 too). * Remove the warnings on underapproximations in `checker.ml`; they are more common than I expected. * Type signature fixes in `Ligo.Map` definitions.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
In the system parameters we keep track of the total amount of kit needed to close all currently open burrows via field
outstanding_kit. However, this is only an approximation of the real amount. Not only we update theburrow.outstanding_kitfield individually (per burrow) when touching burrows, but also burrows or checker might remain untouched for a while, which could introduce further rounding errors.Errors of a few percents per year are NOT acceptable. Errors of 0.1% or so per year would be tolerable.
In order to test this meaningfully we should be able to create scenaria with many burrows, a lot of activity, and let a lot of time pass, which can be quite complicated to manufacture. Furthermore, one must keep track of all the burrows ever created to be able to inspect the storage and aggregate the total amount of outstanding kit for all burrows so that it can be compared to
parameters.oustanding_kit.The text was updated successfully, but these errors were encountered: