Closed Eh2406 closed 8 months ago
Rebased + happy clippy + fmt fix
Pull out some redundant code for checking with sat
This just affects test code, so hopefully we can keep it moving.
Nice! I have an end-of-quarter thing at work to prepare for monday, but after that I’ll fully review here. Also getting a bit hyped up by packaging-con 2023 being announced so I’ll boost my time for pubgrub :)
ping?
Added a profile override to make (sat) tests run faster. Which let me run large_case
on more files I had laying around, some of which had errors that were not NoSolution
.
Backed out the override changes, even with the override the test was still too slow to run without --release
, and with --release
dependencies are already optimized.
Rebased. And this just affects test code, so starting a 10 day timer.
10 day timer is over with no objection. I will muck with settings so that I can self approve.
Removing a dependency from a dependency provider can only make more resolves succeed. Therefore, an interesting test is to look at our error reporting for some resolve, and we resolve with a dependency provider that only has dependency requirements mentioned in our report. If this subsequent run were to pass, then we have demonstrated that the error report did not include enough information to constitute a proof of unsatisfiability.
While working on the code I moved some code out of the proptest macro, as auto complete and formatting works better on non-macro code.
It would be nice to be able to test our proof of unsatisfiability against the failure information from the SAT solver. I was not able to figure out how to do this. I did find several small things that could be cleaned up about the implementation for SAT embedding.