Jacob Finkelman

#104 has made "zuse" less interesting. I have been using a "hobo" example for a while, but it was very slow to run with our benchmarking tools because of synthetic...

Thank you for trying PubGrub! I appreciate the production feedback. I am just giving first reactions here. I may disagree with the opinion stated upon further reflection or the arrival...

> The greater problem is that the version of root is meaningless and propagated through the reporting system. Here's an example of the code I wrote to strip versions from...

After a lot of work, I have been able to reproduce your set up. Here is a profile I was able to collect https://share.firefox.dev/43ioU30 by hacking in a `OfflineDependencyProvider`, adding...

> I think one biggest difference is your incorporation of the `RPITIT` branch. Another safe difference is that large case attempts to resolve every version not just the root version....

That does make `pick_highest_priority_package` `O(1)`, but it also pushes a lot of work into the guts of the resolver. So for example in this case with your FIFO `.map(|priority| priority...

In [my work using](https://github.com/Eh2406/pubgrub-crates-benchmark) PubGrub for Cargo packages I am now also seeing an extraordinary amount of time in `pick_highest_priority_package`. To the point where returning a constant `1` ends up...

Some obvious thoughts about when it is safe/efficient to call the new `simplify`. - If the user provided us with the VS, then we should not simplify. If they wanted...

Some more thoughts: - It is extremely risky to utilize `simplify` in the construction of `partial_solution` or `relation`. It runs the risk of computing inferences that cannot be proven from...

This is closely tied to #19. #19 points out that we can reduce the number of incompatibilities we need to check by preemptively combining ones that have the same semantic...