Ted Kaminski

Just a bit more brain-storming about the potential benefits of this feature: 1. For use-cases where you don't want to check in a corpus, you could seed the corpus with...

Another thought on the potential benefits of this feature: In the situation where you're running a 24/7 fuzzing server, you want to be able to pause fuzzing, update to a...

> For instance, when using a Rng driver, then this feature is essentially impossible to implement. I don't quite follow. Certainly you have a point that it could be difficult...

I wanted to write up what I think this feature might look like from the perspective of the use case I have in mind. A normal bolero harness looks something...

> First is, it just cannot work with the rng-based drivers (used by the proptest generator). Maybe not too big a deal as the proptest generator could just first run...

A note: one might think "perhaps I can just add a translation to the type, oh well if that's not very elegant." But this would exacerbate a build problem. Consider...

Current status: We can supply a basic class (i.e. not parametrized like `Foo`) now. This is accomplished by piggy-backing the information on the `TypeRep`. Ah well. Works. Remaining work: deal...

Ok, post-discussion re-framing: The problem is: 1. Generally the simple way to do things with the IDE is let IDE plug-in developers do IDE things in the most natural way....

At first, I thought maybe this would have to be caught during spec generation time, but I think this is actually something we *can* catch during normal grammar compilation. i.e....

In addition to that comment, and that problem, it looks up productions, not concrete productions specifically.