initial set of edge cases

[anupsv]

Fixes Issue

Fixes #

Changes proposed

Screenshots (Optional)

Note to reviewers

[samlaf]

Looks like useful stuff! Would be nice to have a big comment explaining the different edge cases, otherwise it's very hard to tell whether they are extensive or not.

Also should we turn this into a fuzz test using your edge cases as seed examples?

[anupsv]

Looks like useful stuff! Would be nice to have a big comment explaining the different edge cases, otherwise it's very hard to tell whether they are extensive or not.

Also should we turn this into a fuzz test using your edge cases as seed examples?

I've been looking to create a full end to end fuzzer, i.e integrate proxy with inabox and run the fuzzer through the proxy so the full flow is being tested continuously. Wanted to get these tests in first so that known issues are fixed. Also fuzzing, depending on seed size, would create long PR cycles but i guess people are ok with it since the holesky test takes around 40 minutes. I'll add it in a different PR later on and we can discuss there for runtimes and sizes.

[samlaf]

Looks like useful stuff! Would be nice to have a big comment explaining the different edge cases, otherwise it's very hard to tell whether they are extensive or not.
Also should we turn this into a fuzz test using your edge cases as seed examples?

I've been looking to create a full end to end fuzzer, i.e integrate proxy with inabox and run the fuzzer through the proxy so the full flow is being tested continuously. Wanted to get these tests in first so that known issues are fixed. Also fuzzing, depending on seed size, would create long PR cycles but i guess people are ok with it since the holesky test takes around 40 minutes. I'll add it in a different PR later on and we can discuss there for runtimes and sizes.

Agree to merging this first and then later adding fuzzing using these as base case. golang fuzzer is integrated and pretty fast! You can run it for a fixed amount of time (say 10sec), or run it until its hit n amounts of "surprising" results (say 200). Don't think it should take that long for something simple like this, but I might be wrong...

Giving my stamp of approval, in case this is ready to merge. Prefer smaller PRs that we merge more often, otherwise we have stuff lingering around for weeks on end.