Yeah I’ve lost countless hours to build pipelines.
Custom images/runners is the way to go. Basically docker image that runs your pipeline scripts. Then you have much more control over things.

They will be very useful in source generator scenarios.

Suffering for your friends I see 🙂

Oh ffs. Spelling is hard. Pleasure. It’s a pleasure 😆

It would 🙂

Try LinkedIn, it’s all self-congratulatory posts 🙂

I’ve also had good success using this package: www.nuget.org/packages/Mez...

I am more thinking about what to snapshot test and what not to snapshot test.

E.g., a rule I try to follow is to not include irrelevant (to the test) data in the snapshot, as it makes it harder to validate the snapshot when it changes.

Congrats Scott. You are living the dream!

Actually, I think I usually do a `git clean -fdx` for a "reset the world" operation.

There is a git command that does it, just can’t remember it when I need it, so have to search the web. If your tool had a -all option that would be great 😊

Could it also have an option that does the same a clean git clone?

Sometimes deleting e.g., the .vs folder helps with weirdness in vs.

By using Aspire and related client libraries you get basic telemetry out of the box (without solution specific context), so I’ll do that and then add some/few end-2-end tests for basic protection against regressions.

Solve different problems, so it’s like comparing apples and oranges.

It’s not trivial for sure.

@andrewlock.bsky.social has all the details here: andrewlock.net/creating-a-s...