So yesterday I was playing around figuring out the Repo scheme, how to use everything, and ran into something that bent my brain a little.

Docs say you should be able to keep a clean copy of the Fan distro as your boot Repo, and set a working Repo where you make your changes. Cool.

So, I pulled the tip from the Hg repository, and installed 1.0.45 as the alternate system for rebuilding core pods; did a build (fan buildall.fan full doc) and all works great.

Then I thought, let's try those compiler tweaks, so: I set FAN_REPO to a working directory, used Hg to clone the fan system into my working directory, applied some changes to compiler::Parser.fan, did a build, and things started to get weird.

At the time I was trying to figure out where stuff goes, so I don't know if there's a problem or if I just got things jumbled up. But I was getting what seemed to be clean builds, but nothing would run -- errors about not being able to find fwt.

Best I could tell, I had the newly compiled compiler pod in my working repo, the rest of the core pods (and no others) in the boot repo, and I don't know where the jars were put.

I figure I was probably doing something wrong, but I think the build script did a clean that deleted my boot-repo pods; not really sure. Anyway I restored everything and decided to hold off on that experiment for a minute.

Question, should that work? Or do you have to turn off the working repo if you're working with core system pods?

current design allows you to develop third party 100% fan pods (no native Java or C#) in a working repo, but all the bootstrap stuff requires working in the a single bootstrap repo - should have noted that

No problem, I ended up deciding that must be the case. Things are working great otherwise.