Paul Chiusano

The last idea on this that I recall is that PRs are encoded as namespaces with two subnamespaces, `base`, and `head`. An anonymous namespace is created for this when `pr.create`...

Phase 1 of this should just be a timeboxed design jam / prototyping / implementation sketch of 2 days to decide on 1 or more possible approaches and better estimate...

Phase 1 is a timeboxed design jam (2 meetings, Simon goes off and mocks a few things up between first and second meeting) to figure out what the front-end will...

### Option 1 The simplest fix for this: the pretty-printer just adds the GUID to the declaration it generates when a definition is dumped to a file. ```Haskell unique[someguid] type...

> When we do an update of unique type Direction, we look up the existing decl Direction (no new index is needed, just the existing index by name), and from...

> In HandleInput.hs -> UpdateI, parsing/hashing has already occurred (i.e. nothing new is needed there), we're passed a TypecheckedUnisonFile. We currently have a pure function toSlurpResult which we call here,...

> Aren't the cycles numbering and constructor orders just dependent on the structure of the Decl? or are they dependent on the GUID too? If I have: ```haskell unique type...

I think your idea still works, it's just that the "determine if these two structures are the same" calculation needs to use the algorithm I suggested - unhash the decls...

Idea to increase certainty of this: reduce scope of the issue to a timeboxed investigation of possible fix

This seems to have been miraculously fixed. Perhaps we'll never know how... Tested on Unison Share.