Just one data point...
I am completely overwhelmed by all of the different ways to manage source code in the Pharo universe. I would love to contribute to upstream but this seems way out of reach: I'm not even managing to keep track of the code in my own home directory properly yet.
Just now I am looking at upstream code that should be fixed: CairoLibrary unix32ModuleName and unix64ModuleName are giving higher priority to guessed hard-coded paths than to LD_LIBRARY_PATH and this is backwards i.e. the carefully selected paths a user supplies with LD_LIBRARY_PATH should come first (imho.)
On the one hand it would be fine to dive into the upstream contribution process to work out how to do this. On the other hand I am still trying to master tools like Epica to recover useful code that I have written and lost track of when images crash and get rebuilt etc. I really need to get this under control before spending time learning Iceberg, FogBugz, etc.