As part of the upstream Pistachio release, and the NICTA code drop (#9), we inherited a large, fragmented codebase, in terms of the POWER/PPC64 port, with differently split headers, missing definitions, and redundant versions of common functions, which is rapidly becoming a maintenance nightmare, since we're constantly fixing-and-refixing stuff, and having to identify which layer is responsible for what.
The result of #9 was bemusing, in that backporting from the NICTA archive both fixed, and broke a lot of stuff, and trampled some of our existing fixes. :(
It could be useful to identify what is shared between the POWER3, and POWER4, and G5 implementations, and refactor them into a common files directory.
The PowerPC 32-bit port seems to be a completely-separate, unrelated codebase, with its own problems, but we'll deal with that, later, once we properly get around into adding Wii support (#17).
This effort would probably go a long way towards fixing #5, #6, #15, and #21, with one stone...
As part of the upstream Pistachio release, and the NICTA code drop (#9), we inherited a large, fragmented codebase, in terms of the POWER/PPC64 port, with differently split headers, missing definitions, and redundant versions of common functions, which is rapidly becoming a maintenance nightmare, since we're constantly fixing-and-refixing stuff, and having to identify which layer is responsible for what.
The result of #9 was bemusing, in that backporting from the NICTA archive both fixed, and broke a lot of stuff, and trampled some of our existing fixes. :(
It could be useful to identify what is shared between the POWER3, and POWER4, and G5 implementations, and refactor them into a common files directory.
The PowerPC 32-bit port seems to be a completely-separate, unrelated codebase, with its own problems, but we'll deal with that, later, once we properly get around into adding Wii support (#17).
This effort would probably go a long way towards fixing #5, #6, #15, and #21, with one stone...