This plan is very tentative and only the vision of the current maintainer. It is certainly open to suggestion and will very likely change as new issues surface.
In this release, we should rework the parser to make it less brittle in certain areas. There are several problems that are difficult to knock out with the current parser structure, probably including:
* FR #564641 (bad indentation causes parsing errors)
* FR #951761 (macro expansion in file listing)
* FR #555872 (#define macros with arguments)
* bug #1250441 (incomplete #v substitution support)
* FR #1740549 (handle forward references better)
Not sure how long this one will take, but it's probably something for the distant future. This release is intended to focus on gplink and really soup it up into an optimizing linker. The main goal is to make page and bank switching in the mid-range family completely a thing of the past, with the assembler (when requested) expanding every address access into a page/bank switch, and the linker computing the control flow and optimizing out any unnecessary switches. This or a future release could also really get every last drop out of the code by intelligently relocating linker sections to place tightly bound sections in the same space.
The Diablo project might be useful during this phase.