Today work has focused mostly on ctest stuff; I unexpectedly had to solve an issue with my editor and that consumed a lot of time. I also worked some on getting the diagnostic work on the Idris side of things.
The ctest extension progresses nicely. Today I have further simplified the second core algorithm. This one ensured I collected a list of FfiItems wrapped in a sum type for resolved or unresolved reexports.
Until today, I had attempted to think harder about the case in which the import statement is a grouped import. But I failed to realize that this is still a fairly simple case that does not require much more thought.
Recursing anew in an auxiliary algorithm, while appending the lists resulting from each element of the group gets the job done. Granted, I have yet to implement it, but I believe the idea is not flawed, and is fairly simple.
Beyond that, I think the second core algorithm is finished. Its task was to scan a given reexport, and yield a list of one of resolved or unresolved imports. The former contain another FfiItems with the resolved items.
In terms of side-project work, I could not do much today. I tried to debug an issue with my editor that seemed unrelated the the recent integrations I was building for Idris diagnostics.
As it turns out, there is a bug in the way certain buffer-restoring events triggered on a restart are handled. I tried to further diagnose the core issue by cloning and looking into the codebase.
Unfortunately, I could not get too far. The build requirements are a bother to meet without a package manager. I tried, but I decided it was not worth it to look any further. I have a manual workaround, so that will have to do.
What little work I got done in the Idris diagnostics side had me further refining the parsing logic for diagnostics. There is not much to say here. I have not finished it, so I will put these efforts on halt.
None.
As planned yesterday, I finished the second core algorithm plan today. I will now switch back to working on the first. I will probably have to design some data structure to hold state for the former’s result and the latter’s passes.
In terms of side-project work, tomorrow I will go back to working my way through the Idris crash course. I could not finish the editor integration work today, so I will probably come back to that at some point in the coming weeks.