Architecture review
01M1XJZSDGJ90YGY9ZP6HBYPS9
- Revision
- 1
- Type
- block
- Status
- done
- Assignee
- nobody
- Estimate
- none
- Parent
- none, which is loose rather than invalid
- Labels
No timestamps and no authors. Git holds both, and the log on the overview is where they are read from.
None yet.
The items architecture reviews on 2026-08-29 and 2026-09-02 turned up, the second run
against LCOM over shared module state, a token clone detector, and fan-in and fan-out on
ns. None of them is a bug.
Read an item against the code before working it. Item 15 undercounted its sites by five
and item 18 undercounted its closure by half. ns.db measured clean, 215 keys with four
read outside the folder that declares them, and file-level LCOM4 is 1 almost everywhere.
The numbers on these cards are the numbers they carried in todo.md, kept as labels,
because commit messages reference them and a number that moved would break every one.
Nothing said yet.