Skip to content

Convert neovim config to Lua - #3

Merged
laurigates merged 1 commit into
masterfrom
neovim_lua
Mar 16, 2022
Merged

Convert neovim config to Lua#3
laurigates merged 1 commit into
masterfrom
neovim_lua

Conversation

@laurigates

Copy link
Copy Markdown
Owner

Since LSP configuration was already in Lua I decided to "convert" the
rest of my neovim configuration as well. I stumbled upon the
https://github.com/brainfucksec/neovim-lua repository which I took as a
base and have applied most of my previous plugins and configs that were
not present out of the box. I've left some out for now intentionally and
might add them back in if I miss them.

I had also thought about switching from vim-plug to packer.nvim to
manage plugins so that was achieved here as well. The statusline plugin
changed from lualine to feline which I had also considered testing.

I still left my old configs in place so either can be installed. I'll
get rid of the old stuff when I feel comfortable in doing so. In general
the new configs seem less convoluted than the old.

Since LSP configuration was already in Lua I decided to "convert" the
rest of my neovim configuration as well. I stumbled upon the
https://github.com/brainfucksec/neovim-lua repository which I took as a
base and have applied most of my previous plugins and configs that were
not present out of the box. I've left some out for now intentionally and
might add them back in if I miss them.

I had also thought about switching from vim-plug to packer.nvim to
manage plugins so that was achieved here as well. The statusline plugin
changed from lualine to feline which I had also considered testing.

I still left my old configs in place so either can be installed. I'll
get rid of the old stuff when I feel comfortable in doing so. In general
the new configs seem less convoluted than the old.
@laurigates
laurigates merged commit 7c5a89f into master Mar 16, 2022
@laurigates
laurigates deleted the neovim_lua branch March 16, 2022 17:23
laurigates added a commit that referenced this pull request Jun 29, 2026
…newcomers) (#286)

## Summary

Follow-up #3 of 3 from PR #283. The project pin `.claude/settings.json`
`enabledPlugins` had drifted to **40 keys** vs the **44-key** canonical
marketplace set, silently disabling 4 newcomers in web/remote sessions.

`enabledPlugins` merges via **whole-map replacement** — the project map
REPLACES the global map where `~/.claude` is unavailable
(web/remote/CI), so a plugin absent from the project pin is *disabled*
there regardless of global state. An exhaustive pin must therefore list
the full set.

## Changes

Added the 4 missing marketplace plugins with **dotfiles-appropriate
states** (intentionally diverging from the global overlay, which has all
four `true`):

| Plugin | State | Rationale |
|---|---|---|
| `comfyui-plugin` | `false` | domain-irrelevant to dotfiles |
| `foundryvtt-plugin` | `false` | domain-irrelevant to dotfiles |
| `software-design-plugin` | `false` | not needed here |
| `session-plugin` | `true` | broadly useful (global = true) |

## Verification

- `jq empty .claude/settings.json` → valid
- `jq '.enabledPlugins | length'` → **44**
- `./tests/test-claude-plugin-references.sh` → **4/4** (Check B green —
all 4 newcomers exist in the marketplace)
- `mise run test:plugin-refs` → green; pre-commit (incl. `check-json`)
clean

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
laurigates added a commit that referenced this pull request Jul 12, 2026
…tection to #3 (#321)

Two corrections to `git-hazards.md`, both earned in a session that
recovered an auto-closed PR (laurigates/claude-plugins#2049).

## Hazard #1 — the "immune to drift" claim was wrong

It claimed `git merge-tree --write-tree` containment is *"immune to
squash + drift"*. The squash half is right; **the drift half is not.**
Once `main` moves on over the same files, merging an already-merged
branch back would re-introduce its older versions, so the trees differ
and merge-tree reports **not contained** for work that fully landed. It
misreported three merged branches as unmerged during the audit.

Fixed by reordering the checks by real authority and adding `git
cherry`:

| Check | Strength |
|---|---|
| `gh pr list --state all --head <branch>` | **Authoritative** — a
MERGED PR settles it |
| `git cherry main <branch>` | Marks `-` on patch-equivalent commits —
survives squash **and** cherry-pick, ignores drift |
| `git merge-tree` containment | A match proves containment; **a
non-match proves nothing** |

## Hazard #3 — the auto-close is silent, and only a sweep finds it

No failed check, no notification; the PR list just looks one shorter.
#2049 sat stranded for a day, and the sweep afterwards found **26 dead
branches** — two carrying work that never had a PR opened *at all*.

An event handler on `pull_request: closed` cannot cover this: by the
time it fires the base ref is deleted (the reopen window is gone), and
it is structurally blind to never-PR'd branches. A scheduled sweep is
the only thing that works — `claude-plugins
scripts/check-stranded-work.sh` (laurigates/claude-plugins#2056).

Also records the **accident-vs-decision discriminator**: a
closed-unmerged PR whose base ref **404s** was auto-closed; one whose
base ref is **alive** was closed by a human. 11 of the 26 were
deliberate closes and must not be resurrected.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

https://claude.ai/code/session_01Ap1qbGLEkvv1gaG5hSasps

Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant