The VimL calendar-vim callbacks (nuwiki#diary#calendar_action /
calendar_sign) hardcoded wiki 0, so opening the calendar while editing
ifood_wiki still created/marked diary entries under the first wiki. (The
LSP-driven diary commands already follow the current buffer via its URI;
this brings the calendar in line.)
Track the wiki of the active buffer (nuwiki#diary#track_wiki, driven by
FileType/BufEnter autocmds in plugin/nuwiki.vim) and resolve the calendar
diary path against it, falling back to the first wiki. Verified the
calendar marker/open switch between wikis as you move between their
buffers. test-calendar-vim gains 4 multi-wiki tracking checks (16).
All keymap/config/calendar harnesses green on both clients.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
A Lua-only `setup({ wikis = … })` config is invisible to the VimL side:
the buffer helpers in autoload/nuwiki/config.vim and the vimwiki compat
shim (autoload/vimwiki/vars.vim, which third-party plugins like
vimwiki-sync call via vimwiki#vars#get_wikilocal) read g:nuwiki_wikis. So
a Neovim user who migrated from g:vimwiki_list to native setup({wikis})
would break vimwiki-sync (it'd resolve the default ~/vimwiki).
setup() now publishes the resolved wiki list to g:nuwiki_wikis when the
config came from setup() opts (the only case VimL can't already see —
g:nuwiki_wikis / g:vimwiki_list configs are global and visible). Verified
vimwiki#vars#get_wikilocal('path', n) returns the right roots from a
pure-Lua setup() config. test-global-shorthand grows two bridge checks
(17). Goldens + keymap harnesses green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The g:vimwiki_* drop-in was Vim-only: the Lua client read only setup()
opts and g:nuwiki_wikis, so swapping the vimwiki plugin for nuwiki under
lazy.nvim left the existing vim.g.vimwiki_list ignored and nuwiki pointed
at the default ~/vimwiki.
Mirror autoload/nuwiki/config.vim on the Lua side: config.lua now owns a
single resolver (M.wikis / M.scalar_globals) that resolves setup() wikis →
g:nuwiki_wikis → g:vimwiki_list (translating path->root, path_html->html_path,
template_*, ext->file_extension, …) and folds the global scalar defaults
(g:vimwiki_* then g:nuwiki_* then explicit setup() values; built-in defaults
excluded). lsp.lua (payload) and wiki_cfg/wiki_list (buffer commands) both
route through it, so the server and <Leader>ww agree on the wikis.
Extends test-global-shorthand (15 checks) with the vimwiki_list drop-in +
buffer-command resolution + native-wins cases. Config-parity goldens and
keymap harnesses green on both clients. README + known-issues updated
(drop-in is now Vim & Neovim).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The g:vimwiki_* drop-in only fed the LSP payload (lsp.vim s:settings).
The buffer-side commands — <Leader>ww (wiki_index), :VimwikiSearch,
auto_chdir, diary, completion — resolve the wiki list independently via
s:wiki_cfg / nuwiki#commands#wiki_list, which read only g:nuwiki_wikis /
g:nuwiki_wiki_root. So with a g:vimwiki_list config the server knew the
wikis but <Leader>ww fell back to the default ~/vimwiki and opened an
empty index.
Extract the resolution into a single source of truth,
autoload/nuwiki/config.vim (nuwiki#config#wikis / #scalar_globals), which
resolves g:nuwiki_wikis or an upstream g:vimwiki_list and folds the global
scalar defaults. lsp.vim (payload), commands.vim, and diary.vim all route
through it — also dedups the s:wiki_cfg copy that lived in both
commands.vim and diary.vim.
Regression guard in test-vimwiki-compat-vim: nuwiki#commands#wiki_list()
must resolve the g:vimwiki_list roots (21 checks). All Vim harnesses +
config-parity goldens green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
coc users had to hand-maintain a coc-settings.json `languageserver.nuwiki`
entry, which didn't track g:nuwiki_wikis / g:vimwiki_list / the global
shorthand and meant editing JSON per wiki. nuwiki now registers itself
programmatically: on the first .wiki buffer it calls
coc#config('languageserver.nuwiki', {…}) with the s:settings() payload
(same config the vim-lsp path sends). coc reacts to the languageserver
config change (onDidChangeConfiguration → registerClientsByConfig) and
starts the server for the open buffer.
Details:
- Reliable coc detection via :CocConfig (exists('*coc#config') can't be
trusted — it doesn't trigger autoload in Vim 9.2). The coc#config call
is wrapped in try/catch and autoloads coc.vim itself; falls back to the
printed snippet only if coc genuinely isn't there.
- Deferred to User CocNvimInit when coc's node service isn't up yet.
- Opt out with g:nuwiki_no_coc_register (then we just print the snippet).
Dogfooded in start-vim-coc.sh: dropped the manual coc-settings.json
languageserver block; the harness now relies on auto-registration from
g:nuwiki_* (real-user flow). New harness test-coc-register-vim (7 checks,
stubbed coc#config) asserts the injected payload incl. the folded global
shorthand; wired into CI. README coc section rewritten (zero-config).
Rust 573 passed, clippy clean, all harnesses green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Two related config-ergonomics features for vimwiki migrants, sharing the
same scalar-global machinery.
1. Upstream g:vimwiki_* drop-in (Vim client). When nuwiki isn't configured
natively, the Vim client reads g:vimwiki_list + g:vimwiki_* globals and
translates them into nuwiki's schema: per-wiki path->root,
path_html->html_path, template_*, css_name, auto_export/auto_toc,
syntax, ext->file_extension, index, diary_*, name. g:nuwiki_* always
wins. Lets a vimwiki user drop in nuwiki without rewriting config.
2. Global per-wiki defaults (both clients). A display/generation setting
given once at the top level — g:nuwiki_<key> / setup({<key>=…}) /
g:vimwiki_<key> — is folded into every wiki as a default; a per-wiki
value overrides it (vimwiki's model). Covers toc_header(_level),
toc_link_format, links_header(_level), tags_header(_level),
html_header_numbering(_sym), links_space_char, list_margin, listsyms,
listsym_rejected, and the auto_* toggles. Only values the user
explicitly set fold — built-in defaults don't (added config.user
tracking on the Lua side so a default toc_header_level=1 isn't pushed
onto every wiki). User config tables are never mutated.
Both clients kept in lock-step (config-parity golden enforces identical
payloads). New harnesses test-vimwiki-compat-vim (18) and
test-global-shorthand (8), wired into CI. Server test
vimwiki_compat_payload_sets_toc_level_per_wiki. Docs in README +
known-issues.md (drop-in is currently Vim-only). Rust 573 passed, clippy
clean, all harnesses green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
s:resolve_index referenced an undefined `a:file` (a copy-paste leftover
from wiki_root_for, whose parameter is named a:file) inside the
buffer-owning-wiki loop. Any get_wikilocal() call with no explicit index
while a wiki buffer was current — i.e. the ftplugin BufRead path — hit
`E121: undefined variable a:file`. The local is `l:file`, and the
surrounding `!empty(l:file)` guard already covers emptiness, so the
clause was both wrong and redundant; removed it.
Add development/tests/test-vars-vim.{sh,vim} covering the shim:
per-wiki-index resolution, global-default fallback, out-of-range clamp,
the no-index owning-wiki path (direct guard for this E121), and the
single-wiki shorthand. Wired into CI. 14/14 pass.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- create_link (default true): server resolve_target_uri now only synthesises a
missing-page URI when create_link is set; when off, following a link to a
non-existent page returns no location (no-op).
- dir_link: a [[dir/]] link now resolves to <dir>/<dir_link><ext> when dir_link
is set, else opens the directory itself. Uses the already-parsed
LinkTarget.is_directory.
- auto_header (default off): new wiki pages get a `= Stem =` header derived from
the filename, skipping the wiki index + diary index. Client-side BufNewFile
autocmd in both clients (init.lua + plugin/nuwiki.vim), with auto_header
handlers in both command layers (trailing-slash-robust root matching).
Tests: cmd.auto_header_inserts + cmd.auto_header_skips_diary_index (nvim) /
cmd.auto_header_skips_index (vim). Full rust suite + both harnesses green
(Neovim 307, Vim 301/18/21).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Both remaining mapping gaps (client-side, both clients):
- Visual <CR>: upstream binds it to NormalizeLinkVisualCR (== visual `+`).
Added `x <CR>` -> normalize_link(true) (wrap the selection as a wikilink)
in both clients.
- Insert <S-CR>: upstream's multiline-list-item continuation (VimwikiReturn
2 2 -> kbd_cr with no new marker). New smart_shift_return (both clients,
<expr> insert mapping): on a list item returns <CR> + spaces aligning under
the item text (marker width + 1, +4 for a `[ ] ` checkbox), no marker; off a
list item, a plain <CR>. Mirrors smart_return's auto-indent handling.
Tests: map.visual_cr_wraps_selection + map.insert_shift_cr_multiline
(test-keymaps.lua, the latter exercising the handler directly since <S-CR>
doesn't round-trip headless feedkeys), and map.visual_cr_wraps_selection +
cr.shift_cr_multiline{,_checkbox} (test-keymaps-vim.vim). Docs (README +
doc/nuwiki.txt) and the gap doc updated. Neovim 305, Vim 299/18/21 pass.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Re-audit of the just-landed command work found one real bug I introduced
and a convention gap:
- VimwikiSearch / VWS raised a raw E480 on a no-match search; upstream
reports it gracefully. Both clients now catch it (try/catch in Vim,
pcall in Neovim) and emit "nuwiki: no match for <pat>".
- Added a NuwikiSearch alias alongside VimwikiSearch/VWS (every other
command carries a Nuwiki* form).
Gap-doc note corrected: nuwiki's search populates + opens the location
list (shows all matches) where upstream jumps to the first match — logged
as a minor presentation divergence rather than claiming exact parity.
Config + mappings re-audits came back clean (no new gaps, no regressions;
table-mapping rewire verified non-recursive, gLH/gLL/gLR present).
Tests: cmd.VimwikiSearch_no_match_graceful, surface.{Vimwiki,Nuwiki}Search.
Neovim 301, Vim 294/18/21 pass.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
All six remaining command gaps (client-side; no server change):
- VimwikiRenameFile: bare -> -nargs=? (arg accepted-ignored; rename still
delegates to the LSP rename prompt). No more E488.
- VimwikiVar / NuwikiVar: get/set the client config (upstream vimwiki#vars#cmd).
No arg prints config (Vim: g:nuwiki_*; Neovim: resolved setup() opts), one arg
gets a key, two+ sets it. Global command, both clients.
- VimwikiReturn / NuwikiReturn: command form of the smart <CR> continuation —
enters insert at EOL and fires the <CR> mapping (return_cmd / vimwiki_return).
Upstream's mode-flag args accepted but unused.
- VimwikiTableAlignQ vs AlignW: corrected — upstream's gqq/gww are identical on
a table; the only difference is the non-table fallback (native `normal! gqq`
vs `gww`). New table_align_or_cmd(cmd) aligns on a table row, else runs
`normal! <cmd>`. Q/gqq/gq1 -> gqq; W/gww/gw1 -> gww; NuwikiTableAlign -> gqq.
- VimwikiSearch / VWS: scope the lvimgrep to the current wiki's root (resolved
client-side) over <root>/**/*<ext>, not CWD `**`; empty pattern reuses the
last search. Both clients.
- VimwikiIndex family: add global Vimwiki/NuwikiIndex + TabIndex entry points
(plugin/nuwiki.vim + init.lua _setup_global_commands, with -count), so a wiki
opens from any buffer. Buffer-local copies still shadow inside wiki buffers.
Tests: surface.{Vimwiki,Nuwiki}{Return,Var}, cmd.VimwikiVar_sets_*,
cmd.global_entry_points (both harnesses). Docs (README + doc/nuwiki.txt) and the
gap doc updated. Neovim 299, Vim 291/18/21 pass.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Closes the actionable fifth-pass findings (gap doc updated):
- VimwikiGoto / NuwikiGoto: -nargs=1 -> -nargs=* in all four defs. Was a
real bug — `:VimwikiGoto` raised E471, `:VimwikiGoto My Page` raised E488,
and the handler's empty-arg prompt fallback was unreachable. Both clients
already prompt on empty, so -nargs=* (keeping -complete=pages) restores
upstream behavior.
- autowriteall (upstream default ON): while a wiki buffer is current, mirror
Vim's global 'autowriteall' (auto-save on page switch / :make), restoring
the prior value on leave. Neovim via ftplugin.lua setup_autowriteall
(BufEnter/BufLeave) gated by the `autowriteall` setup option; Vim via the
NuwikiAutoWriteAll augroup gated by g:nuwiki_autowriteall.
- table_auto_fmt (upstream default ON): re-align the table under the cursor
on InsertLeave. Both clients, gated by `table_auto_fmt` / g:nuwiki_table_auto_fmt,
guarded on the cursor line containing `|` before the async table_align.
Also logged (no code): VimwikiRenameFile missing -nargs=? (subsumed by the
LSP-rename divergence), table_reduce_last_col, user_htmls, hl_headers/
hl_cb_checked, and a commentstring value-mismatch note. Mappings audit clean.
The Neovim harness disables table_auto_fmt in setup (its async re-align would
perturb the deterministic table-nav keystroke cases); its wiring is verified
on a throwaway scratch buffer instead. Docs (README + doc/nuwiki.txt) updated.
Tests: cmd.VimwikiGoto_accepts_multiword, config.autowriteall_applied,
config.table_auto_fmt_autocmd. Neovim 293, Vim 285/18/21 pass.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Closes four low-risk P3 parity items (gap doc updated):
- VimwikiShowVersion / NuwikiShowVersion: new global command (both clients)
echoing g:nuwiki_version (0.1.0) + the host editor, via
nuwiki#commands#show_version / commands.show_version.
- gLH / gLL / gLR: upstream's case-variant aliases of gLh/gLl/gLr (dedent /
indent whole item, renumber all lists) added in both clients.
- VimwikiSearch / VWS: -nargs=1 -> -nargs=* (Vim + Neovim), so a multi-word
search no longer raises E488 and an empty invocation reuses the last search
pattern. (The lvimgrep-vs-vimwiki-engine difference stays open.)
- VimwikiTableAlignQ / AlignW / NuwikiTableAlign: bare -> -nargs=?, so
`:VimwikiTableAlignQ 2` no longer raises E488 (optional column arg is
accepted-and-ignored; the gqq-align vs gww-align-without-resize split
stays open).
Docs (doc/nuwiki.txt + README) updated for the command and the aliases.
Tests: surface.{Vimwiki,Nuwiki}ShowVersion, map[n].gL{H,L,R} (both
harnesses), cmd.search_and_tablealign_nargs (vim). Neovim 291, Vim
282/18/21 pass.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Implements the remaining fourth-pass findings (gap doc updated):
- list_margin: rework to upstream's buffer-side meaning. Drop the
nuwiki-only HTML em-margin (renderer field/method + render_page_html
param removed) and prepend max(0, list_margin) leading spaces to every
generated bullet in build_toc_text / build_links_text /
build_tag_links_text / diary::build_index_body; headings stay at col 0.
From<RawWiki> derives 0 for markdown wikis when unset. Negatives can't
resolve 'shiftwidth' server-side, so they collapse to zero indent
(documented divergence).
- diary_months: per-wiki Vec<String> (default 12 English names), threaded
into the diary-index month labels; missing/empty slots fall back to the
English name.
- diary_caption_level: widen u8 -> i8 so vimwiki's -1 (min: -1) parses;
build_index_body clamps < 0 to base tree level 0.
- VimwikiRemoveDone: regain upstream's -range. All four defs are now
-bang -range, dispatched via remove_done(bang, range, l1, l2) in both
clients: ! -> whole buffer, explicit range -> new list_remove_done_range
({range:[l1-1,l2-1]}), else current list. Server remove_done_edit gained
an Option<(u32,u32)> range that filters whole-doc victims by start line.
markdown_header_style is deferred: the generators emit vimwiki syntax only
(caption_line never writes markdown headers), so there's no markdown header
to attach the style to. Logged as the "generated-content is vimwiki-only"
intentional divergence pending a later markdown generated-content effort.
Tests: list_margin indent, markdown list_margin-0 default, diary_months
custom + fallback, negative caption_level clamp/parse, ranged remove-done
(server + both keymap harnesses). 553 lsp/core tests pass; clippy clean.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Upstream's :VimwikiToggleRejectedListItem is -range and binds glx in
visual mode; nuwiki's four defs were bare, so a ranged invocation raised
E481 and acted cursor-only. Same -range class already fixed for
ToggleListItem/Increment/Decrement -- ToggleRejected was missed.
All four defs (Vim+Neovim x Vimwiki*/Nuwiki*) are now -range, routed
through a new reject_list_item_range(l1, l2) looping over_range in both
clients, mirroring toggle_list_item_range. Visual glx stays cursor-only,
consistent with its <C-Space>/gln/glp siblings.
Also logs the 2026-06-02 re-audit findings in development/vimwiki-gap.md:
RemoveDone lost -range (low pri); config gaps list_margin
semantics/markdown-default, diary_months, markdown_header_style,
diary_caption_level=-1; and the path_html/template_path default-location
divergences. Mappings audit came back clean.
Tests: cmd.reject_list_item_range (test-keymaps.lua),
cmd.VimwikiToggleRejectedListItem_accepts_range (test-keymaps-vim.vim).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add headless harnesses covering the calendar-vim integration, wired into
CI alongside the existing keymap/config suites:
- test-calendar-vim.sh/.vim (+ -optout.vim): pure-VimL diary hooks
(calendar_sign markers, calendar_action path resolution from wiki root,
calendar window close), plugin/nuwiki.vim hook auto-wiring that
overrides calendar-vim's defaults, and opt-out via g:nuwiki_use_calendar
and g:nuwiki_no_calendar. Uses a stub calendar-vim on the runtimepath.
- test-calendar.sh/.lua: Neovim setup() wiring + window-close under the
real window API, and opt-out via use_calendar = false.
Fixes surfaced by the suite:
- diary.vim: new diary entries set filetype 'vimwiki' (was the raw
extension, e.g. 'wiki'); the stray FileType autocmd also blocked the
calendar window from closing.
- init.lua: drop the redundant Neovim FileType autocmd (plugin/nuwiki.vim
already wires both editors) and translate use_calendar=false into
g:nuwiki_no_calendar so the opt-out actually disables wiring.
2da2168's diary test edits landed unevenly: the lua
cmd.VimwikiMakeDiaryNote_has_count case got inserted twice, and the two vim
cases (cmd.Vimwiki{MakeDiaryNote,DiaryIndex}_accepts_count) the commit message
referenced never landed (the Edit had no-matched). Remove the lua duplicate and
add the two vim cases before the harness wrap-up. lua 284, vim 274/18/21, all
0 failed.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
874bdd0 committed the diary -count feature with several Edits that had
silently no-matched, leaving the tree non-compiling and the clients
inconsistent (CI run 236 failed). This completes it:
- commands.rs: add the `wiki` field to OptionalUriArg (the previous edit
targeted a wrong struct name) and pass None to the six resolve_diary_wiki
callers that take no selector (date/list/step paths). Server builds clean.
- autoload/nuwiki/commands.vim + lua/nuwiki/commands.lua: actually thread the
count through s:diary_open / _diary_open and the diary_today/today_tab/
yesterday/tomorrow/index handlers (these edits had failed before, so the
ftplugin defs were calling handlers that ignored/rejected the arg).
- Add the test cases the prior commit referenced but never landed:
cmd.VimwikiMakeDiaryNote_has_count (test-keymaps.lua) +
cmd.Vimwiki{MakeDiaryNote,DiaryIndex}_accepts_count (test-keymaps-vim.vim).
- Fix the OptUriArg→OptionalUriArg name in the gap-doc note.
Verified with CI flags: workspace test/clippy/fmt clean; lua 284, vim 272/18/21,
all 0 failed; the three new diary cases pass.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Closes the last command-cluster gap. Upstream's diary-note family
(:Vimwiki{Make,TabMake,MakeYesterday,MakeTomorrow}DiaryNote + VimwikiDiaryIndex,
and the Nuwiki* forms) carry -count=0 where the count selects the wiki number;
nuwiki's were bare. Implemented as a real end-to-end selector, reusing existing
infrastructure:
- Server (commands.rs): OptUriArg gained an optional `wiki` selector;
resolve_diary_wiki(backend, uri, wiki) now prefers it (via the existing
resolve_wiki_selector, which already handles a 0-indexed wiki number) over
the buffer URI. Used by diary_open_relative + diary_open_index.
- Clients: s:diary_open / _diary_open (both) and diary_today/today_tab/
yesterday/tomorrow/index take a count and send {wiki: count-1} when >0. All
22 diary-note + DiaryIndex command defs across the 4 contexts are now
-count=0 passing <count>. diary_next/prev ignore the count.
Tests: cmd.VimwikiMakeDiaryNote_has_count (test-keymaps.lua) +
cmd.Vimwiki{MakeDiaryNote,DiaryIndex}_accepts_count (test-keymaps-vim.vim).
Workspace test/clippy/fmt clean with CI flags; lua 284, vim 272/18/21.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Closes the three new command gaps from the 2026-05-31 re-audit (all real,
client-side, both clients × Vimwiki/Nuwiki):
- VimwikiChangeSymbolTo / ListChangeSymbolI / NuwikiChangeSymbol are now
-range -nargs=1 via a new list_change_symbol_range(symbol,l1,l2) helper that
loops over_range; was -nargs=1 only → E481 on a visual selection.
ChangeSymbolInListTo/InList stay range-less.
Bonus: the Neovim VimwikiChangeSymbolInListTo + NuwikiChangeSymbolInList defs
passed whole_list=false, changing only the current item — corrected to true.
- VimwikiGenerateLinks / NuwikiGenerateLinks are now -nargs=?; with an arg the
client dispatches the existing server command nuwiki.links.generateForPath
({path}) for real subtree-scoped link generation, else nuwiki.links.generate.
Was bare → E488 on an arg.
- Added NuwikiGenerateTags (Vim + Neovim) mirroring VimwikiGenerateTags, with
-nargs=? -complete=…tags — closes the :Vimwiki*↔:Nuwiki* alias asymmetry.
Tests: cmd.ChangeSymbolTo_range_sets_marker, cmd.GenerateLinks_accepts_optional_path,
cmd.NuwikiGenerateTags_exists (test-keymaps.lua, 283 pass) +
cmd.VimwikiChangeSymbolTo_accepts_range, cmd.VimwikiGenerateLinks_accepts_path,
cmd.NuwikiGenerateTags_exists (test-keymaps-vim.vim, 272/18/21). fmt clean.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The previous commit's test-insert Edit silently no-matched (wrong anchor
text), so cmd.VimwikiFollowLink_wraps_bare_word and
cmd.VimwikiSplitLink_wraps_bare_word were referenced in the commit message and
gap doc but never added. Add them now at the correct anchor; the lua harness
goes 278 → 280, both new cases pass.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Upstream attaches -complete= to several commands; nuwiki had none, so <Tab>
at the : line never completed page/tag/colour names. New
autoload/nuwiki/complete.vim adds pure client-side, synchronous completers
(config + filesystem, never the LSP — so they work under vim-lsp, coc and
Neovim), wired via -complete=customlist into all command contexts:
- nuwiki#complete#pages → VimwikiGoto/NuwikiGoto (glob wiki roots for page names)
- nuwiki#complete#colors → VimwikiColorize/NuwikiColorize (color_dic keys +
colours already used in buffer color:… spans)
- nuwiki#complete#tags → VimwikiSearchTags/GenerateTagLinks/GenerateTags (+Nuwiki)
(scan wiki files for :tag: runs — the tag index is server-side and a
synchronous completer can't query it, so we read the same source it indexes)
Documented divergence: no file completer for VimwikiRenameFile — nuwiki
delegates renaming to the LSP rename request (no filename arg to complete),
unlike upstream which renames itself.
Tests: complete.{pages_globs_and_filters,tags_scans_wiki_files,
colors_from_buffer_spans} (test-keymaps-vim.vim). Both harnesses green.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Upstream's :Vimwiki[V]SplitLink take optional reuse_other_split_window +
move_cursor flags (the link is always the one under the cursor); nuwiki's defs
took no args, so `:VimwikiSplitLink 0 1` raised E488. All eight defs (both
clients × Split/VSplit × Vimwiki/Nuwiki) are now -nargs=*, restoring signature
parity for migrated configs/scripts.
Documented divergence: the two flags are accepted but not acted on — nuwiki
always opens a fresh split and follows in it. Honoring move_cursor/reuse would
require threading a completion callback through the async LSP follow path,
disproportionate for this rarely-used arg form; the no-arg path (99% case) is
fully equivalent. The gap-doc note (which wrongly described the args as a
"target link") is corrected.
Test: cmd.VimwikiSplitLink_accepts_args (test-keymaps-vim.vim, no E488).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Upstream is `-range -nargs=+` (change_level(line1, line2, direction,
plus_children)); nuwiki was -nargs=? and range-less, so `:1,3VimwikiListChangeLvl
increase 0` raised E481/E488 and the level change never spanned a selection.
All four defs (Vim+Neovim × Vimwiki+Nuwiki) are now -range -nargs=+ and forward
<line1>, <line2>, <f-args>. list_change_lvl(line1, line2, direction,
[plus_children]) (both clients) parses the required direction + optional subtree
flag and applies the change to every list item in the range via the existing
over_range helper.
Tests: cmd.ListChangeLvl_range_indents (test-keymaps.lua) +
cmd.VimwikiListChangeLvl_accepts_range_and_args (test-keymaps-vim.vim).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Two command-attribute parity fixes (audited against upstream vimwiki master):
- VimwikiTable / NuwikiTable were -nargs=1 and called table_insert() with no
args, so `:VimwikiTable 4 3` raised E488 and cols/rows were unreachable.
All four defs (Vim+Neovim × Vimwiki+Nuwiki) are now -nargs=* and forward
<f-args>; the backing table_insert already parsed cols (a:1) + rows (a:2).
- VimwikiIndex/TabIndex (and Nuwiki forms) switch from bare -count to
upstream's -count=0. Behavior is unchanged (both default count to 0;
wiki_index already maps a count to the 1-indexed wiki number) — this just
matches the upstream declaration exactly.
Tests: cmd.VimwikiTable_passes_cols_and_rows (test-keymaps.lua — asserts a
4-col/3-row table) and cmd.VimwikiTable_accepts_cols_rows (test-keymaps-vim.vim
— no E488). gap doc updated.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
A map_prefix parity audit against upstream vimwiki found the Neovim global
entry-point block in lua/nuwiki/init.lua carried the same collision the Vim
global + buffer-local maps were already fixed for: <prefix><Leader>t was
bound to tomorrow and <prefix><Leader>m (tomorrow) was missing entirely.
Now <prefix><Leader>t opens today in a new tab and <prefix><Leader>m is
tomorrow — consistent with plugin/nuwiki.vim, the buffer-local maps, and
upstream (VimwikiTabMakeDiaryNote / VimwikiMakeTomorrowDiaryNote).
Adds global_maps.diary_tab_and_tomorrow_targets to test-keymaps.lua (queries
in a scratch buffer so buffer-local maps don't shadow the globals). Logs the
fix plus a new "RSS feed structure fidelity" gap (the audit confirmed the
custom_wiki2html arg contract and base_url scope are faithful; only the RSS
document shape diverges) in development/vimwiki-gap.md.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Mirror vimwiki's g:vimwiki_map_prefix. The whole <Leader>w* family was
hardcoded across the map layer; it is now built from a configurable prefix:
- Neovim: `map_prefix` setup() option (default '<Leader>w'), read by
lua/nuwiki/keymaps.lua (buffer-local) and lua/nuwiki/init.lua (global
entry points).
- Vim: g:nuwiki_map_prefix, applied via :exe in plugin/nuwiki.vim (global)
and ftplugin/vimwiki.vim (buffer-local).
Setting a custom prefix relocates <prefix>{w,t,s,i,n,d,r,c,h,hh,ha} and
<prefix><Leader>{w,y,t,m,i}, leaving nothing under the old <Leader>w*.
Non-prefix maps (<CR>, gl*, headers, …) are untouched.
Tests: new test-keymaps-vim-prefix.vim harness (21 cases, wired into
test-keymaps-vim.sh) + map_prefix.relocates_wiki_family in test-keymaps.lua.
Docs: README, doc/nuwiki.txt, lua config comment, vimwiki-gap.md ticked.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Implements the four items previously left as "needs handler work" — all
fully wired, not cosmetic.
- ToggleListItem / IncrementListItem / DecrementListItem are now -range
(both clients). The client loops the per-line op over [<line1>,<line2>]
(toggle_list_item_range / list_cycle_symbol_range). The server's edits are
version-less single-line WorkspaceEdits, so each line applies independently
— no server protocol change. `:'<,'>VimwikiToggleListItem` toggles the whole
selection.
- VimwikiRebuildTags gains -bang: `:…RebuildTags!` passes { all: true } and
the server (tags_rebuild) re-indexes every configured wiki via
wikis_snapshot(), not just the current one.
- VimwikiNormalizeLink is -nargs=?: with `1` (upstream's visual flag; the
x-mode `+` mapping now passes it) it wraps the '< / '> selection as
[[selection]] (new wrap_visual_as_wikilink in both clients).
- VimwikiCheckLinks is -range: a ranged invocation filters the broken-link
report to the current buffer's selected lines (client-side filter in
results_to_qf / check_links).
Tests: cmd.toggle_list_item_range + cmd.normalize_link_visual_wraps_selection
(test-keymaps.lua), normalize.visual_wraps_selection (test-keymaps-vim.vim).
README + doc/nuwiki.txt + gap doc updated. fmt/clippy clean; nvim 274,
vim 266+18, all Rust tests + config-parity green.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Clean, fully-wired parity fixes (skipping the accept-but-ignore ones):
- <D-CR> (macOS Cmd+Return) → follow_link_drop, alongside <C-S-CR>, in both
clients. Upstream binds both to VimwikiTabDropLink.
- plugin/nuwiki.vim global map block: fix the same diary collision the
buffer-local maps had — <Leader>w<Leader>t now opens today in a new tab and
<Leader>w<Leader>m (tomorrow) is added.
- Neovim TabIndex: add -count to NuwikiTabIndex and switch both
Vimwiki/NuwikiTabIndex from vim.v.count (always 0 in command context) to
<count>, so a count is actually honored.
- Converge VimwikiTableMoveColumn{Left,Right}: the Vim-branch commands now call
the table_move_column_left/right aliases, matching the Neovim branch.
- VimwikiColorize/NuwikiColorize: -nargs=1 → -nargs=* (all four defs) for
upstream parity; a bare :VimwikiColorize now prompts for the colour.
Docs (README + doc/nuwiki.txt) updated; mapping surfaces gain <D-CR> and
<Leader>w<Leader>m in both harnesses. vim 265+18, nvim 272, all green.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Two upstream-parity keymap fixes (confirmed against upstream
ftplugin/vimwiki.vim):
1. <Leader>w<Leader>t collision — nuwiki bound both <Leader>w<Leader>t and
<Leader>w<Leader>m to tomorrow's diary. Upstream: <Leader>w<Leader>t opens
today's diary in a NEW TAB, <Leader>w<Leader>m opens tomorrow. <Leader>w
<Leader>t now calls diary_today_tab; <Leader>w<Leader>m stays tomorrow.
2. <M-CR> badd-link — adding a link target to the buffer list was mouse-only
(<MiddleMouse>) / command-only (:VimwikiBaddLink). Upstream binds
<M-CR> -> VimwikiBaddLink. Added normal-mode <M-CR> -> badd_link to the
links group of both clients.
Both clients (lua/nuwiki/keymaps.lua, ftplugin/vimwiki.vim); README + doc/
nuwiki.txt updated. Tests: map[n].<M-CR> in both harnesses;
diary.leader_t_opens_today_in_tab / diary.leader_m_opens_tomorrow (vim).
nvim 270, vim 263+18, all green.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
A literal `<span style="color:…">` in prose (the sample wiki's own colorize
instructions used one, with a `…` ellipsis as the colour) was matched by the
scanner and fed to `highlight guifg=…`, which raised E254 on the un-silenced
guifg line. That aborted the whole nuwiki#colors#refresh(), so on the Vim
clients NO spans got concealed (start-vim) and follow-link crashed mid
FileType autocmd (start-vim-coc). Neovim happened to dodge it depending on
which buffer was scanned.
s:define() now:
- validates the colour (only #hex or an alphabetic name) and skips anything
else — rgb()/hsl() functions and prose examples no longer create bogus
conceal regions or errors;
- runs both `highlight` calls under `silent!`, so a valid-format but unknown
name can't abort the refresh either (tags still conceal, text just isn't
recoloured).
Also reworded the sample wiki's colorize note so it no longer embeds a
literal span tag.
Test: colorize.skips_unallocatable_colour (vim) — an rgb() span before a real
one: refresh doesn't error, the rgb span isn't concealed, the real span still
is. nvim 269, vim 260+18, all green.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
A :VimwikiColorize'd word showed the literal
`<span style="color:NAME">TEXT</span>` markup. Now the tags are concealed
and TEXT is painted in NAME, so the buffer shows just the coloured word —
matching the HTML export.
New autoload/nuwiki/colors.vim: nuwiki#colors#refresh() scans the buffer
and, per distinct colour, defines a syntax region that conceals the
`<span …>` / `</span>` tags (concealends) and highlights the body with the
span's actual colour (guifg always; ctermfg for named colours). Idempotent
(clears the group before redefining) and tracked in b:nuwiki_color_seen.
Refresh is driven from three places so it's both correct and immediate:
- syntax/vimwiki.vim: initial pass + re-establish after :colorscheme reload
(resets the per-buffer cache since :syntax clear drops the regions);
- ftplugin TextChanged/InsertLeave autocmd for spans that arrive via paste/
undo/external edits;
- colorize() itself calls refresh() right after wrapping, so a freshly
colorized word conceals instantly (TextChanged doesn't fire reliably mid
command / in headless ex-mode).
The LSP @vimwikiColor token no longer links to Constant, so in Neovim it
doesn't override the real per-span colour on the concealed text. Pure syntax
+ :highlight — works identically in Vim, coc.nvim, and Neovim, no LSP needed.
Tests: colorize.conceal_hides_tags_shows_text (both harnesses) and
colorize.command_conceals_immediately (vim). Sample wiki's colorize section
notes the conceal. nvim 269, vim 259+18, config-parity green.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Selecting text and running :VimwikiColorize {color} raised
"E481: No range allowed" — Vim prepends '<,'> to the command when invoked
from a visual selection, but the command was defined without -range.
All four colorize command defs (Vim + Neovim, Vimwiki* + Nuwiki*) are now
-range and forward the range count to the handler. A ranged (visual)
invocation wraps the '< / '> selection; a bare invocation wraps the cword.
The x-mode <Leader>wc mapping passes the visual flag explicitly.
The Vim-branch colorize() gained selection support (it was cword-only),
matching the Neovim branch, so both clients wrap a visual selection
identically. visual_selection_range() (Lua) no longer gates on
vim.fn.visualmode() — the explicit visual flag now carries that intent, and
the marks check guards unset selections.
Tests: cmd.Colorize_range_wraps_selection (nvim) and
colorize.{command_wraps_cword,visual_wraps_selection,ranged_command_no_error}
(vim). nvim 268, vim 257+18, all green.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The Neovim command defs for :VimwikiColorize / :NuwikiColorize were
-nargs=1 but called colorize() without <q-args>, so the colour name was
silently dropped and the user was prompted instead (the Vim branch passed
it correctly). Pass <q-args>.
While verifying, found a second bug in the Lua colorize(): it inferred
visual-vs-normal from vim.fn.visualmode(), which returns the *last* visual
mode used in the session along with stale '< / '> marks. So running
:VimwikiColorize in normal mode after any earlier visual selection wrapped
that leftover selection instead of the word under the cursor (produced an
empty, misplaced span). colorize(color, visual) now takes an explicit
visual flag, passed only by the x-mode <Leader>wc mapping; the command and
the normal-mode mapping always wrap the cword.
Tests: cmd.VimwikiColorize_uses_arg, cmd.NuwikiColorize_uses_arg,
cmd.Colorize_ignores_stale_visual_selection, colorize.visual_wraps_selection
in test-keymaps.lua. nvim suite 267, vim suite 254+18, all green.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The Vim follow-link path delegated to the LSP client's jump UI
(:LspDefinition / coc jumpDefinition), which broke two ways for wiki
link-following:
1. **No create-on-follow (vim-lsp + coc).** vim-lsp's location handler
readfile()s the target to build a quickfix entry; for a not-yet-created
page that throws E484 and aborts the jump, so <CR> on [[NewPage]] did
nothing. The server was correct throughout — textDocument/definition
returns a *synthesised* location for missing pages (verified over
JSON-RPC).
2. **Wrong window placement (coc).** Without an explicit open command coc
fell back to coc.preferences.jumpCommand (e.g. a tab command).
Stop delegating: nuwiki#commands#follow_link_or_create() /
follow_link_drop() now resolve textDocument/definition directly and open
the URI themselves via a shared s:open_definition(open_cmd) helper — :edit
for <CR>, `tab drop` for <C-S-CR>. :edit opens a buffer for a missing path
(save creates it, matching vimwiki) and gives exact current-window
placement regardless of the user's coc jumpCommand. Removes
s:jump_definition / s:drop_from_response.
Also fix the coc action name: CocAction('getDefinition') does not exist
(E605 "Action getDefinition does not exist") — the registered action is
'definitions'. Fixed in s:open_definition and badd_link.
coc config delivery: the setup guidance shipped without
initializationOptions, so coc users never sent wiki_root and the server
resolved links against its default root (existing pages flagged broken,
follow-to-create landed in the cwd). The printed snippet
(autoload/nuwiki/lsp.vim) now emits initializationOptions populated with
the resolved settings, and the README coc snippet documents it as required.
Add development/start-vim-coc.sh — a coc.nvim dev launcher (sibling of
start-vim.sh) that wires coc's prebuilt release branch + a generated
coc-settings.json (with initializationOptions), deliberately omitting
vim-lsp so the coc path is exercised; doubles as a reproducer.
Tests: new cr.follow_opens_missing_page_in_current_window case in
test-keymaps-vim.vim (stubs CocAction to a missing page, asserts
current-window :edit). Keymap suite 254+18 green; config-parity (Vim +
Neovim) green. Neovim path unchanged and re-confirmed.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Swap the gl<Space>/gL<Space> surface assertions for bare gl/gL; add behavioral
cases for gl (item) and gL (whole list) remove-checkbox, and drive remove-done
through :NuwikiRemoveDone / :NuwikiRemoveDone! now that it is command-only.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Add SplitLink/VSplitLink/TabnewLink/TabDropLink/GoBackLink to both
keymap harness command lists, plus a smoke check that
nuwiki.commands.follow_link_drop is callable.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The plain-Vim ftplugin only exposed a whole-layer
g:nuwiki_no_default_mappings gate, so Vim users could not drop a single
keymap group the way Neovim users can via mappings.<group> = false. Add
g:nuwiki_no_{wiki_prefix,links,lists,headers,table_editing,diary,
html_export,text_objects}_mappings, each wrapping its group block, to
reach parity with lua/nuwiki/keymaps.lua. Cover them with a dedicated
opt-out harness and document them in the README and help file.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Both editor harnesses now emit diagnostic.link_severity and exercise a
non-default override ('error') in the sample scenario, diffed against the
shared golden. Drop the README "not wired" caveat now that the value flows
through, and document the g:nuwiki_link_severity global.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Expose the init_options payload builders publicly in both clients
(M.init_options/M.server_settings in Lua, nuwiki#lsp#settings() in Vim) so
they can be inspected. Add editor harnesses that dump each client's
server-bound config as deterministic key=value lines and diff against a
shared golden file: nvim == golden and vim == golden together prove the two
clients send identical config. Add a server-side test asserting the Vim flat
shape and the Neovim {nuwiki:{...}} wrapper desugar to the same WikiConfig.
Wire both harnesses into CI.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Update the README command tables for the full :Nuwiki*/:Vimwiki* surface
and add command-registration assertions (removeCheckbox, catUrl, and the
other parity entry points) to both the Vim and Neovim keymap harnesses.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
gffranco