rebrand to notevi: one CLI over the jj sidecar

vr and vrsite become a single binary. vrsite/ folds into a web package in
one module (0x4200.cafe/notevi); "notevi web" serves and exports exactly
what vrsite did, and "notevi read/grep/note/query" is unchanged.

The sidecar is renamed with it: notevi_log, notevi-log.jsonl, and the
description "private: notevi log". The pre-rebrand names are still
recognized, so an old repository opens and reads; it is renamed in place
on the first write, or up front with "notevi migrate DIR...".

That rename cannot be a single mv inside jj run. jj only auto-tracks a
*new* file in the run working copy below a size limit it does not take
from the command line, so writing a whole log under a name the change has
never held is silently dropped while jj reports success. ensureLogFile
creates the file empty first and lets every later byte be a modification
of a tracked file, which snapshots at any size; that also fixes the same
latent bug when importing a large legacy vr-log.jsonl.

Adds a bem-te-vi mark (favicon and nav brand) and a README.

Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>

1 files changed, 65 insertions, 0 deletions

diff --git a/web/assets/zig.scm b/web/assets/zig.scm
new file mode 100644
index 0000000..a85383a
--- /dev/null
+++ b/web/assets/zig.scm
@@ -0,0 +1,65 @@
+; Zig highlights, from tree-sitter-grammars/tree-sitter-zig v1.1.2
+; queries/highlights.scm (MIT). Reordered for later-patterns-win, capture names
+; mapped onto zed syntax keys — upstream's keyword.* split collapses to
+; @keyword, which zed themes colour as one — and the #lua-match? identifier
+; heuristics rewritten as #match?, which this engine evaluates.
+
+["(" ")" "[" "]" "{" "}"] @punctuation.bracket
+(payload "|" @punctuation.bracket)
+[";" "," "." ":" "=>" "->"] @punctuation.delimiter
+["=" "==" "!=" "<" ">" "<=" ">=" "+" "+%" "+|" "-" "-%" "-|" "*" "*%" "*|"
+ "**" "/" "%" "&" "|" "^" "~" "!" "||" "<<" "<<|" ">>" ".." ".*" ".?" "?"
+ "+=" "+%=" "+|=" "-=" "-%=" "-|=" "*=" "*%=" "*|=" "/=" "%=" "&=" "|=" "^="
+ "<<=" "<<|=" ">>="] @operator
+
+["addrspace" "align" "allowzero" "and" "asm" "async" "await" "break"
+ "callconv" "catch" "comptime" "const" "continue" "defer" "else" "enum"
+ "errdefer" "error" "export" "extern" "fn" "for" "if" "inline" "linksection"
+ "noalias" "noinline" "nosuspend" "opaque" "or" "orelse" "packed" "pub"
+ "resume" "return" "struct" "suspend" "switch" "test" "threadlocal" "try"
+ "union" "usingnamespace" "var" "volatile" "while"] @keyword
+
+(builtin_type) @type
+["anyframe" "anytype"] @type
+(parameter type: (identifier) @type)
+(variable_declaration (identifier) @type "="
+  [(struct_declaration) (enum_declaration) (union_declaration)
+   (opaque_declaration)])
+
+["null" "undefined" "unreachable"] @constant
+(boolean) @constant
+; an enum literal — .ok — is a member with nothing to its left
+(field_expression . member: (identifier) @constant)
+(enum_declaration (container_field type: (identifier) @constant))
+
+(container_field name: (identifier) @property)
+(field_initializer . (identifier) @property)
+(field_expression (_) member: (identifier) @property)
+; .x = 1 inside a Point{...} is a field, not the enum literal above
+(initializer_list
+  (assignment_expression
+    left: (field_expression . member: (identifier) @property)))
+(parameter name: (identifier) @variable.parameter)
+(block_label (identifier) @label)
+(break_label (identifier) @label)
+
+(function_declaration name: (identifier) @function)
+(call_expression function: (identifier) @function)
+(call_expression function: (field_expression member: (identifier) @function.method))
+(builtin_identifier) @function.builtin
+
+; Zig has no type-identifier node, only the TitleCase convention, and these
+; land after the call patterns so that the convention wins the way upstream's
+; first-pattern-wins order has it: std.ArrayList(u8) is a type, not a call.
+((identifier) @type
+ (#match? @type "^[A-Z][a-zA-Z0-9_]*$"))
+((identifier) @constant
+ (#match? @constant "^[A-Z][A-Z0-9_]+$"))
+
+[(integer) (float)] @number
+[(string) (multiline_string) (character)] @string
+(escape_sequence) @string.escape
+
+(comment) @comment
+((comment) @comment.doc
+ (#match? @comment.doc "^//[/!]"))