grammar: a #tag starts after whitespace and with a letter, on every surface

The shared fixture went red on the server (run 8468: 5 failed), because the three
tag rules disagreed:
- core and web: any non-tag character counts as a boundary, so `(#todo)`
  and `end.#tag` are tags, and so is the `/#section` of a pasted URL;
- server: only whitespace counts, but `#1st` and `#_x` are tags.

A note's labels could therefore change every time it synced.

All three now share the strict rule: start of line or whitespace, then a
letter, then letters, digits, `_` and `-`. Nothing becomes a tag that wasn't
already one everywhere, and URL anchors stop becoming labels on desktop and
Android. The server's existing `http://x/#nope` test already expected this.

derive.rs's boundary, markdown.ts's lookbehind and tags.py's regex change
together; `_is_tag` goes because the regex now requires the letter. The
fixture flips `(#todo)`, `end.#tag` and its lift case, and adds the URL case.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

core/src/local/derive.rs

@@ -2,8 +2,10 @@
 //! computes on save. Pure string scanning (no regex dependency), kept in lockstep
 //! with the frontend's inline rules (see frontend notes/markdown.ts):
 //!
-//! - `#tag`: `#` at a word boundary followed by tag characters (letter first).
-//!   On save these become labels attached with `via_tag = true`.
+//! - `#tag`: `#` at the start of a line or after whitespace, then a letter, then
+//!   letters, digits, `_` and `-`. On save these become labels attached with
+//!   `via_tag = true`. The server and the web run the same cases
+//!   (core/testdata/grammar.json).
 //! - `- [ ] item`: a checklist item. The body IS the checklist (M304) — there is no
 //!   table of items beside it, so a list can sit between two paragraphs instead of
 //!   only after them.
@@ -29,7 +31,11 @@ fn line_tags(chars: &[char]) -> Vec<(usize, usize, String)> {
     let mut i = 0;
     while i < chars.len() {
         if chars[i] == '#' {
-            let boundary = i == 0 || (!is_tag_char(chars[i - 1]) && chars[i - 1] != '#');
+            // Start of line or after whitespace, and nothing else. Any non-tag
+            // character used to count, which made `(#todo)` a tag here and plain
+            // text on the server, and made the `/#section` of a pasted URL a label.
+            // Whitespace is the rule all three implementations now share (#5166).
+            let boundary = i == 0 || chars[i - 1].is_whitespace();
             // A tag must start with a letter (so "#1" or a bare "#" is not a tag).
             if boundary && i + 1 < chars.len() && chars[i + 1].is_alphabetic() {
                 let mut j = i + 1;

core/testdata/grammar.json

@@ -43,8 +43,9 @@
     { "body": "#tag, and #more.", "tags": ["tag", "more"] },
     { "body": "#x-y_z9", "tags": ["x-y_z9"] },
     { "body": "#café #über", "tags": ["café", "über"] },
-    { "body": "(#todo)", "tags": ["todo"] },
-    { "body": "end.#tag", "tags": ["tag"] },
+    { "body": "(#todo)", "tags": [] },
+    { "body": "end.#tag", "tags": [] },
+    { "body": "https://x.com/#section", "tags": [] },
     { "body": "#tag#more", "tags": ["tag"] },
     { "body": "a#b", "tags": [] },
     { "body": "x-#tag", "tags": [] },
@@ -64,7 +65,7 @@
     { "body": "#todo\ncall #todo later", "standalone": [], "inline": ["todo"], "lifted": "call #todo later" },
     { "body": "#a #b", "standalone": [], "inline": ["a", "b"], "lifted": "#a #b" },
     { "body": "```\n# comment #todo\n```", "standalone": [], "inline": ["todo"], "lifted": "```\n# comment #todo\n```" },
-    { "body": "(#todo)\nnotes", "standalone": [], "inline": ["todo"], "lifted": "(#todo)\nnotes" }
+    { "body": "(#todo)\nnotes", "standalone": [], "inline": [], "lifted": "(#todo)\nnotes" }
   ],
 
   "tint": {