Add #roomname references in chat messages (#47)

Mirrors the existing @-mention system's shape: a regex finds #roomname
tokens, extract_referenced_room_ids validates them against rooms the
*sender* actually belongs to (mirrors mentions' "must be a real member"
rule -- referencing a private room the sender isn't in would otherwise
leak its existence), and a MessageRoomReference join row is stored per
match in create_message. No notification/unread layer, unlike mentions --
referencing a room has no "you were referenced" semantics.

Rendering is the same markdown-link rewrite trick MessageContent.tsx
already uses for mentions (#username -> [#username](mention:username)),
but resolved against the *viewer's* own room list (threaded down from
ChatShellPage's room state through ChatPane/MessageList) rather than the
stored server-side reference -- a reference to a room the current viewer
isn't in quietly renders as plain text instead of a link, same as an
@mention of someone outside the room does. The href scheme renders a
real react-router Link instead of mentions' inert span, since a room
reference is meant to be navigable.

mention_service.strip_code_spans (was _strip_code_spans) is now shared
between both extraction paths rather than private to one module.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

backend/app/services/mention_service.py

@@ -9,12 +9,14 @@ from app.models import RoomMembership, User
 MENTION_PATTERN = re.compile(r"@([a-zA-Z0-9_.-]+)")
 
 
-def _strip_code_spans(content: str) -> str:
+def strip_code_spans(content: str) -> str:
     """Blanks out fenced code blocks and inline code spans (replacing with
-    equal-length whitespace, so a bare '@' in pasted code -- a decorator, an
-    email fragment -- doesn't page someone. Mirrors the same skip logic
+    equal-length whitespace, so a bare '@'/'#' in pasted code -- a
+    decorator, an email fragment, a shell comment -- doesn't trigger a
+    mention or room reference. Mirrors the same skip logic
     frontend/src/components/MessageContent.tsx already uses for emoji
-    shortcode conversion."""
+    shortcode conversion. Shared with room_reference_service, not private
+    to this module anymore."""
     lines = content.split("\n")
     in_fence = False
     out = []
@@ -37,7 +39,7 @@ async def extract_mentioned_user_ids(
     """Resolves `@username` tokens in `content` against this room's actual
     members -- a bare '@' followed by prose that happens to not match
     anyone's username is just text, not a mention."""
-    usernames = set(MENTION_PATTERN.findall(_strip_code_spans(content)))
+    usernames = set(MENTION_PATTERN.findall(strip_code_spans(content)))
     if not usernames:
         return set()