fix(feed): only tooltip the channel name where it's actually cut off

I set the attribute unconditionally because the one-line row truncates
hard, and never looked at the other four views — where the name fits and
the tooltip just repeated what was already on screen. It now shares the
title's clamped check.

frontend/src/components/VideoCard.tsx

@@ -164,6 +164,9 @@ function RowMetaCells({ video }: { video: Video }) {
           ? { Icon: VideoIcon, cls: "text-accent", label: t("card.stream") }
           : null;
   const StatusIcon = status?.Icon;
+  // 4rem + 1rem + 1rem + two 4px gaps = 104px, which the one-line row's width tally counts on —
+  // widen a slot and the title column silently pays for it. rem, not px, so the whole cell tracks
+  // the text-size setting the same way its contents do.
   return (
     <div className="shrink-0 grid grid-cols-[4rem_1rem_1rem] gap-1 items-center text-sm text-muted">
       {/* Right-aligned: a duration is a number, so the digits line up on their own edge (30px for
@@ -356,20 +359,27 @@ function VideoCard({
       {video.published_at && <>{" · "}{formatDate(video.published_at, i18n.language)}</>}
     </>
   );
-  // Show the full title in a native tooltip only when it's actually cut off — vertically where it
-  // clamps to 2 lines, horizontally where the one-line row truncates it.
+  // Show the full text in a native tooltip only when it's actually cut off — vertically where the
+  // title clamps to 2 lines, horizontally where a fixed column truncates it. Shared by the title
+  // and the channel name: both truncate in some views and fit in others, and a tooltip that just
+  // repeats what's already legible is noise.
   const titleRef = useRef<HTMLAnchorElement>(null);
+  const channelRef = useRef<HTMLButtonElement>(null);
   const [titleClamped, setTitleClamped] = useState(false);
+  const [channelClamped, setChannelClamped] = useState(false);
   useEffect(() => {
-    const el = titleRef.current;
-    if (!el) return;
-    const check = () =>
-      setTitleClamped(el.scrollHeight > el.clientHeight + 1 || el.scrollWidth > el.clientWidth + 1);
+    const cut = (el: HTMLElement) =>
+      el.scrollHeight > el.clientHeight + 1 || el.scrollWidth > el.clientWidth + 1;
+    const els: [HTMLElement | null, (v: boolean) => void][] = [
+      [titleRef.current, setTitleClamped],
+      [channelRef.current, setChannelClamped],
+    ];
+    const check = () => els.forEach(([el, set]) => el && set(cut(el)));
     check();
     const ro = new ResizeObserver(check);
-    ro.observe(el);
+    els.forEach(([el]) => el && ro.observe(el));
     return () => ro.disconnect();
-  }, [video.title]);
+  }, [video.title, video.channel_title]);
   const title = (
     <a
       ref={titleRef}
@@ -388,13 +398,12 @@ function VideoCard({
   );
   const channelButton = (
     <button
+      ref={channelRef}
       onClick={(e) => {
         e.stopPropagation();
         onOpenChannel?.(video.channel_id, video.channel_title ?? t("card.thisChannel"));
       }}
-      // The name truncates in every layout (hard in the one-line row's fixed column), so it always
-      // carries the full text — unlike the title, which only earns a tooltip once it's clamped.
-      title={video.channel_title ?? undefined}
+      title={channelClamped ? video.channel_title ?? undefined : undefined}
       className="text-sm text-muted truncate block w-fit max-w-full text-left hover:text-fg"
     >
       {video.channel_title}
@@ -418,8 +427,9 @@ function VideoCard({
 
   if (spec.family === "row") {
     // A table without being one. Every column is a FIXED width, which is what makes the same datum
-    // land on the same vertical line in every row — a CSS grid can't do it for us, because
-    // VirtualFeed renders each row as its own subtree, so there is no shared grid to align to.
+    // land on the same vertical line in every row. A grid can't align ACROSS rows here — VirtualFeed
+    // renders each as its own subtree, so there's no shared parent to be the grid — but a grid is
+    // still right WITHIN a cell (see RowMetaCells): fixed slots, so they line up across rows too.
     // Widths are measured, not chosen by eye — against the worst case in the LIBRARY and in BOTH
     // languages, since both have caught me out here:
     //   actions 188 (6 x 28 + 5 x 4)  ·  meta cells 104 (see RowMetaCells)  ·  views 139 ("101.7K