A tag write past the limit says tag: no space: over 4096 bytes, which 9ns maps to ENOSPC as the docs say

The tree answers ENOSPC, but through a mount 9ns maps the words, and tag: over 4096 bytes carried none it knows, so a shell saw EIO. The words now hold no space; docs/fs.md says ENOSPC again.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

1 files changed, 4 insertions, 2 deletions

diff --git a/src/ninep/pane.zig b/src/ninep/pane.zig
index bb4ebf86..3c1e181e 100644
--- a/src/ninep/pane.zig
+++ b/src/ninep/pane.zig
@@ -571,7 +571,7 @@ fn writeTag(req: Req, pane: *Pane) Reply {
     const had = tagline.curTail(pane).len;
     const room = limits.max_tag_tail -| had -| @intFromBool(pf.tag_held_newline);
     // Whole or not at all: a write that would pass the limit changes nothing.
-    if (req.data.len > room) return tree.failText(req.tag, E.NOSPC, std.fmt.comptimePrint("tag: over {d} bytes", .{limits.max_tag_tail}));
+    if (req.data.len > room) return tree.failText(req.tag, E.NOSPC, std.fmt.comptimePrint("tag: no space: over {d} bytes", .{limits.max_tag_tail}));
     const take = wholeUtf8(req.data);
     // A truncating write drops ONE trailing newline, its whole text's: a
     // newline held from a write before goes in once more text follows it.
@@ -1214,7 +1214,9 @@ test "a tag write past the limit is refused whole, naming the limit" {
     const big = "w" ** (limits.max_tag_tail + 10);
     const r = wr(p, tag, big);
     try testing.expectEqual(E.NOSPC, r.errno());
-    try testing.expect(std.mem.startsWith(u8, r.reply.ename, "tag: over "));
+    // `no space` is what 9ns maps to ENOSPC, so a shell through a mount
+    // sees the errno the tree gives.
+    try testing.expect(std.mem.startsWith(u8, r.reply.ename, "tag: no space: over "));
     try testing.expectEqualStrings(before, rd(p, tag, 0, 1 << 16).bytes);
 }