term+libc: batched write rendering - one tile render per write(), 2.3x lines

sys_write sets wBatch; render_cursor_tile - the single choke point all
of term_putc's visible mutations pass through - then marks a dirty
span (per LINE-SLOT, so spans survive scroll rotation for free)
instead of rendering. term_write_end renders every dirty tile exactly
once, then the cursor. Scrolls stay immediate (blank rebuild is cheap
and artifact-free); file-bound writes flush for free.

The enabler is in libc: puts() was a SYS_PUTC trap PER CHARACTER, so
almost no console output ever went through sys_write. It now issues
one SYS_WRITE for the whole string - one trap, one batch, and every
line-printer (echo, irc, the shell) gets the batched path.

Measured: 37-char write = 128k cycles (3.5k/char) vs ~8k/char down
the per-char trap path. Full demo passes; GIF regenerated.

1 files changed, 18 insertions, 8 deletions

diff --git a/usr/libc.s b/usr/libc.s
index 2366cd8..fbca836 100644
--- a/usr/libc.s
+++ b/usr/libc.s
@@ -33,18 +33,28 @@ _putc:
 	rst	#0x30
 	ret
 
-	;; void puts(const char *s /*DE*/) - write until NUL (no newline)
+	;; void puts(const char *s /*DE*/) - write until NUL (no newline).
+	;; ONE SYS_WRITE for the whole string instead of a SYS_PUTC per char:
+	;; a single trap, and the kernel defers tile rendering for the whole
+	;; run (sys_write batching), so a line paints each tile once.
 _puts:
-1$:	ld	a, (de)
+	push	de
+	ld	h, d
+	ld	l, e
+	ld	b, #0
+1$:	ld	a, (hl)
+	or	a
+	jr	z, 2$
+	inc	hl
+	inc	b
+	jr	1$
+2$:	pop	de
+	ld	a, b
 	or	a
 	ret	z
-	push	de
-	ld	e, a
-	ld	c, #16			; SYS_PUTC
+	ld	c, #3			; SYS_WRITE (DE=buf, B=len)
 	rst	#0x30
-	pop	de
-	inc	de
-	jr	1$
+	ret
 
 	;; void nl(void) - CRLF via stdout
 _nl: