Phase 7h: string slicing s[a..b] (retire substr)

`s[a..b]` is a fresh substring of the bytes [a, b) — the modern, end-based form
of the C-style `substr(s, start, count)`:

  substr(src, start, i - start)   -> src[start..i]
  substr(t, 2, len(t) - 2)        -> t[2..len(t)]
  substr(src, i, 2)               -> src[i..i + 2]

Mechanics: a new E_SLICE postfix (`base[lo..hi]`, distinct from `base[i]`
indexing) lowers to a @fn_str_slice prelude (malloc + copy + terminate), emitted
once into any program that slices. Two reseeds: add the syntax + prelude, then
migrate the 22 substr calls and delete substr. The migrator recognises the
common `count == end - start` shape and emits the clean `s[start..end]` rather
than `s[start..start + (end - start)]`.

examples/strings.ludic gains slicing (now prints 1..9). Reseeded (22673 lines);
C-free fixpoint holds; goldens identical; 18/18; vocab + doc-fences clean.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

selfhost/emit_decl.ludic

@@ -69,6 +69,7 @@ fn emit_program() -> void {
   code = buf_new()
   g_uses_str = false
   g_uses_intstr = false
+  g_uses_strslice = false
   loc_name = new []ptr;  loc_reg = new []ptr;  loc_ty = new []ptr;  loc_mut = new []int
   brk_lbl = new []ptr;   cnt_lbl = new []ptr
   self_stk = new []ptr
@@ -86,6 +87,7 @@ fn emit_program() -> void {
   }
   if g_uses_str { emit_str_prelude() }          # @fn_str_eq / @fn_str_concat, after all uses are seen
   if g_uses_intstr { emit_int_str() }           # @fn_int_str, for str(int) in interpolation
+  if g_uses_strslice { emit_str_slice() }       # @fn_str_slice, for s[a..b]
 }
 
 # Flush the emitted IR. With a null path it goes to stdout (the pipe the shell