feat(lora): log exact side-channel applies

The exact factor path was the only apply route with no log line; its
success read as silence. Track layers taking it beside the hosted and
fallback lists and report all three as key=value apply lines
(apply=exact/hosted/requantize); the stack fallback notices use the
same form. The suite pins its stack-mode baseline to sum so a mode
left set in user config cannot reroute tests that assume plain
summation.

modules/lora/lora_stack.py

@@ -78,10 +78,10 @@ def active_select(n_loaded):
     if m not in SELECT_MODES:
         return False
     if n_loaded != 2:
-        warn_once('select-count', f'Network stack: mode={m} networks={n_loaded} requires exactly 2, using sum')
+        warn_once('select-count', f'Network stack: mode={m} networks={n_loaded} required=2 fallback=sum')
         return False
     if select_blocked():
-        warn_once('select-compile', f'Network stack: mode={m} disabled with model compile, using sum')
+        warn_once('select-compile', f'Network stack: mode={m} compile=model fallback=sum')
         return False
     return True
 
@@ -290,7 +290,7 @@ def weight_selection(module, entry, winner):
     from modules.lora.lora_apply import network_apply_weights
     backup = getattr(module, 'network_weights_backup', None)
     if not isinstance(backup, torch.Tensor): # fuse mode keeps a bool sentinel, not a pristine copy
-        warn_once('select-nobackup', 'Network stack: select flip skipped, no weight backup')
+        warn_once('select-nobackup', 'Network stack: flip=skipped backup=none')
         return
     net = next((n for n in l.loaded_networks if n.name == entry['nets'][winner]), None)
     net_module = net.modules.get(entry['layer'], None) if net is not None else None