engine: an okurigana mark does not outlive its kana

Shift on a romaji letter marks where the okurigana begins, and Backspace
shortened the reading through im.l->back without clearing the mark.
dictqjp only applies a mark that still falls inside the reading, so the
stale index lay dormant while the reading was short and fired again the
moment it grew back past it: kaKu, Backspace, then nzi offered
噛んじ 兼んじ 漢字 幹事 感じ, where kanzi typed plainly offers 漢字 first.
A Backspace edits the reading, so the split marked in it is gone.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

strans.c

@@ -970,6 +970,7 @@ transition(u32int ks, u32int mod, Str *com)
 		else if(ks == Kesc)
 			reset();
 		else{
+			okuriat = -1;	/* the mark goes with the kana it marked */
 			im.l->back(&im);
 			if(haspre(&im))
 				dictqjp();