> Mhhh... wasn't there a way to do a longjmp from a signal handler > (|siglongjmp()|) ? If AST |malloc()| is (still) async-signal safe > (that's a question for kpv) then a builtin could _optionally_ use this > way to enable all signals and do a |siglongjmp()| to a safe point > within itself for resource cleanup... > ... would that work ?
Some space may be lost forever from getting reused because the respective data structures will remain locked after a longjmp. Otherwise, safe. Phong _______________________________________________ ast-developers mailing list ast-developers@research.att.com https://mailman.research.att.com/mailman/listinfo/ast-developers
