Re: routine_id() is not working
- Posted by mattlewis (admin) Apr 11, 2013
- 1452 views
since the sleep in os.e is already namespaced away from misc.e,
That's not quite how euphoria namespaces work.
Yes, that "already namespaced away" struck me as a very odd thing to say. My tuppenceworth:
If the compiler finds a conflict,
But the thing is, the way Matt explained it, there is a conflict, and no error was thrown. The sleep in os could be referred to as os:sleep, but it was still in the same namespace as misc's sleep.
No error was thrown when the two files both included a sleep, and none was thrown when routine_id() was used on sleep.
Pete's explanation was a little odd, too, in that he referred to the compiler. A routine_id call is entirely handled at runtime. So it's really the interpreter dealing with this at runtime. The error is returned as -1 (albeit without much help at figuring out why).
So i also suggested, and maybe to limit the scope of the post i shouldn't have, mentioned it might be more helpful for the crash-on-start to list the routine_id(sleep) lines when it found two sleeps in the same namespace.
We could add warnings, perhaps, for routine_id calls with literals that we know cannot be resolved. Since you can dynamically build the name of the routine, we obviously cannot do this for all calls to routine_id.
Matt
