Jonathan Nieder wrote:
> Yep, especially in the light of [1], this sounds like a bug.
>
> When running a special builtin, exverror will longjmp past the
> handling of eflag back to the cmdloop, and the latter does not have
> enough information to tell whether it is time to terminate the shell.
Side note:
POSIX is not as clear as one would like about this. XCU ยง2.8.1
"Consequences of Shell Errors" says:
In all of the cases shown in the table, an interactive shell
shall write a diagnostic message to standard error without
exiting.
This does not sit well with my intuition that in interactive shells,
"dot script not found" should be an error like any other (and in
particular cause the shell to exit if the -e flag is set). A
clarification to the standard would be nice.
> [1] http://austingroupbugs.net/view.php?id=114
--
To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]
