On Tue, Feb 17, 2015 at 04:29:29PM -0500, Daniel Kahn Gillmor wrote: > it looks like if you send emacs a SIGINT, it will clean up nicely and > terminate with a return code of 0.
That seems like suboptimal behavior. Exiting due to a signal doesn't seem like a "successful" exit. It seems emacs isn't very consistent about this. Running emacsclient to connect to an existing daemon means that frame exits with an exit code of 130, as expected. Running a standalone “emacs -nw”, SIGINT gets caught and doesn't exit. Running a standalone X emacs frame does exit cleanly as you describe. > if i run a bts command that drops me into my editor in an X11 > environment, emacs spawns as a separate window, while my shell is > blocked by the bts invocation (as it should be). > > if i hit ctrl-C in the shell, though, apparently bts passes the SIGINT > through to my editor, and then proceeds to use the contents of the > message as prepared. Yes, Perl ignores SIGINT/SIGQUIT while running the external process, so that process gets the signal. I'll adjust bts to inspect the exit code of the process so it doesn't treat a non-zero exit code as success. However, emacs will need to be fixed to not exit with a zero when killed. As an alternative, you can tell bts to cancel by saving an empty file. Cheers, -- James GPG Key: 4096R/331BA3DB 2011-12-05 James McCoy <[email protected]>
signature.asc
Description: Digital signature
_______________________________________________ devscripts-devel mailing list [email protected] http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/devscripts-devel
