Package: lsb-base Version: 3.2-16 Severity: grave This change in lsb-base 3.2-16
* Fix behavior of killproc and pidofproc when no pidfile is passed in. breaks the stop action in many init scripts, as evidenced by bugs such as #494268 and #495437 coming in now. The consensus appears to be that using the --name option of start-stop-daemon is evil. I also don't understand what this "fix" is trying to accomplish. (No bug number is listed.) The LSB spec states: """ The start_daemon, killproc and pidofproc functions shall use the following algorithm for determining the status and the process identifiers of the specified program. 1. If the -p pidfile option is specified, and the named pidfile exists, a single line at the start of the pidfile shall be read. If this line contains one or more numeric values, separated by spaces, these values shall be used. If the -p pidfile option is specified and the named pidfile does not exist, the functions shall assume that the daemon is not running. 2. Otherwise, /var/run/basename.pid shall be read in a similar fashion. If this contains one or more numeric values on the first line, these values shall be used. Optionally, implementations may use unspecified additional methods to locate the process identifiers required. """ What the old version did is #2, and at least in case of dirmngr (bug #495437) this is exactly the behavior the init script wanted. The current version does something that is not in agreement with the LSB spec and breaks init scripts quite trivially. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
