(0004216)
eblake (manager)
2019-01-21 16:51
edited on: 2019-01-21 16:59
|
Interpretation response
------------------------
The standard clearly states that dd shall terminate abnormally as if by SIGINT, and conforming implementations must conform to this.
Rationale:
-------------
Being able to use a 'trap' handler to cater to an abnormal termination of dd in a long-running script is a useful behavior. Additionally, the standard should allow for existing practice of some implementations providing a way to get ongoing status reports from dd without terminating the process.
Notes to the Editor (not part of this interpretation):
-------------------------------------------------------
At line 86094 (ASYNCHRONOUS EVENTS), change:
For SIGINT, the dd utility shall interrupt its current processing, write status information to standard error, and exit as though terminated by SIGINT. It shall take the standard action for all other signals; see the ASYNCHRONOUS EVENTS section in Section 1.4 (on page 2336). to:For SIGINT, the dd utility shall interrupt its current processing, write status information to standard error, and terminate abnormally as if by the default action for SIGINT. One or more implementation defined non-job-control signals other than SIGABRT, SIGHUP, and SIGTERM may write status information to standard error and continue processing. All other signals (including job control signals, SIGABRT, SIGHUP, and SIGTERM) shall take their default action; see the ASYNCHRONOUS EVENTS section in Section 1.4 (on page 2336).
At line 86205 FUTURE DIRECTIONS, change:
None. to:A future version of this standard may introduce the SIGINFO signal; on platforms where such a signal is available, it is recommended that this signal be used for reporting status without terminating the process.
|