Petr Pisar
unread,Nov 8, 2017, 11:58:35 AM11/8/17You do not have permission to delete messages in this group
Either email addresses are anonymous for this group or you need the view member email addresses permission to view the original message
to bug-gn...@gnu.org
The recently released time-1.8 reports a command failure if -q option is
not specified. This happens even if -p option is specified:
$ ./time-1.8/time -p /usr/bin/false
Command exited with non-zero status 1
real 0.00
user 0.00
sys 0.00
I think the first line should not be here when POSIX mode is request
with the -p option. POSIX says:
If −p is specified, the following format shall be used in the POSIX
locale:
"real %f\nuser %f\nsys %f\n", <real seconds>, <user seconds>,
<system seconds>
[...] The implementation may
append white space and additional information following the format shown
here. The implementation may also prepend a single empty line before the
format shown here.
In the previous 1.7 release, Fedora patched time to report the command
failure only if -v option was specified, so that "time -p" also did not
report it.
But because 1.8 started to control the command failure with -q option,
I feel users will complain that "time -p" from time-1.8 does not behave
as prescribed by the POSIX.
I'd like to apply this change the corrects it:
--- a/src/time.c
+++ b/src/time.c
@@ -431,7 +431,7 @@ summarize (fp, fmt, command, resp)
unsigned long ru; /* Elapsed real microseconds. */
unsigned long vu; /* Elapsed virtual (CPU) microseconds. */
- if (!quiet)
+ if (!quiet && output_format != posix_format)
{
if (WIFSTOPPED (resp->waitstatus))
fprintf (fp, "Command stopped by signal %d\n",
-- Petr