Am 11.08.2011 um 15:35 schrieb Dave Love:
> Reuti <re...@staff.uni-marburg.de> writes:
>
>> Yeah, exactly. But although it's tricky to use two buffers trace_str and
>> trace_str2 for alternating assignments, I would suggest to use only one for
>> the final string.
>>
>> strcat(trace_str, "INFINITY");
>>
>>
>>> }
>>> else {
>>> strcat(trace_str, limit_fmt) ;
>>> sprintf(trace_str2, trace_str, dlp.rlim_cur ) ;
>>
>> sprintf(&trace_str[strlen(trace_str)], limit_fmt, dlp.rlim_cur);
>
> [Yes, that approach is used in admin_mail.c (?), but I think was
> violating C rules as originally coded.]
>
>> Maybe Dave can make a statement what he prefers putting into the repository.
>
> Thanks for the suggestions. I'll have to look at it later, though I
> wonder if it's simplest just to strip the backspace sequences post hoc.
>
> I'm actually more interested in under what circumstances you get the
> shepherd trace in the admin mail.
Problem in the starter_method or prolog/epilog of the queue can cause it.
-- Reuti
> At least with 6.2u5 and, I think,
> 8.0.0, I've only ever seen empty output for the shepherd data
> <https://arc.liv.ac.uk/trac/SGE/ticket/1307>, in contrast to what 6.0u8
> did in the same circumstances.
_______________________________________________
users mailing list
users@gridengine.org
https://gridengine.org/mailman/listinfo/users
