>> makes more visual sense.  it might be fun to extend the %[npu] fields
>> with things like d (for a number, always, as in %u), o or x for the
>> pid (making it %xp), etc.  maybe even add %t for current time_t.  then
>> i could have a coredump string of
>>
>>  /home/%u/.cores/%n.%08xt.%05p
>> 
>> so that none of my coredumps get smashed.
>
>a couple of things here:
>
>* i'm not sure i'd bother with the precision/padding specs.  just pick
>some sane default, and use it.  this doesn't need to be infinitely
>flexible, just serve its purpose, and that seems overkill (that could
>be added later if people really needed it).

that's certainly syntactic sugar...

>* it's probably useful to add a specifier for the process start time
>(p->p_stats->p_start) as well.

...but everyone seems to have comments on what would be nice given the
facility to control core dump naming.

with a little work, i'm sure we could make the format string as
convoluted as the posix printf semantics (with the $ stuff and
everything).

-- 
|-----< "CODE WARRIOR" >-----|
codewarrior@daemon.org             * "ah!  i see you have the internet
twofsonet@graffiti.com (Andrew Brown)                that goes *ping*!"
andrew@crossbar.com       * "information is power -- share the wealth."