Re: How to use Procmail to remove messages from server after x number of

On Tue, Mar 01, 2011 at 04:35:59PM -0500, Richard Ball wrote:
> On 1-3-2011 16:31, LuKreme wrote:
> > -exec rm {} \; is simply a long way of spawning a shell and executing rm for 
> > old, OLD versions of find that don't have -delete.
> Which will break if too many files are found. (where "too many" means 
> the length of the string of filenames exceeds the buffer length 
> allocated for commands on the system).
I think you're thinking of "rm `find ... `".  According to my experience 
and the documentation of the POSIX behavior of "find" 
(http://pubs.opengroup.org/onlinepubs/009695399/utilities/find.html),
"find ... -exec rm {} \;" should run rm once for each pathname, and thus 
should never exceed the buffer length.
I'm glad I looked that up, because now I know about "find ... -exec rm {}+". 
It avoids the heavyweight one-process-per-file of "-exec rm {} \;", 
batching them in sets, but limits the number of pathnames in each set 
specifically to avoid exceeding the buffer limit.
And since there's no intervening pipeline (unlike "find ... -print | 
xargs rm"), pathnames with spaces and newlines don't cause problems.
--
Ed Blackman

signature.txt
Description: Digital signature

____________________________________________________________
procmail mailing list   Procmail homepage: http://www.procmail.org/
procmail(_at_)lists(_dot_)RWTH-Aachen(_dot_)de
http://mailman.rwth-aachen.de/mailman/listinfo/procmail