OK

> the error message is coming from mem.c.
> 
> Bill will have to look at this; I have no idea where to start.
> 
> > I use $results->NextResult to get data
> > and save in perl hash only 10-100 results needed for display 
> > result page.
> > I think if error message 
> > 	err: Ran out of memory (could not allocate 262144 more bytes)!
> > created in SWISH::API, then trouble in search or in process 
> search results
> > from API function. 
> > I use 5 indexes and result use merge although I don't known 
> help you it.
> > 
> > 
> >>>>>May be use search from perl require much more memory than 
> >>>>
> >>>>use from command line?
> >>>>
> >>>>depends on what you're doing. Are you using swish.cgi? or a 
> >>>>custom app with the 
> >>>>SWISH::API?
> >>>
> >>>SWISH::API call from perl.
> >>
> >>Bill will answer this more intelligently than me. Until then, 
> >>you get my dumb 
> >>answers.
> >>
> >>Perl API shouldn't use appreciably more memory than swish-e 
> >>binary since Perl is 
> >>using the same C library that swish-e is. I would guess that 
> >>whatever you *do* 
> >>with the results (save to a Perl hash, etc) is where the 
> >>memory difference would 
> >>come in, since Perl is notoriously greedy with memory.
> >>
> >>I can only suggest the typical Perl debugging strategies:
> >>
> >>use warnings
> >>use strict
> >>use Devel::Peek (for memory debugging)
> >>try and limit your tests to a small set of docs or small set 
> >>of results and use 
> >>lots of print() or warn() statements to help tell you where you are.
> >>
> >>I've never had a problem with running out of memory using the 
> >>API; I think your 
> >>guess about user limits is probably right; check with your 
> >>sysadmin (unless you 
> >>are the sysadmin, in which case check yourself and then 
> check google).
> >>
> >>
> >>-- 
> >>Peter Karman  .  http://peknet.com/  .  peter(at)not-real.peknet.com
> >>
> 
> -- 
> Peter Karman  .  http://peknet.com/  .  peter(at)not-real.peknet.com
>