httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Dean Gaudet <>
Subject Re: Apache 2.0/NSPR
Date Mon, 14 Sep 1998 02:07:24 GMT
How about someone describe the actual difficulties that lead to this whole
gc discussion before we take it any further?

As far as I can recall, the only storage leaks that have ever occured in
Apache were due to people using make_sub_pool(NULL).  Something which I
made impossible in apache-nspr (or at least I think I did, I intended

Or was it me talking about reference counting buffer pages?  Gah... 
reference counts are trivial for that usage... why look further? 


On Sun, 13 Sep 1998, Rodent of Unusual Size wrote:

> Jim Jagielski wrote:
> > 
> > > Personally I find even the idea of adding pfree to the pool
> > > abstraction distasteful.
> > 
> > I do too, but I don't feel qualified say exactly why :)
> Hmm.  Irrational, unfounded belief structure, Captain.  Fascinating.
> :-)
> Well, if we're going to do pools in shared memory (I know at least
> one other person studying it), that means bounds.  And bounded
> allocation of shared structures implies garbage collection.  Or
> terminal constipation, take your pick. :-)
> Personally, I like the Java approach: garbage collection is done
> by the 'system' and the application has no direct access to it,
> other than by nuking all handles to allocated blocks so the
> GC knows they're candidates when it needs to reclaim.  Applying
> that to our current pool system means changing the semantics
> fairly drastically.  But bounded, shared pools are a pretty
> drastic change, too, so maybe we can do it right -- and make it
> httpd-neutral at the same time.  An excellent 'port' candidate,
> eh?
> Or am I talking through my hat?
> #ken	P-)}
> Ken Coar                    <http://Web.Golux.Com/coar/>
> Apache Group member         <>
> "Apache Server for Dummies" <http://Web.Golux.Com/coar/ASFD/>

View raw message