On Wed, Dec 27, 2000 at 05:06:48PM -0600, Mark J. Roberts wrote:
> On Wed, 27 Dec 2000, Mark J. Roberts wrote:
>
> > > > To add support for MSKs to RequestClient, the obvious place to add it is
> > > > ClientUtil.getKey, but I would have to overload it with another version
> > > > that takes a MapHandler as an argument, and that's ugly and inefficient,
> > > > because the key might not even be a MSK. So I could add a isMSK method to
> > > > ClientUtil that takes the key string and checks. I'd do it in fillBuckets
> > > > before the ClientKey is initialized, and if it returns true, I'd
> > > > instantiate a MapHandler and run the key string through it first.
> > >
> > > ClientUtil.getKey is the right place, I think, but the place for the
> > > MapHandler is in MapHandler.instance (static variable). You only need one
> > > map handler for the whole JVM even if you have multiple things running and
> > > you don't need to mess with all that other crazy stuff.
> >
> > So that's how you do that! Java's so kewl...
>
> Well, that's all working great for me now. The only problem is that FProxy
> leaks memory so terribly that it would be impossible to surf Freenet for
> more than 10 minutes without swapping and thrashing.
>
> PID USER PRI NI SIZE RSS SHARE STAT LIB %CPU %MEM TIME COMMAND
> 444 root 9 0 87460 73M 380 S 0 0.0 59.9 0:00 java
>
> (I have 128 megs of RAM!)
>
> This is fucking unbelievable. I want to kill somebody. At first I thought
> it was my fault, and after fooling around with my code for an hour, I
> checked out a fresh, unaltered tree -- and it was EXACTLY THE SAME!
> (Actually, Mr. Bad's SSK subspace, with about 10-15 images, leaked about
> the same as or a little more than my hashtable in a mapspace with 35
> images (full shift-key reload for both). This is strong evidence in favor
> of MSKs.)
>
> Is this the fault of SimplifiedClient and ProxyClient? Brandon, do you
> have any idea what's causing this? Should someone who knows a lot about
> memory and Java rewrite SimplifiedClient or something? Or is this a bigger
> problem?
This kind of stuff really makes me wonder about the state of
programming these days. It seems as if programmers have forgotten
that computers have limited amounts of RAM and swap space and that
processors are not infinitely fast. Just yesterday, one of my friends
of "real life" showed me the game Command & Conquer Red Alert II by
Westwood. It required at least 128 MB of RAM and a 330 MHz x86
processor! If you aren't doing something like encoding music in
realtime into a lossy format like MP3, Ogg Vorbis, or MP4, a program
shouldn't require that amount of speed. If you aren't doing something
like raytracing huge scenes, you shouldn't require that amount of
space. The idea that a simple realtime strategy game should require
that kind of power is just crazy. And this doesn't appear to be a
game which uses 3D rendering in realtime without hardware
acceleration! Am I one of the few programmers these days who gives a
shit about memory usage and efficiency?!!
--
Travis Bemann
Sendmail is still screwed up on my box.
My email address is really [EMAIL PROTECTED]
PGP signature
