Hey, d00ds:
Freenet development has moved to the very fine [EMAIL PROTECTED]
list, here:
http://www.uprizer.com/mailman/listinfo/devl
Things are going quite spiffily.
~Mr. Bad
--
~
/\/\ Mr. Bad <[EM
OS> issues).
I believe there would be PLENTY of I18N issues that fall under a
development purview, but I agree that we should table those until 0.4
is out.
~Mr. Bad
--
~~~~~~~~~~~~~
/\/\ Mr. Bad <[EMAIL PROTECTED]>
\ /
mised.
I agree with Oskar that what I described wouldn't solve this
problem. The only thing I can think of that would is having the
routing algorithms try to preserve diversity in the routing table --
which would probably be hard to do right.
~Mr. Bad
--
~
this, I can say fairly certainly that Fred does verify *all*
SGM> the data passing through it.
As usual, I stand corrected!
~Mr. Bad
--
~~~~~~~~~
/\/\ Mr. Bad <[EMAIL PROTECTED]>
\ / Pigdog Journal | http://pigdog.
>>>>> "OS" == Oskar Sandberg <[EMAIL PROTECTED]> writes:
OS> We have no solution to the problem of the honest cancer (not
OS> even Mr. Bad's "lets break freenet" proposal really
OS> h
KSKs, I don't think
they're the only problem: As far as I can tell, Fred does -not- check
CHKs or do other verification on data it gets.
MT> The public network will always be vulnerable to such attacks,
MT> it i
A - B - C - D
OK, every node here is shy (except maybe B, who we don't really care
about. Screw B!). Unfortunately, A and C have been a little
promiscuous with their trust and provided an entry into Freenet for a
cancer node, B. The B stands for "bad" (and not of the Mr. Ba
ial rather
than a technical solution, but it might be the only one that would
even help.
~Mr. Bad
--
~~~~~~~~~
/\/\ Mr. Bad <[EMAIL PROTECTED]>
\ / Pigdog Journal | http://pigdog.org/ | *Stay*Real*Bad*
| (X \x)
n somebody routes the
bug reports to the developer most likely to understand/fix the problem
and...
~Mr. Bad
--
~~~~~~~~~
/\/\ Mr. Bad <[EMAIL PROTECTED]>
\ / Pigdog Journal | http://pigdog.org/ | *Stay*Real*Bad*
| (X
e to see some of the newer stuff, like MSKs and date-based
redirects go into 0.4, and really try to wind 0.3 down to just bug
fixing and egregious problems like the ref blocking you fixed.
If we have TWO branches adding new features, we are going to have a
hell of a time reconciling the
rches for "car", they
AV> won't find it.
Yeah, but also if people search on "automobilo" and there's a keyword
"automobiloj", they're probably not going to hit, either.
~Mr. Bad
--
~~~
tenPort" in
your freenet.ini. Then, restart Freenet and try it again.
If that *still* doesn't work... try going to a DOS window, cd to your
freenet directory, and try:
frequest gpl.txt
That should -definitely- work. If not, post back!
~Mr. Bad
--
~
's been great fun
so far, and I really think it's going strong.
Beaujolais to you! You're heroes of Information Freedom.
~Mr. Bad
--
~~~~~~~~~
/\/\ Mr. Bad <[EMAIL PROTECTED]>
\ / Pigdog Journal | http://pi
s worms, and tech companies are amoral
money-grubbers.
I am *really* interested in seeing which of us is correct come New
Year's Eve, 2001.
~Mr. Bad
--
~
/\/\ Mr. Bad <[EMAIL PROTECTED]>
\ / Pigdog Journ
ing method):
Spock spock = new Spock();
synchronized(spock) {
spock.gar();
}
I don't know if it even matters, really. 'Swhy I asked in the first
place. B-)
~Mr. Bad
--
~
/\/\
o you can
SGM> control access to the object rather than the method.
No, I understand that. What I was talking about was exactly what I
wrote -- that the synchonized block uses the monitor belonging to
"this",
) {
// stuff happens here;
}
? The first is an idiom I've run across a few times in Fred code, and
I've never seen it used before.
~Mr. Bad
--
~
/\____/\ Mr. Bad <[EMAIL PROTECTED]>
\ / Pigdo
imenting with
BC> this. (and MSKs, too.)
Well, I'd love to give it a shot, since I don't get this redirecting
hooha at all. But if you want to, why don't you muck with the
freenetmirror code in the freenet CVS area? It's at...
cvs -d
>>>>> "OS" == Oskar Sandberg <[EMAIL PROTECTED]> writes:
OS> You couldn't offend me if you tried. :-)
You might be the only one.
OK, I think you guys are right on this one. B-) And that's
s might be another nice place to use Hprof. Anyone with a Windows
machine ever tried it?
~Mr. Bad
P.S. freenet-dev was my most active mailing list over the last 48
hours, except for sf-raves. We must be onto something if we talk about
it this much. I've inserted a Content-Type=beer/draft into Fr
Memory."
I think it stands for "I Pretend To Support Software Freedom With
Linux Ports and Fancy-Pants Alphaworks Stuff But Then I Go Do
Something Stupid Like Try To Get Copy Protection Added To Every Hard
Drive In The World."
Oh, wait... I think that might be too long.
~Mr.
My point still stands, though: I think "stable" should be for stable
stuff and main trunk should be for development.
I also think we should add all the new features for pseudoupdating and
filtering to the 0.4 feature list.
~Mr. Bad
--
~~~~~~~~~~
ets a ref collision as a bug
fix.
Ian, did you test this at all? Like actually run it and make sure it
worked?
~Mr. Bad
--
~~~~~~~~~
/\/\ Mr. Bad <[EMAIL PROTECTED]>
\ / Pigdog Journal | http://pigdog.org/ | *Stay*Re
"
nodes to make it work. Nobody's suggested yet that the only way to
keep from going to jail or getting killed is if we encourage lots of
other people to buy and sell drugs out in the open. We've found other
ways to keep from going to jail or getting killed.
Strange, that.
~Mr.
not harmful
Oskar, it was a joke. You don't have to respond.
OS> (ie, not your clusters)...
They're not "my" clusters. B-)
~Mr. Bad
--
~~~~~~~~~
/\/\ Mr. Bad <[EMAIL PROTECTED]>
\ / Pigd
>>>>> "SGM" == Scott Gregory Miller <[EMAIL PROTECTED]> writes:
SGM> No, you're blowing that out of proportion. The idea is give
SGM> people technologies that are less difficult to fuck up.
...or no technol
NALITY of Fred was broken in the official release for over a
month. The whole point of "stable" is so we can fix stuff like that
really quickly.
~Mr. Bad
--
~
/\/\ Mr. Bad <[EMAIL PROTECTED]>
\ / Pigdog
heads.
This has got to be the most back-assward reasoning I've ever
heard. "Most people will misuse this feature, so don't give it to
anybody."
~Mr. Bad
--
~~~~~~~~~
/\/\ Mr. Bad <[EMAIL PROTECTED]>
\ /
>>>>> "BC" == Benjamin Coates <[EMAIL PROTECTED]> writes:
BC> Mr. Bad: Can you support date-based redirects in
BC> freenetmirror? That would make for a farily straightforward
BC> package to create a freenet-website...
Not yet, no. Rememb
e.
Wow! That doesn't seem really appropriate at all... It's a change in
behavior, not a bug fix, and it's pretty much untested.
It seems like it makes more sense to hold this kind of thing for the
next release rather than putting it in stable. BTW, I volunteered to
ma
;s also a nice profile output analyzer called HAT
available from Sun, just for heap analysis. Here the URL:
http://java.sun.com/people/billf/heap/
I did some investigation of this stuff and I didn't see anything that
was glaringly out of whack, but I'm not really a performance tuner.
G
ree. However, the person I was responding to (gosh, lost
that message already) had said that Freenet now was a "zero-trust"
system. I think it's more like a 100%-trust system.
~Mr. Bad
--
~~~~~~~~~
/\/\ Mr. Bad <[EMAIL
g. BBSes in Freenet is going to -require- client-side
code, and doing it in the browser is the best way right now.
~Mr. Bad
--
~~~~~~~~~
/\/\ Mr. Bad <[EMAIL PROTECTED]>
\ / Pigdog Journal | http://pigdog.org/ | *Stay*Re
'd prefer to
choose who I let do that -- expecially when I know that there are
people out there who will abuse that privilege in order to shut down
my Net connection.
~Mr. Bad
--
~~~~~~~~~~~~~
/\/\ Mr. Bad <[EMAIL PROTECTED]
e node list is not
B> seized. Therefore creating clusters reduces liability because
B> less nodes are detectable via port scanning or running an evil
B> node.
OK, this sounds right now. B-)
~Mr. Bad
--
~~~~~~~~~
/\/
>>>>> "IC" == Ian Clarke <[EMAIL PROTECTED]> writes:
IC> Comments? Ideas? Insults? Thoughts?
Do the insults have to pertain to this particular proposal?
~Mr. Bad
--
~~~~~~~~~
/\
me point."
That was the basis of my point, which was more humorous than anything
else. I tend to be of the school that likes to throw in small features
if they are a) up to the user/admin to choose and b) relatively
low-impact for other functionality.
But that's just me.
~Mr. Bad
--
deployment. We don't have this, and I
don't think we're going to have it before a technology like
copyright.net's starts attacking nodes.
~Mr. Bad
--
~
/\/\ Mr. Bad <[EMAIL PROTECTED]>
\ /
>>>>> "IC" == Ian Clarke <[EMAIL PROTECTED]> writes:
IC> So the Unix and Windows files are now up, and I have sent an
IC> announcement request to Freshmeat.
Ian: I did the branch point already for "stable." This was still t
bly in higher proportion than if it wasn't a gateway,
but it's still not totally obvious that the clustered nodes even
exist.
~Mr. Bad
--
~
/\/\ Mr. Bad <[EMAIL PROTECTED]>
\ / Pigdog Journal | http
ses the "Media
IC> Enforcer" problem?
I think it does, partially.
~Mr. Bad
--
~
/\/\ Mr. Bad <[EMAIL PROTECTED]>
\ / Pigdog Journal | http://pigdog.org/ | *Stay*Real*Bad*
| (X \x)
(
here you have no relationship with
SGM> references on your node.
This is a very good point. Having evidence that you've made some
out-of-band arrangements with other node operators would be somewhat
incriminating -- probably some kind of collusion.
However, does
ld do a good job of
replacing "clustering" without actually doing any clustering. B-)
Although it -does- kind of draw more attention to a shield node than a
clustering system would (since no one would know that a gateway was
actually a gateway, but a shield node's IP address goes out with
er (and by "liability" I think you mean
"risk"). It seems right now that the best-case scenario is more
probable than the worst-case one.
Also, I don't quite believe that in a country where machines get
seized, there's going to be
Public
B> key crypto defeats the MediaEnforcer attack (scanning IPs
B> doesn't work).
I think the word you're looking for is "seize," not "cease."
~Mr. Bad
--
~
/\/\ Mr. Bad <[E
l
OS> refutation of this in the arguments given, and I'm afraid that
OS> the burden of proof falls on Brandon and Mr. Bad here - there
OS> are several bases for simulators out there (Serapis, that
OS> which Theo wrote for his analysis, and little birds whisper
OS&g
>>>>> "TM" == Timm Murray <[EMAIL PROTECTED]> writes:
TM> If gatewayed private networks ever get implemented and Freenet
TM> becomes very widespread, then this discussion is just the
TM> first in a very long list of
>>>>> "BC" == Benjamin Coates <[EMAIL PROTECTED]> writes:
BC> p.s. Wow, you have the google 'I'm feeling lucky' for "decss".
Yeah, because practically every other DeCSS site has been shut down.
~Mr. Bad
--
~~~~~
raints of what I have said above, I am more
IC> than happy to provide a detailed examination of this issue -
IC> but any solution should be classified as an extension of
IC> Freenet's current aims, not a "security fix".
How about just, "This is a risk you r
don and I are saying is that we won't have ANY
network if we don't make a priority of protecting node operators or at
least giving them some ability to protect themselves -- even at the
possible expense of speed or scalability.
I'd like to think there's some compromise space i
BC> happened already?)
OK, I just made a "stable" branch at the rel-0-3-6 point. The mainline
trunk (regular CVS, no changes to your working dir or whatever) will
be "development."
I need to talk with whoever does the builds and snapshots, so that we
can s
r "Dont let anyone find
SGM> out we're running Freenet" which is an arguable problem.
I think that's the "cluster" idea. "Don't talk to strangers" means:
"Even if someone knows I'm running Freenet, don't let them use my node
to
se, is a different story B-).)
"They" are perfectly happy to throw tons of money out there to fuck
us. Experience seems to show that "they" can take us out if we do this
stuff above ground. It's sad, but I can't help feeling that it's true.
Count
plan to do about the risk, even if we
choose not to do anything and bank on the integrity and courage of
ISPs and universities. But even if we haven't decided what we're going
to do, we should point out the potential hazard.
~Mr. Bad
--
~~~~~~~~~~
-in-the-sky ideas based on a) widely-available free ISPs and b)
unmatchable account-to-IP-address for cable modems. For people like
me who have DSL accounts (there's a 3+ month wait for DSL in Northern
California) neither one is workable.
~Mr. Bad
--
~~
>>>>> "SGM" == Scott G Miller <[EMAIL PROTECTED]> writes:
SGM> The mediaenforcer attack is resisted by PK (as well as it can
SGM> be)
I don't understand how this is true.
~Mr. Bad
--
~~~~~~~
o remove users for running Napster, citing
IC> free-speech concerns, if these concerns applied to Napster,
IC> then they would *definitely* apply to Freenet.
Many, many universities shut down Napster even before getting RIAA
letters. Those gutsy schools are the exception, not the rule.
e
behind Fred was *not* a knock on those projects. It was supposed to
point out that a) implementing the Freenet protocol is HARD and b)
there's only one implementation that actually works right now, and
it's GPL'd.
I think all the others are GPL'd, too, so that may be a moot pt
it's good
for dissidents, they're going to get shot, and they're going to die.
Optimizing for scalability is so unimportant compared to this, Ian.
~Mr. Bad
--
~
/\/\ Mr. Bad <[EMAIL PROTECTED]>
\ /
;d have to put it under GPL -- since Fred is GPL and not LGPL.
I guess I just wanted to loft that balloon out there. If there's a
nice FSF lawyer or someone who could send a letter to these guys, it
might be a good stalling tactic. Hell, I dunno.
~Mr. Bad
P.S. Oh, crap -- looks like they ru
ly to be a
E> problem in large networks, but it seems to me to be a problem.
Yeah, it's a problem all right.
~Mr. Bad
P.S. Glad to see we're talking in the same language now!
--
~~~~~~~~~
/\/\ Mr. Bad <[EMA
u to legal
liability. You are responsible (and could be held liable to
others) for all submissions from your Service account."
I just don't think the evidence holds up that ISPs wouldn't bother to
act on cease-and-desist letters.
~Mr. Bad
--
~~
quest it from downstream nodes, then put me in
jail for providing him with child porn.
It also may help against some other attacks.
Both of these make for slower networks. However, they decrease the
propagation of addresses, and make nodes (not publishers or readers)
more anonymous.
~Mr.
caches (?).
MJR> Of a total of 305 references, 287 are to the same node.
Our simulations show that this doesn't happen. B-)
~Mr. Bad
--
~~~~~~~~~
/\/\ Mr. Bad <[EMAIL PROTECTED]>
\ / Pigdog Jo
.
I don't think it would be a popularity contest at all. Even if I was
very trustworthy, I wouldn't set up an agreement with YOU if you were
untrustworthy.
~Mr. Bad
--
~~~~~
/\/\ Mr. Bad <[EMAIL PRO
ystem for
distributing files in a hostile environment. It seems like paranoia is
built in from the ground up -- this optional feature would be at most
a few more lines of code, and would give some vital protection.
I dunno, I guess I think it's a good idea.
~Mr. Bad
--
~~~
hat is borked right now please speak out.
IC> I will release on Saturday evening if there are no objections.
Ian,
It's Sunday afternoon on this side of the pond, and I can't find 0.3.6
on the Freenet site. Were there ob
sidents are really just guerrillas without guns.
~Mr. Bad
--
~~~~~~~~~
/\/\ Mr. Bad <[EMAIL PROTECTED]>
\ / Pigdog Journal | http://pigdog.org/ | *Stay*Real*Bad*
| (X \x)
(((**) "If it's not bad, do
only trust other cluster computers, and then gateways that trust their
cluster and a handful of other gateways.
~Mr. Bad
--
~~~~~~~~~
/\/\ Mr. Bad <[EMAIL PROTECTED]>
\ / Pigdog Journal | http://pigdog.org/ |
So, there's just a ton of dead code in the CVS repository -- stuff
like Freenet/client/old/*, lots of build.sh.old things and etc.
Are there good reasons not to cvs remove these? They'll still be
available for retrieval if necessary, but they sure do make build
scripts complicated.
70 matches
Mail list logo
