In local.freenet, you wrote:
> The superdir of your Freenet directory of course. Depending on what you
> are building, you may also need to make the Contrib/javax and
> Contrib/junit classes available.
Thanks, I was lacking the javax-stuff (which is kind of counter intuitive
as I thought that jav
In local.freenet, you wrote:
> A related question: how can this be accomplished while allowing Freenets
> to exist in private blocks with the ability to connect to the larger
> network as well as directly with each other?
That's hard to domaybe you could add a command line switch which tell
Zitiere David McNab <[EMAIL PROTECTED]>:
> thanks ken for the patch
second that
> regrettably it failed.
>
> can you please ensure you've got latest version of the fcptools files
> from
> cvs, and re-generate the patch.
Please generete the unified diffs with
diff -ruN
thats
cvs diff -uN
for C
You can hack the freenet.ini config file, and manually change that value to
something greater.
In the meantime, I or Sebastian will need to add code to the config program
to allow sizes > 2047 if the disk is NTFS/
David
- Original Message -
From: "Lee Gagner"
To:
Sent: Monday, October
> In the meantime, I or Sebastian will need to add code to the config program
> to allow sizes > 2047 if the disk is NTFS/
I thought the 2GB limit was due to the Java VM, as well as the OS.
(i.e. If the OS allows it, the VM might not)
-- Emil
___
The only real concern is that it is connecting anywhere 26 times in less
than a minute. That some refs will be broken is unavoidable - it doesn't
matter whether they are IPs on the wrong network or simply names that
don't resolve, they _should_ be flushed out soon enough.
On Mon, Oct 22, 2001 at
thanks ken for the patch
regrettably it failed.
can you please ensure you've got latest version of the fcptools files from
cvs, and re-generate the patch.
seems the patch only applies to makefile.am
otherwise, you could just tar up the modified files and send them to me
cheers
david
__
> I had a packet catcher running while I was checkingo out a suspicion I
> had and I stumbled across some private addresses. Sure enough, my node
> was trying to connect to 192.168.0.1 -- 26 times in under a minute. What
> is being done to stop private addresses from polluting the routing tables
A non-text attachment was scrubbed...
Name: msg.pgp
Type: application/pgp
Size: 7105 bytes
Desc: not available
URL:
<https://emu.freenetproject.org/pipermail/devl/attachments/20011022/c73a1d38/attachment.bin>
I downloaded the Freenet sources from the CVS and noticed that the ant
build script didn't function anymore. I made a new one based on the
Makefile and now I could update it to CVS but I don't have write access
there.
I suppose we can have both the Ant and Make build scripts since they are
quite
msg.pgp
On Mon, Oct 22, 2001 at 03:21:19PM +0200, Volker Stolz wrote:
> Hi, I was browsing through the networking code and noticed an odd thing:
> TCP sockets support a listen-backlog, i.e. the number of maximum pending
> copnnections. This is initialised to 50 and never changed.
>
> Wouldn't it be intere
Hi, I was browsing through the networking code and noticed an odd thing:
TCP sockets support a listen-backlog, i.e. the number of maximum pending
copnnections. This is initialised to 50 and never changed.
Wouldn't it be interesting to see if it would be sensible to modify the
backlog depending on
> > In the meantime, I or Sebastian will need to add code to the config program
> > to allow sizes > 2047 if the disk is NTFS/
>
> I thought the 2GB limit was due to the Java VM, as well as the OS.
> (i.e. If the OS allows it, the VM might not)
I think the Java standards specify that file sizes h
> > This isn't a Freenet problem really, it's a problem with the Windows file
> system. Freenet 0.4 stores all data in a single file, but Windows can't
> make > a file larger then 2 GB (2048 MB). Note that this is the limit of a
> 32-bit signed integer. If Microsoft were to increase this to 64-b
On Mon, Oct 22, 2001 at 08:59:44PM +1000, Emil Mikulic wrote:
> > In the meantime, I or Sebastian will need to add code to the config program
> > to allow sizes > 2047 if the disk is NTFS/
>
> I thought the 2GB limit was due to the Java VM, as well as the OS.
> (i.e. If the OS allows it, the VM mi
I had a packet catcher running while I was checkingo out a suspicion I
had and I stumbled across some private addresses. Sure enough, my node
was trying to connect to 192.168.0.1 -- 26 times in under a minute. What
is being done to stop private addresses from polluting the routing tables?
A rel
> I had a packet catcher running while I was checkingo out a suspicion I
> had and I stumbled across some private addresses. Sure enough, my node
> was trying to connect to 192.168.0.1 -- 26 times in under a minute. What
> is being done to stop private addresses from polluting the routing table
The only real concern is that it is connecting anywhere 26 times in less
than a minute. That some refs will be broken is unavoidable - it doesn't
matter whether they are IPs on the wrong network or simply names that
don't resolve, they _should_ be flushed out soon enough.
On Mon, Oct 22, 2001 a
I had a packet catcher running while I was checkingo out a suspicion I
had and I stumbled across some private addresses. Sure enough, my node
was trying to connect to 192.168.0.1 -- 26 times in under a minute. What
is being done to stop private addresses from polluting the routing tables?
A re
From: "Jan-Thomas Czornack"
> You can make larger nodes on Windows NT/2000 with NTFS. I tried it with
5Gb
> and it worked fine.
It's almost a case for converting my server over to win2k.
Emphasis on 'almost'.
Whereas linux requires just about every bleeding thing to be recompiled with
Clibs fo
> > > In the meantime, I or Sebastian will need to add code to the config program
> > > to allow sizes > 2047 if the disk is NTFS/
> >
> > I thought the 2GB limit was due to the Java VM, as well as the OS.
> > (i.e. If the OS allows it, the VM might not)
>
> I think the Java standards specify th
> > In the meantime, I or Sebastian will need to add code to the config program
> > to allow sizes > 2047 if the disk is NTFS/
>
> I thought the 2GB limit was due to the Java VM, as well as the OS.
> (i.e. If the OS allows it, the VM might not)
I think the Java standards specify that file sizes
> > This isn't a Freenet problem really, it's a problem with the Windows file
> system. Freenet 0.4 stores all data in a single file, but Windows can't
> make > a file larger then 2 GB (2048 MB). Note that this is the limit of a
> 32-bit signed integer. If Microsoft were to increase this to 64-
On Mon, Oct 22, 2001 at 03:21:19PM +0200, Volker Stolz wrote:
> Hi, I was browsing through the networking code and noticed an odd thing:
> TCP sockets support a listen-backlog, i.e. the number of maximum pending
> copnnections. This is initialised to 50 and never changed.
>
> Wouldn't it be inter
Hi, I was browsing through the networking code and noticed an odd thing:
TCP sockets support a listen-backlog, i.e. the number of maximum pending
copnnections. This is initialised to 50 and never changed.
Wouldn't it be interesting to see if it would be sensible to modify the
backlog depending on
On Mon, Oct 22, 2001 at 08:59:44PM +1000, Emil Mikulic wrote:
> > In the meantime, I or Sebastian will need to add code to the config program
> > to allow sizes > 2047 if the disk is NTFS/
>
> I thought the 2GB limit was due to the Java VM, as well as the OS.
> (i.e. If the OS allows it, the VM m
> In the meantime, I or Sebastian will need to add code to the config program
> to allow sizes > 2047 if the disk is NTFS/
I thought the 2GB limit was due to the Java VM, as well as the OS.
(i.e. If the OS allows it, the VM might not)
-- Emil
__
Hi
I've got a CVS on sourceforge and wanted to add a new file (with cvs add)
and commited the changes (cvs commit). The result was following message:
cvs: commit.c:2104: checkaddfile: Assertion `*rcsnode == ((void *)0)'
failed.
Terminated with fatal signal 6
Now I always get following message if
You can hack the freenet.ini config file, and manually change that value to
something greater.
In the meantime, I or Sebastian will need to add code to the config program
to allow sizes > 2047 if the disk is NTFS/
David
- Original Message -
From: "Lee Gagner" <[EMAIL PROTECTED]>
To: <[E
Actually, on second thought, this is almost certainly due to the thread
problems, because when the threadpool is maxed out the aggregating task
that is scheduled every minute cannot be run on time. If it concerns
you, disable the stats until the current problems are resolved.
On Mon, Oct 22, 2001
If it comes up every five minutes then it is hard to say whether it is
bad or not a problem. It either means that about one in five times
aggregating the last minutes stats takes a little more than one minute,
or it means that every minute takes 5 minutes and it is slipping longer
and longer behin
Well, if it isn't a windows problem, then it is probably a problem with
that particular JVM. Freenet uses signed 64 bit numbers for all data
sizes.
On Sun, Oct 21, 2001 at 10:10:45PM +0100, Dave Hooper wrote:
<>
> Are you sure? How can you explain the following API which has been in
> existenc
Someone wrote:
> I recently setup 0.4 (14102001 for Windows I think) and noticed that,
where
> you put in the maximum size of of the store, it whouln't take anything
above
> 2047M (0x7fff bytes). any possibility of scaling this by Ks or
clusters
> or something to allow larger (20-30Gb) stores?
This is the weekly Freenet Project job jar update.
Comments to devl at freenetproject.org or jobjar at hawk.freenetproject.org
Check out the FNBTS bug tracker at http://hawk.freenetproject.org/bts/bts.php
OPEN JOBS
-
* Produce more thorough and accurate newbie instructions for
gett
35 matches
Mail list logo
