[freenet-dev] [freenet-cvs] r19567 - trunk/freenet/src/freenet/store
On Wednesday 30 April 2008 19:34, Matthew Toseland wrote: > On Saturday 26 April 2008 04:57, j16sdiz at freenetproject.org wrote: > > Author: j16sdiz > > Date: 2008-04-26 03:57:34 + (Sat, 26 Apr 2008) > > New Revision: 19567 > > > > Modified: > >trunk/freenet/src/freenet/store/BerkeleyDBFreenetStore.java > > Log: > > use FileChannel.read()/write() in BDBFS > > > All good stuff, however: > - Don't use allocateDirect() unless you plan to keep the returned ByteBuffer > around for a long time. They are immobile and therefore hard to garbage > collect. Sorry... I should really queue my replies until I reach the end of the commit log. :| > - ClosedChannelException, AsynchronousCloseException and > ClosedByInterruptException should be caught and cause a silent failure for a > smooth shutdown. > This is however still valid. -- next part -- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available URL: <https://emu.freenetproject.org/pipermail/devl/attachments/20080430/6da2f56d/attachment.pgp>
[freenet-dev] [freenet-cvs] r19596 - trunk/freenet/src/freenet/node
On Monday 28 April 2008 17:59, nextgens at freenetproject.org wrote:
> Author: nextgens
> Date: 2008-04-28 16:59:25 + (Mon, 28 Apr 2008)
> New Revision: 19596
>
> Modified:
>trunk/freenet/src/freenet/node/KeyTracker.java
> Log:
> ditto
>
> Modified: trunk/freenet/src/freenet/node/KeyTracker.java
> ===
> --- trunk/freenet/src/freenet/node/KeyTracker.java2008-04-28 16:57:21 UTC
(rev 19595)
> +++ trunk/freenet/src/freenet/node/KeyTracker.java2008-04-28 16:59:25 UTC
(rev 19596)
> @@ -728,7 +728,7 @@
> return acks;
> }
>
> - public void requeueForgot(int[] forgotPackets, int start, int length) {
> + public synchronized void requeueForgot(int[] forgotPackets, int start,
> int
length) {
Why?
> synchronized(forgottenQueue) { // It doesn't do anything else
> does it?
REDFLAG
> for(int i=start;i<start+length;i++) {
> queueForgotten(i, false);
-- next part --
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL:
<https://emu.freenetproject.org/pipermail/devl/attachments/20080430/bd8caa76/attachment.pgp>
[freenet-dev] [freenet-cvs] r19589 - trunk/freenet/src/freenet/node
On Monday 28 April 2008 17:17, nextgens at freenetproject.org wrote:
> Author: nextgens
> Date: 2008-04-28 16:17:10 + (Mon, 28 Apr 2008)
> New Revision: 19589
>
> Modified:
>trunk/freenet/src/freenet/node/PeerManager.java
> Log:
> Fix a deadlock in PeerManager: writePeersInner will request PeerNode.this
*but* the calling thread might not have it
We don't really want to hold the PeerManager lock while doing I/O now do we?
How did this deadlock work?
>
> Modified: trunk/freenet/src/freenet/node/PeerManager.java
> ===
> --- trunk/freenet/src/freenet/node/PeerManager.java 2008-04-28 15:45:40 UTC
(rev 19588)
> +++ trunk/freenet/src/freenet/node/PeerManager.java 2008-04-28 16:17:10 UTC
(rev 19589)
> @@ -1008,6 +1008,7 @@
> }
>
> private void writePeersInner() {
> + synchronized(this) {
> synchronized(writePeersSync) {
> if(darkFilename != null)
> writePeersInner(darkFilename, getDarknetPeers());
> @@ -1018,6 +1019,7 @@
> writePeersInner(om.getOldPeersFilename(),
> om.getOldPeers());
> }
> }
> + }
> }
>
> /**
>
> ___
> cvs mailing list
> cvs at freenetproject.org
> http://emu.freenetproject.org/cgi-bin/mailman/listinfo/cvs
>
>
-- next part --
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL:
<https://emu.freenetproject.org/pipermail/devl/attachments/20080430/07871e1d/attachment.pgp>
[freenet-dev] [freenet-cvs] r19567 - trunk/freenet/src/freenet/store
On Saturday 26 April 2008 04:57, j16sdiz at freenetproject.org wrote: > Author: j16sdiz > Date: 2008-04-26 03:57:34 + (Sat, 26 Apr 2008) > New Revision: 19567 > > Modified: >trunk/freenet/src/freenet/store/BerkeleyDBFreenetStore.java > Log: > use FileChannel.read()/write() in BDBFS > All good stuff, however: - Don't use allocateDirect() unless you plan to keep the returned ByteBuffer around for a long time. They are immobile and therefore hard to garbage collect. - ClosedChannelException, AsynchronousCloseException and ClosedByInterruptException should be caught and cause a silent failure for a smooth shutdown. -- next part -- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available URL: <https://emu.freenetproject.org/pipermail/devl/attachments/20080430/0c10bbb3/attachment.pgp>
[freenet-dev] Mac using volunteers needed!
Could anyone who has access to a MacOS/X box please contact us? We need to test the new startup script for the installer, so that when we release 0.7.0, it will work out of the box even with a reboot (like it does on windows). -- next part -- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available URL: <https://emu.freenetproject.org/pipermail/devl/attachments/20080430/de217a19/attachment.pgp>
[freenet-dev] Config page usability
On Tuesday 29 April 2008 23:33:59 Matthew Toseland wrote: > IMHO there is a difference between a geek who knows what an IP address is, > and an experienced freenet user. Duh, nobody doubts that. But most of the people using freenet will probably think "why should I only look at the medium settings when I can also look at the advanced settings" and act accordingly. It makes the people feel more clever. They like that. That's why I think that the medium settings page is pretty much a waste of time. David -- next part -- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 197 bytes Desc: This is a digitally signed message part. URL: <https://emu.freenetproject.org/pipermail/devl/attachments/20080430/440ed741/attachment.pgp>
[freenet-dev] Config page usability
On Tuesday 29 April 2008 23:37, David ?Bombe? Roden wrote: > On Tuesday 29 April 2008 23:33:59 Matthew Toseland wrote: > > > IMHO there is a difference between a geek who knows what an IP address is, > > and an experienced freenet user. > > Duh, nobody doubts that. But most of the people using freenet will probably > think "why should I only look at the medium settings when I can also look at > the advanced settings" and act accordingly. It makes the people feel more > clever. They like that. > > That's why I think that the medium settings page is pretty much a waste of > time. Well, the main purpose for the medium settings page is to have the small number of settings that are useful to the average geek easily visible and not hidden amongst the large number of advanced options. To try to minimise the number of people who come to us asking how to open fproxy to their LAN. -- next part -- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available URL: <https://emu.freenetproject.org/pipermail/devl/attachments/20080430/3aa65744/attachment.pgp>
[freenet-dev] Mac using volunteers needed!
Could anyone who has access to a MacOS/X box please contact us? We need to test the new startup script for the installer, so that when we release 0.7.0, it will work out of the box even with a reboot (like it does on windows). pgpS71IYXUp1t.pgp Description: PGP signature ___ Devl mailing list Devl@freenetproject.org http://emu.freenetproject.org/cgi-bin/mailman/listinfo/devl
Re: [freenet-dev] [freenet-cvs] r19567 - trunk/freenet/src/freenet/store
On Saturday 26 April 2008 04:57, [EMAIL PROTECTED] wrote: Author: j16sdiz Date: 2008-04-26 03:57:34 + (Sat, 26 Apr 2008) New Revision: 19567 Modified: trunk/freenet/src/freenet/store/BerkeleyDBFreenetStore.java Log: use FileChannel.read()/write() in BDBFS All good stuff, however: - Don't use allocateDirect() unless you plan to keep the returned ByteBuffer around for a long time. They are immobile and therefore hard to garbage collect. - ClosedChannelException, AsynchronousCloseException and ClosedByInterruptException should be caught and cause a silent failure for a smooth shutdown. pgpTyDDd41dYP.pgp Description: PGP signature ___ Devl mailing list Devl@freenetproject.org http://emu.freenetproject.org/cgi-bin/mailman/listinfo/devl
Re: [freenet-dev] [freenet-cvs] r19589 - trunk/freenet/src/freenet/node
On Monday 28 April 2008 17:17, [EMAIL PROTECTED] wrote:
Author: nextgens
Date: 2008-04-28 16:17:10 + (Mon, 28 Apr 2008)
New Revision: 19589
Modified:
trunk/freenet/src/freenet/node/PeerManager.java
Log:
Fix a deadlock in PeerManager: writePeersInner will request PeerNode.this
*but* the calling thread might not have it
We don't really want to hold the PeerManager lock while doing I/O now do we?
How did this deadlock work?
Modified: trunk/freenet/src/freenet/node/PeerManager.java
===
--- trunk/freenet/src/freenet/node/PeerManager.java 2008-04-28 15:45:40 UTC
(rev 19588)
+++ trunk/freenet/src/freenet/node/PeerManager.java 2008-04-28 16:17:10 UTC
(rev 19589)
@@ -1008,6 +1008,7 @@
}
private void writePeersInner() {
+ synchronized(this) {
synchronized(writePeersSync) {
if(darkFilename != null)
writePeersInner(darkFilename, getDarknetPeers());
@@ -1018,6 +1019,7 @@
writePeersInner(om.getOldPeersFilename(),
om.getOldPeers());
}
}
+ }
}
/**
___
cvs mailing list
[EMAIL PROTECTED]
http://emu.freenetproject.org/cgi-bin/mailman/listinfo/cvs
pgpKd2Dr4jsWu.pgp
Description: PGP signature
___
Devl mailing list
Devl@freenetproject.org
http://emu.freenetproject.org/cgi-bin/mailman/listinfo/devl
Re: [freenet-dev] [freenet-cvs] r19596 - trunk/freenet/src/freenet/node
On Monday 28 April 2008 17:59, [EMAIL PROTECTED] wrote:
Author: nextgens
Date: 2008-04-28 16:59:25 + (Mon, 28 Apr 2008)
New Revision: 19596
Modified:
trunk/freenet/src/freenet/node/KeyTracker.java
Log:
ditto
Modified: trunk/freenet/src/freenet/node/KeyTracker.java
===
--- trunk/freenet/src/freenet/node/KeyTracker.java2008-04-28 16:57:21 UTC
(rev 19595)
+++ trunk/freenet/src/freenet/node/KeyTracker.java2008-04-28 16:59:25 UTC
(rev 19596)
@@ -728,7 +728,7 @@
return acks;
}
- public void requeueForgot(int[] forgotPackets, int start, int length) {
+ public synchronized void requeueForgot(int[] forgotPackets, int start,
int
length) {
Why?
synchronized(forgottenQueue) { // It doesn't do anything else
does it?
REDFLAG
for(int i=start;istart+length;i++) {
queueForgotten(i, false);
pgpARvKjacmtm.pgp
Description: PGP signature
___
Devl mailing list
Devl@freenetproject.org
http://emu.freenetproject.org/cgi-bin/mailman/listinfo/devl
Re: [freenet-dev] [freenet-cvs] r19567 - trunk/freenet/src/freenet/store
On Wednesday 30 April 2008 19:34, Matthew Toseland wrote: On Saturday 26 April 2008 04:57, [EMAIL PROTECTED] wrote: Author: j16sdiz Date: 2008-04-26 03:57:34 + (Sat, 26 Apr 2008) New Revision: 19567 Modified: trunk/freenet/src/freenet/store/BerkeleyDBFreenetStore.java Log: use FileChannel.read()/write() in BDBFS All good stuff, however: - Don't use allocateDirect() unless you plan to keep the returned ByteBuffer around for a long time. They are immobile and therefore hard to garbage collect. Sorry... I should really queue my replies until I reach the end of the commit log. :| - ClosedChannelException, AsynchronousCloseException and ClosedByInterruptException should be caught and cause a silent failure for a smooth shutdown. This is however still valid. pgpq2sGMwQfFj.pgp Description: PGP signature ___ Devl mailing list Devl@freenetproject.org http://emu.freenetproject.org/cgi-bin/mailman/listinfo/devl
Re: [freenet-dev] [freenet-cvs] r19567 - trunk/freenet/src/freenet/store
On Thu, May 1, 2008 at 4:50 AM, Matthew Toseland [EMAIL PROTECTED] wrote: On Wednesday 30 April 2008 19:34, Matthew Toseland wrote: On Saturday 26 April 2008 04:57, [EMAIL PROTECTED] wrote: Author: j16sdiz Date: 2008-04-26 03:57:34 + (Sat, 26 Apr 2008) New Revision: 19567 Modified: trunk/freenet/src/freenet/store/BerkeleyDBFreenetStore.java Log: use FileChannel.read()/write() in BDBFS All good stuff, however: - Don't use allocateDirect() unless you plan to keep the returned ByteBuffer around for a long time. They are immobile and therefore hard to garbage collect. Sorry... I should really queue my replies until I reach the end of the commit log. :| - ClosedChannelException, AsynchronousCloseException and ClosedByInterruptException should be caught and cause a silent failure for a smooth shutdown. This is however still valid. fixed in r19641 ClosedChannelException is the super class of other two. ___ Devl mailing list Devl@freenetproject.org http://emu.freenetproject.org/cgi-bin/mailman/listinfo/devl
