You can subscribe to this list here.
2000 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
(9) |
---|---|---|---|---|---|---|---|---|---|---|---|---|
2001 |
Jan
(46) |
Feb
(37) |
Mar
(8) |
Apr
(76) |
May
(63) |
Jun
(9) |
Jul
(18) |
Aug
(7) |
Sep
(8) |
Oct
|
Nov
|
Dec
(7) |
2002 |
Jan
(1) |
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
From: David F. <dav...@ya...> - 2001-05-22 05:17:07
|
On Tuesday 22 May 2001 15:11, Brandon hit his keyboard with a hammer and this was the result: > > Inserted KSK@brandon-1 with a default htl. In tests with mjr, I have > > found that with inserts of default htl, and requests of htl 100 we could > > get a file between us in 2 minutes. > > At HTL 100 I got the file, so we've established that our nodes can talk to > each other. This is good since I never got any of your previously inserted > files. Now we can start testing actual mail. Have you sent me a mail? David -- David Findlay ---------- Email: dav...@ya... Homepage: http://users.bigpond.com/nedz/ Segmentation Fault. (Core dumped) |
From: Brandon <bl...@ut...> - 2001-05-22 05:11:47
|
> Inserted KSK@brandon-1 with a default htl. In tests with mjr, I have found > that with inserts of default htl, and requests of htl 100 we could get a file > between us in 2 minutes. At HTL 100 I got the file, so we've established that our nodes can talk to each other. This is good since I never got any of your previously inserted files. Now we can start testing actual mail. |
From: Brandon <bl...@ut...> - 2001-05-22 05:05:09
|
> I don't see much of a difference on Freenet between DNS and document naming. > I figure we better make it work, I'll hack Freenet to do it ;-) You don't have to hack Freenet to do it. It can be implemented entirely as a client-side solution that works on top of the existing Freenet. |
From: Brandon <bl...@ut...> - 2001-05-22 05:03:37
|
> Is anybody coding on it? I'd like to work on it in C with -lfreenet if not. > I realize we'd probably scrap the first few tries, but I have a sort of > masochistic tendency to want to write things more than once anyway (tortured > artist syndrome). The original coder has stated that he's given up on it, so I think it's fair game. Give me your sourceforge name and I'll give you CVS access. Don't worry about dead ends and rewriting stuff. This project is for fun, so do it the right way and don't compromise. libfreenet is nice, but I'd recommend you use the C XML-RPC interface instead. Unless you're coding under Windows since it doesn't compile there yet. I'd love to hear your ideas about how FNS should work. We have, of course, already talked about a whole lot of issues and may already have the answers to some of your concerns. |
From: <ia...@ha...> - 2001-05-22 02:03:19
|
On Mon, May 21, 2001 at 09:55:34PM -0400, Jay 'Eraserhead' Felice wrote: > On Mon, May 21, 2001 at 06:09:34PM -0700, Ian Clarke wrote: > > I don't think that DNS will work too well over the current freenet, > > however a special application, which uses Freenet's routing and caching, > > along with some of it's concepts (subspaces and date-based-redirects), > > but with a much more lightweight protocol, perhaps based around UDP, now > > that could work. > I don't see much of a difference on Freenet between DNS and document naming. > I figure we better make it work, I'll hack Freenet to do it ;-) DNS could work over Freenet, but the Freenet protocol is really *way* too bloated to make it work in practical terms. > Seriously, if it requires a UDP-based protocol for performance reasons, it > should be a UDP-based protocol which somehow caches the Freenet data. I do > _not_ want to go build another (even similar) Freenet. Believe me, it will be *much* easier to rebuild a cut-down UDP version than to try to hack Freenet, you will end up removing 95% of the code,and rewriting the other 5% anyway. It is actually quite straight-forward to reimplement the guts of the Freenet algorithm, and there are a number of simplifications you can make (you don't need to worry about streaming since the data will be so small etc). Ian. |
From: Jay 'E. F. <era...@er...> - 2001-05-22 01:56:40
|
On Mon, May 21, 2001 at 06:09:34PM -0700, Ian Clarke wrote: > I don't think that DNS will work too well over the current freenet, > however a special application, which uses Freenet's routing and caching, > along with some of it's concepts (subspaces and date-based-redirects), > but with a much more lightweight protocol, perhaps based around UDP, now > that could work. I don't see much of a difference on Freenet between DNS and document naming. I figure we better make it work, I'll hack Freenet to do it ;-) Seriously, if it requires a UDP-based protocol for performance reasons, it should be a UDP-based protocol which somehow caches the Freenet data. I do _not_ want to go build another (even similar) Freenet. > > You should speak to the guys at "OpenNic" though to see some of the > issues they have had to address with respect to domain registration. Agreed. Thanks. > > Ian. > |
From: David F. <dav...@ya...> - 2001-05-22 01:21:23
|
On Monday 21 May 2001 20:30, Brandon hit his keyboard with a hammer and this was the result: > Let's start with that incremental KSK testing again. Insert files for me > at KSK@brandon-1, KSK@brandon-2 with incremental HTLs and let me know when > you've done it. Then I'll download them, see how many hops you are from > me, and that way we can set our mail HTLs so that I'm sure that your mails > will get to me. Then we can start testing the mail software itself. > Hopefully by the end of the month we can have something suitable to > replace the ad-hoc freesite feedback mechanisms. Could you just give the fmail a go? I am in #eof David -- David Findlay ---------- Email: dav...@ya... Homepage: http://users.bigpond.com/nedz/ Segmentation Fault. (Core dumped) |
From: David F. <dav...@ya...> - 2001-05-22 01:15:38
|
On Monday 21 May 2001 20:30, Brandon hit his keyboard with a hammer and this was the result: > Let's start with that incremental KSK testing again. Insert files for me > at KSK@brandon-1, KSK@brandon-2 with incremental HTLs and let me know when > you've done it. Then I'll download them, see how many hops you are from > me, and that way we can set our mail HTLs so that I'm sure that your mails > will get to me. Then we can start testing the mail software itself. > Hopefully by the end of the month we can have something suitable to > replace the ad-hoc freesite feedback mechanisms. Inserted KSK@brandon-1 with a default htl. In tests with mjr, I have found that with inserts of default htl, and requests of htl 100 we could get a file between us in 2 minutes. David -- David Findlay ---------- Email: dav...@ya... Homepage: http://users.bigpond.com/nedz/ Segmentation Fault. (Core dumped) |
From: <ia...@ha...> - 2001-05-22 01:13:46
|
I don't think that DNS will work too well over the current freenet, however a special application, which uses Freenet's routing and caching, along with some of it's concepts (subspaces and date-based-redirects), but with a much more lightweight protocol, perhaps based around UDP, now that could work. You should speak to the guys at "OpenNic" though to see some of the issues they have had to address with respect to domain registration. Ian. On Mon, May 21, 2001 at 09:09:29PM -0400, Jay 'Eraserhead' Felice wrote: > Just curious, I bopped around to the Free DNS Project, then to here where= I > saw the FNS spec. >=20 > This is something I'd like to work on/theorize about/code. 'Specially the > latter. The spec that I saw was a great start, but I've already got a bu= nch > of additions for how it should and shouldn't work. >=20 > Is anybody coding on it? I'd like to work on it in C with -lfreenet if n= ot.=20 > I realize we'd probably scrap the first few tries, but I have a sort of > masochistic tendency to want to write things more than once anyway (tortu= red > artist syndrome). >=20 > I can't start coding for another three weeks, probably, but I'd like to p= ut > the feelers out now... >=20 > -Jay 'Eraserhead' Felice >=20 > _______________________________________________ > Eof-dev mailing list > Eo...@li... > http://lists.sourceforge.net/lists/listinfo/eof-dev |
From: Jay 'E. F. <era...@er...> - 2001-05-22 01:10:34
|
Just curious, I bopped around to the Free DNS Project, then to here where I saw the FNS spec. This is something I'd like to work on/theorize about/code. 'Specially the latter. The spec that I saw was a great start, but I've already got a bunch of additions for how it should and shouldn't work. Is anybody coding on it? I'd like to work on it in C with -lfreenet if not. I realize we'd probably scrap the first few tries, but I have a sort of masochistic tendency to want to write things more than once anyway (tortured artist syndrome). I can't start coding for another three weeks, probably, but I'd like to put the feelers out now... -Jay 'Eraserhead' Felice |
From: Brandon <bl...@ut...> - 2001-05-21 10:30:21
|
> Can we please start doing formal testing of eof mail, with a view towards a > release in a month or so? I am yet to recieve an email via eof, but I have > been able to use incrementing KSKs to get messages to people using the > standard CLI tools of Freenet. We need to get rid of the vapourware tag. Yes! I have been busy graduating, but that is now over and I'm ready to work. Let's start with that incremental KSK testing again. Insert files for me at KSK@brandon-1, KSK@brandon-2 with incremental HTLs and let me know when you've done it. Then I'll download them, see how many hops you are from me, and that way we can set our mail HTLs so that I'm sure that your mails will get to me. Then we can start testing the mail software itself. Hopefully by the end of the month we can have something suitable to replace the ad-hoc freesite feedback mechanisms. |
From: David F. <dav...@ya...> - 2001-05-20 21:47:11
|
Can we please start doing formal testing of eof mail, with a view towards a release in a month or so? I am yet to recieve an email via eof, but I have been able to use incrementing KSKs to get messages to people using the standard CLI tools of Freenet. We need to get rid of the vapourware tag. Please fmail me: dav...@fr...ee David -- David Findlay ---------- Email: dav...@ya... Homepage: http://users.bigpond.com/nedz/ Segmentation Fault. (Core dumped) |
From: David F. <dav...@ya...> - 2001-05-20 11:57:47
|
This just fixes a small compilialtion error for GUI chat. David -- David Findlay ---------- Email: dav...@ya... Homepage: http://users.bigpond.com/nedz/ Segmentation Fault. (Core dumped) |
From: Timm M. <ti...@ma...> - 2001-05-10 15:01:23
|
I've nearly completed a GUI for EOF Chat. The only things left is to parse commands (like /nick) passed to it durring run time and to actualy connect through the EOF Chat client *g*. Unfortunatly, I have no code commit to CVS right now. The only computer that has the code on it's hard drive has no network connection, and it's floppy drive was fraged while I moved that computer into a new case. As soon as I get the floppy drive up again, I'll be sure to commit it. Timm Murray -------------------- Life is like a perl script: Really short and messy. |
From: David F. <dav...@ya...> - 2001-05-09 21:29:11
|
I just tried sending a message to myself - can't even manage to receive that back. There has to be some major bug stopping us from receiving mails. David -- David Findlay ---------- Email: dav...@ya... Homepage: http://users.bigpond.com/nedz/ Segmentation Fault. (Core dumped) |
From: Brandon <bl...@ut...> - 2001-05-08 16:36:57
|
> different note, I've been trying to figure out how to inititalize > NewsFolder properly, for it requires three pieces of information (date > of latest post requested, index of latest post requested, index of > lastest post today (not necessarily requested yet)). The problem here > is that if a standard sort of mailbox is used with freenet-news, this > info will not be stored. But then, it is advantageous to use a > standard sort of mailbox with freenet-news for various reasons. You should try looking at eof.mail.MailIndex. It handles writing and reading these things from a file. You can either use eof.mail.IndexManager to handle interface to the MailIndex, in which case it will puts your messages in a standard mailbox, or else you can copy this code into a separate eof.news class if for some reason you want to handle things differently than with mail. |
From: Travis B. <be...@ex...> - 2001-05-08 11:56:08
|
On Mon, May 07, 2001 at 11:52:54PM -0500, Brandon wrote: >=20 > > I honestly think that the arrangement now used by IndexClient is > > rather nonideal. Why not use a constant date in milliseconds instead > > of an offset from the current date in milliseconds? >=20 > You're totally right. I must not have been thinking clearly when I did > that. I'll fix it today or tomorrow. I'll then have to change NewsFolder again, but so what. On a different note, I've been trying to figure out how to inititalize NewsFolder properly, for it requires three pieces of information (date of latest post requested, index of latest post requested, index of lastest post today (not necessarily requested yet)). The problem here is that if a standard sort of mailbox is used with freenet-news, this info will not be stored. But then, it is advantageous to use a standard sort of mailbox with freenet-news for various reasons. --=20 Yes, I know my enemies. They're the teachers who tell me to fight me. Compromise, conformity, assimilation, submission, ignorance, hypocrisy, brutality, the elite. All of which are American dreams. - Rage Against The Machine |
From: Brandon <bl...@ut...> - 2001-05-08 04:52:31
|
> I honestly think that the arrangement now used by IndexClient is > rather nonideal. Why not use a constant date in milliseconds instead > of an offset from the current date in milliseconds? You're totally right. I must not have been thinking clearly when I did that. I'll fix it today or tomorrow. |
From: Brandon <bl...@ut...> - 2001-05-08 04:51:30
|
> What is the status of transient nodes running EOF? If I don't run my node > today, will I be able to receive messages with today's date? There is currently no support to pick up where you left off and catch up to today. So if you run your node every other day, you'll get mail that was sent every other day and not get the rest of the mail. This shouldn't be too hard to add. I just haven't gotten to it yet. I'll put it on my TODO list. > P.S. Blanu: have you got those keys yet? I haven't had time as I got deluged in work all of the sudden. I'll get back to you at the end of this week when my craziness is over. :-) |
From: Brandon <bl...@ut...> - 2001-05-07 19:21:27
|
> What do those hashes mean on the freenet_server console when using freenet > mail? IndexClient puts . when it gets a file and # when it fails to get one. Three #'s and it quits. This is really only useful when using it from the command line and I should take it out. I just keep it in for debugging. You can tell when you've received mail (.'s) even if it doesn't show up in your mailbox. |
From: Travis B. <be...@ex...> - 2001-05-07 03:58:59
|
On Sun, May 06, 2001 at 09:32:04PM -0400, Travis Bemann wrote: > I've now changed NewsFolder to use IndexClient, and now it is much > simpler than before. I also found that the problem of IndexClient > using relative dates rather than absolute dates actually was not very > significant. I haven't yet changed NewsStore to properly initialize > NewsFolder classes (they now require some new parameters that they > didn't before), but changing NewsStore to properly work with > NewsFolder shouldn't be much of a hassle. I've now changed NewsFolder to fix a bug that I introduced just earlier today; a silly mistake. Not much of a change. --=20 Yes, I know my enemies. They're the teachers who tell me to fight me. Compromise, conformity, assimilation, submission, ignorance, hypocrisy, brutality, the elite. All of which are American dreams. - Rage Against The Machine |
From: David F. <dav...@ya...> - 2001-05-07 03:16:04
|
What do those hashes mean on the freenet_server console when using freenet mail? David -- David Findlay ---------- Email: dav...@ya... Homepage: http://users.bigpond.com/nedz/ Segmentation Fault. (Core dumped) |
From: Travis B. <be...@ex...> - 2001-05-07 01:32:14
|
I've now changed NewsFolder to use IndexClient, and now it is much simpler than before. I also found that the problem of IndexClient using relative dates rather than absolute dates actually was not very significant. I haven't yet changed NewsStore to properly initialize NewsFolder classes (they now require some new parameters that they didn't before), but changing NewsStore to properly work with NewsFolder shouldn't be much of a hassle. --=20 Yes, I know my enemies. They're the teachers who tell me to fight me. Compromise, conformity, assimilation, submission, ignorance, hypocrisy, brutality, the elite. All of which are American dreams. - Rage Against The Machine |
From: Travis B. <be...@ex...> - 2001-05-07 00:10:23
|
On Sun, May 06, 2001 at 11:52:14AM -0500, Brandon wrote: >=20 > IndexClient in CVS now has -future and -past flags for the command line an > an extended API which allows for specifying an arbitrary date offset. I honestly think that the arrangement now used by IndexClient is rather nonideal. Why not use a constant date in milliseconds instead of an offset from the current date in milliseconds? The problem is that the time is constantly changing while the program runs, so if the GMT date changes while a program is running, it could screw up the operation of that program if the program is not properly designed to handle this situation. Therefore, it is easier to write something which uses an IndexClient which uses constant dates rather than something which uses an IndexClient which uses offsets. The only problem with using constant dates is the 32-bit Unix date rollover. --=20 Yes, I know my enemies. They're the teachers who tell me to fight me. Compromise, conformity, assimilation, submission, ignorance, hypocrisy, brutality, the elite. All of which are American dreams. - Rage Against The Machine |
From: David F. <dav...@ya...> - 2001-05-06 23:46:56
|
On Monday 07 May 2001 02:52, Brandon hit his keyboard with a hammer and this was the result: > IndexClient in CVS now has -future and -past flags for the command line an > an extended API which allows for specifying an arbitrary date offset. What is the status of transient nodes running EOF? If I don't run my node today, will I be able to receive messages with today's date? David P.S. Blanu: have you got those keys yet? -- David Findlay ---------- Email: dav...@ya... Homepage: http://users.bigpond.com/nedz/ Segmentation Fault. (Core dumped) |