Well, gyus, i filled in the form. What next?
Yesterday i played a bit with AROSTCP source code and here is my first progress report: it builds. It even starts up. Then it crashes at NETTRACE (syslog server) startup, don't know why yet, too tired and went asleep.
I specified the deadline as 2 months. I hope the period extension won't be required.

good news, i hope you'll make it Sonic, thanks for taking care of it :)

Very interesting news .. some super mos devel should offer you some help i think.

Quote:

Sonic keep in mind you need to meet these if you port the AROS stack even. DHCP is a must. The GUI can be left out as long as setup is simple for even a novice user of MOS I would think.

I know that GUI is needed. And i absolutely agree. Of course first GUI will be minimal. Krashan already gave me his working MUI-based syslog viewer, i will use it.
AmiTCP of course has DHCP support. I see no reason why it should not work in AROSTCP.

[ Edited by Sonic on 2005/9/8 22:17 ]

Good news. Thank you for looking at taking this on Sonic. Everything else is basically formalities (putting your name on the page), assuming that no one demands my head for allowing it

goodluck sonic I hope forthe best =)

So do you accept me? I see that the bounty is still open...

as far as I can tell, the aros tcp/ip stack has no support for DHCP. If you think you can upgrade it to be able to utilize dhcp (etc), then we'll kick that bounty closed, erm, assigned! Also, I would like to see Krashan's work continued in this somehow. I'll make those sources public again (was just playing with the permissions, and that made a good test subject).

I wonder if we can talk CISC into chiming in here

I think you should be extremely careful reassigning this bounty .. the first thing you should do is check with everyone that donated that they are ok with downgrading Krashan's efforts (which after all is the reason this bounty is so big now) to a mere port of an ancient version of AmiTCP...

If there is ambiguous support for this, maybe a compromise could be done, like f.ex. setting off a certain amount from the bounty to a new bounty for this port (again, this is something which should be decided by the donators), and reserve the major remains to when/if someone takes up again the initial project.


- CISC

I agree with CISC, ~800$ for just a quick port of aros tcp stack is too much, i would never donated anything for this.

I think a new bounty should be created and keep EzTCP
as it is.

To me this looks like its just a quick earn, and a spit
in Krashans face.
I belive the reason for the huge donations was because it was Krashan doing it and he wrote it from scratch.

What else have this Sonic_ru(?) done so we can atleast
know what quality coder he is ?

An interesting debate for sure but if the program fulfils the goal of the original bounty then it should surelybe paid, no matter what the source of the code (no pun intended)

I have to agree with JKD here; the terms where set, and people donated for the product, not the coder. Even though I can see people doniting (more) because Krashan took the bounty, it still was a donation for the project as laid down in the said terms. If someone can fullfill those terms, he has all the rights to claim the bounty.

You have to understand that alot of people probably donated when they saw what Krashan was doing, and forwarded their money on the basis of that .. if this isn't cleared with the donators themselves you might face several refund claims, and we don't want that do we?


- CISC

I see the hotbed, but still any coder that meets the bounty terms has the right to claim the bounty. If people think they should draw back from the bounty because they don't have faith in the coder/donated for another reason than the plain fullfillment of the bounty, they should.

Ideal situation? No, but fair.

I have to agree with jkd and the others here. As it is AmiTCP does NOT meet the guide lines of the bounty. It would have to be updated and modernized a lot. DHCP is a must as is a workable GUI. A quick port of the AROS stack will not cut it.

As to the money, well it was for the bounty not the coder even if those that put it in only did so for Krashan. They had plenty of time to request a refund as in the rules of the system. If I recall Targhan even asked about refunds. But if something needs to be worked out with people I am sure we can come to some sort of agreement everyone can be happy with.

So in the end if Sonic does the job and meets to goals then he can claim it.

[ Edited by Acill on 2005/9/10 6:13 ]

Acill.. I do othervice agree fully, but..

Has anyone checked how old Berkeley-Netstack -code Amitcp (or it's aros variant) are based on? I haven't checked Aros code, but I know how old original 68k AmiTCP is, and it's *ancient*.

IMHO it would be quite quite questionable to support re-port of decade (or two) old code.. There's been plenty of buxfies, new features etc made to Berkeley Stack since AmiTCP was made.

I have donated $200 for the ezTCP bounty (and I planned to donate a lot more as things shaped up and started to materialize). I did it because of two reasons (in order):

1) I feel a desperate need for a PPC native TCP/IP stack with working DHCP

2) Krashan really sold his project to me, I liked his ideas and his approach.

I don't know the reasons to Krashans sudden stop, but I think it was very sad. But where to go from here?

Well, I don't know the chances of Krashan (or anyone else capable of understand Krashans ideas and approach good enough to make something *very similar*) will pick up the "original" ezTCP project. If chances are small (which I think they are), then what's the point in keeping a bounty alive for it anyway? Wouldn't the money make more use elsewhere, like in a bounty for a TCP/IP stack that actually has a chance of materializing and be released? Because even if I think it was a shame that Krashan stopped his work, the first reason to why I donated to the bounty still apply: I feel a desperate need for a PPC native TCP/IP stack!

But I see the points raised by people in this thread. Hmm, maybe the ones still having hopes of someone actually picking up Krashans sources to make a stack very similar to what was *originally* "announced" could keep their money in that bounty, everyone else can withdraw theirs, and then a *new* bounty (for *another* stack) could start? Would that be a good compromise?

I mean, if you clever people here in this thread can come to a an agreement that the AROS TCP/IP stack is *not* totally ancient, and that a PPC native version of it would be at least *better* than the 68k Amiga stacks we currently use, then feel free to use my money to kickstart a new bounty to simply port the AROS stack! In the current TCP/IP situation for MorphOS I see no place for asking for absolute perfection; just give me a goddamn PPC TCP/IP stack! Now! Basta!

[ Edited by takemehomegrandma on 2005/9/10 14:01 ]

The bounty for ezTCP stack isnt jus tthe one Krashan did, its the PPC native stack bounty. It can stay as is since all the guides for it are the same. If a new project starts then thats fine. I dont think anyone can take over Krashans code myself, but a new start or a very modern update to the AROS stack is fine as long as all the rules of the bounty are met.

Well, well...
Yesterday i compiled the working version. It started ans succesfully pinged itself (127.0.0.1)
The roadmap for a project was:
1. Build a working kernel.
2. Create a modularized GUI consisting at least of 3 components:
a) small control panel (capable of controlling every interface, putting it online and offline.
b) syslog viewer (this is already taken from Krashan's work, needs only to be integrated).
c) configuration editor, very much like one from Genesis.
The whole system could be linked together with events notification system. AROSTCP code is very well written, it can be easily extended and adding events is not an impossible task, current code allows it, it is just not done.
3. Check DHCP, fix (implement) it.
4. At this point the bounty is complete, project goes opensource. Further suggested tasks are (not in order of priority):
1. Rewrite PPP device from scratch. Add a good dialer (like in Miami, not like in Genesis).
2. Write PPPoE device from scratch.
3. Re-add inetd dropped in AROS distribution.
4. Fixes and improvements, your suggestions, etc.
Well, well... I see you wish to pay Krashan but don't wish to pay me just because it's me. Ok, you'll get what you want. I HALT THE PROJECT!!! I wouldn't like to waste 2 months and when everything is done hear something like "code is bad" (but the actual work is done, it's usable, clever, right?). Or get $50 from some new bounty. Well, pay to someone you WISH to pay to...
At least thanks for that you told it NOW and i didn't lose my time. I'll better use my evenings for php coding and industrial Pentium motherboards repair because someone pays me for this. Not so much, but they DO unlike you.
I didn't force anyone, YOU suggested the bounty. And usually it's suggested to answer for your words... Don't promise anything if you don't want to keep your promise...
Of course i don't erase my code, i can continue at any time. But i will do it only when you accept me as a project coder and it the amount will be reasonable (in case if you remove all the money and start new bounty).
Krashan's work is not dead, i use parts of it.

[ Edited by Sonic on 2005/9/12 9:42 ]

Privet Sonic

Take it easy now.. this thread is about discussion the bounty. So let's do that, discuss. Remember, different people have different opinions, and not all of them necessarily are similar to yours.

You made it sound like people dont want to pay because they dont like you.. I'm pretty sure thats not the case here

MY take on this: A working PPC TCP/IP stack with high reliability should be enough to claim the bounty. No matter WHO started the project.

Well I must say that isn't this a little childish (sp?), I mean hera a guy appears use some open source code from aros, which is another amiga clone, which I thought the Morph community actuly liked, and thenm he can't collect the bounty? Don't be stupod! Who of you are actuly going to creat a stack instead?

I've checked the code, there's really no DHCP in current code. But it's fixable. There are two ways:
1. Build in some DHCP client, probably based on Krashan's code (there are at least parts).
2. Take UNIX dhclient, port it and bind to the suite.
The configuration of the interfaces will be performed by some script (or program) basing on some configuration file like:
eth0 IP=192.168.0.1 NETMASK=255.255.255.0 UP
eth1 DHCP
The script will read this file, parse it and configure the interface via ifconfig or dhclient. This process will be performed transparently to the user. The user will see only "Online" and "Offline" buttons and status LED.

Well, as I've said earlier, I really think that any stack/project meeting the bounty requirements is entitled to the money. No matter who does it, no matter what code is involved.

Besides that, I have full confidence in Sonic's skills

@Sonic

My share of the bounty (~22%) will go to *anyone* releasing a good performing, PPC native TCP/IP stack, capable of using DHCP and handling several network interfaces. I hope that will be you!