New Ixcoin fork -> I0coin

[Technocrat]

Do you realize how false your sentence is?

Actually I don't. Maybe I'm just stupid. So you can still send I0coins?

[d.james]

The founder has disappeared. Is I0coins dead?

Yes bitcoin is dead because Satoshi disappeared.

[ArtForz]

Actually, just realized something... as ix will also switch to 3-min blocks but has 96 coins/block... that's twice our inflation rate... can't have that.
so... either also switch to 96 coins/block (boooring!) or go to 90 seconds/block target while keeping the 48 coin reward.
At 90 sec/block... change retarget to every 120 blocks (= every 3h nominal), keep the *1.1 /4 limit per retarget, change block acceptance rules so blocks can't be more than 5min or so in the future (we still use bitcoins 120min, but we have way more accurate time thanks to integrated NTP...)
And while we're at it, go to diff 256 instead of 4096 after block 14639.
"special" enough?

[d.james]

sounds good on paper, at difficulty of 256, it's going to lure a small number miners to get in on this.
wish we can test it out in someway, but this thing might actually turn out to be an improved version of solidcoin after all.
we'll just have to wait and see.

[fusebox]

Actually, just realized something... as ix will also switch to 3-min blocks but has 96 coins/block... that's twice our inflation rate... can't have that.
so... either also switch to 96 coins/block (boooring!) or go to 90 seconds/block target while keeping the 48 coin reward.
At 90 sec/block... change retarget to every 120 blocks (= every 3h nominal), keep the *1.1 /4 limit per retarget, change block acceptance rules so blocks can't be more than 5min or so in the future (we still use bitcoins 120min, but we have way more accurate time thanks to integrated NTP...)
And while we're at it, go to diff 256 instead of 4096 after block 14639.
"special" enough?

Fine with me.. can you tell me the code to change and I'll do it.. or maybe there's a way to add you to the project.... I'm still working on getting this damn thing to compile ( worked on the prereqs this morning before life happened and fininishing that up now).

[doublec]

Well, once theres reasonable consensus on what the new algo should be, it's just a matter of changing the constants and finding a suck... masochi... kind soul to build binaries.
Probably not a too good idea to open a exchange back up before "everyone" switched and the network gets at least a few dozen Ghps.

Yeah, I'll enable the exchange once it's been running for a bit and seems to be working ok. I'll update my pool with the new code though to help move it along.

[ditchmagnet]

So can we actually mine IOcoin yet?  I didn't want to read all through the pages, and there is nothing about it in the OP.

[Starlightbreaker]

So can we actually mine IOcoin yet?  I didn't want to read all through the pages, and there is nothing about it in the OP.

you're late to the party.

[fusebox]

Actually, just realized something... as ix will also switch to 3-min blocks but has 96 coins/block... that's twice our inflation rate... can't have that.
so... either also switch to 96 coins/block (boooring!) or go to 90 seconds/block target while keeping the 48 coin reward.
At 90 sec/block... change retarget to every 120 blocks (= every 3h nominal), keep the *1.1 /4 limit per retarget, change block acceptance rules so blocks can't be more than 5min or so in the future (we still use bitcoins 120min, but we have way more accurate time thanks to integrated NTP...)
And while we're at it, go to diff 256 instead of 4096 after block 14639.
"special" enough?

Ok, checked in most these changes... except..
>And while we're at it, go to diff 256 instead of 4096 after block 14639.

Not exactly sure how to do that.

[kano]

Firstly https://i0exchange.bitparking.com/ doesn't seem to work any more (after you login it fails)

Secondly, why on earth are people hacking with the code other than to fix the current problems?

Seriously, just use that patch, make a new one and we'll be done.

Copying some other coin ideas and agreeing coz one person posted a comment here on the forum is not a good excuse to be causing more problems
Every other coin who has been making changes keeps screwing up - so don't follow in their footsteps and just fix the current problem.

[MaGNeT]

This one still stands

I would like to setup a bounty for fixing i0coins:

200 i0coins for the one who fixes it!

Who wants to add some more? Quote and add your name to the list!

Bounty list:
MaGNeT - 200 i0coins
d.james - 400 i0coins

Total - 600 i0coins

Remember, we can only pay the bounty if the network is up again  

[fusebox]

Secondly, why on earth are people hacking with the code other than to fix the current problems?

Because the entire network is down and now is the time to fix it (ie no one has published working binaries yet which everyone will need to upgrade too).   Either way these were trivial changes.  Now back to getting this windows binary compiling (I suspect linux users probabaly can easily rebuild and restart thier nodes anytime they wish and will once some people confirm they are up and generating blocks again).

Unless someone posts a quick way to reset the difficulty target for restart in the code.. the probabbaly is the last change (unless someone else takes the reigns and makes thier own forks).

[doublec]

Firstly https://i0exchange.bitparking.com/ doesn't seem to work any more (after you login it fails)

Works fine for me, PM me your error if you're getting one.

[doublec]

Unless someone posts a quick way to reset the difficulty target for restart in the code.. the probabbaly is the last change (unless someone else takes the reigns and makes thier own forks).

The NTP code was broken in the original release. It would cause i0coind to hang if the NTP server didn't respond. It'd be good if that can be fixed. I disabled the NTP code on my systems otherwise it was too reliable.

[ArtForz]

for the switch-to-256 thing...

Code:

--- a/src/main.cpp
+++ b/src/main.cpp
@@ -726,6 +726,9 @@ unsigned int static GetNextWorkRequired(const CBlockIndex* pindexLast)
 //okay, maybe not this line
     if ((pindexLast->nHeight+1) < 14640)
        return GetNextWorkRequired_OLD(pindexLast);
+    //hardcoded switch to 256.0 difficulty at block 14639
+    if ((pindexLast->nHeight+1) == 14640)
+        return 0x1C00FFFF;
 
     // Only change once per interval
     if ((pindexLast->nHeight+1) % nInterval != 0)

should work ?

[ArtForz]

right, the NTP hangs... let's see how hard adding a short timeout would be...

[fusebox]

for the switch-to-256 thing...

Code:

--- a/src/main.cpp
+++ b/src/main.cpp
@@ -726,6 +726,9 @@ unsigned int static GetNextWorkRequired(const CBlockIndex* pindexLast)
 //okay, maybe not this line
     if ((pindexLast->nHeight+1) < 14640)
        return GetNextWorkRequired_OLD(pindexLast);
+    //hardcoded switch to 256.0 difficulty at block 14639
+    if ((pindexLast->nHeight+1) == 14640)
+        return 0x1C00FFFF;
 
     // Only change once per interval
     if ((pindexLast->nHeight+1) % nInterval != 0)

should work ?

ok.. duh... should have figured that out... ... committed.  thanks.

[caston]

Do you realise how ironic this is?

[ArtForz]

First stab at adding a 3s recv timeout to the NTP stuff, I *hate* asio.

Code:

diff --git a/src/util.cpp b/src/util.cpp
index 31dcb57..cfd1932 100644
--- a/src/util.cpp
+++ b/src/util.cpp
@@ -839,6 +839,17 @@ time_t GetNTPTime( const char* ntpServer )
 	socket.send_to(
 			boost::asio::buffer(data),
 			receiver_endpoint);
+	//ugly hack, CBA to switch to asio async I/O just to get a 3 sec timeout
+	int hSocket = socket.native();
+	fd_set fdset;
+	FD_ZERO(&fdset);
+	FD_SET(hSocket, &fdset);
+	struct timeval t;
+	t.tv_sec = 3;
+	t.tv_usec = 0;
+	int n_readable = select(hSocket+1, &fdset, NULL, NULL, &t);
+	if (!n_readable)
+		return time_t(0);
 	socket.receive_from(
 			boost::asio::buffer(data),
 			sender_endpoint);
@@ -890,10 +901,17 @@ int64 GetAdjustedTime()
 	
 	if((time - nTimeNTPLastSync) >= (60 * 5)) // Calculate the NTP offset once every 5 min
 	{
-		int64 ntpTime = GetNTPTime();	
-		nTimeNTPOffset = ntpTime - time;
 		nTimeNTPLastSync = time;
-		printf("nTimeNTPOffset=%d\n",nTimeNTPOffset);
+		int64 ntpTime = GetNTPTime();	
+		if (ntpTime)
+		{
+			nTimeNTPOffset = ntpTime - time;
+			printf("nTimeNTPOffset=%d\n",nTimeNTPOffset);
+		}
+		else
+		{
+			printf("NTP timeout\n");
+		}
 	}
 	
     return time + nTimeNTPOffset;

it compiles, seems to work and doesn't lock up or crash the client... so far.

[caston]

Please set the block number for the change over a bit higher before relaunch.