Are "seedAddrs" in "SeedPeers.java" still online?

[Green Lantern]

I'm trying to use SeedPeers.java in android app for connection to bitcoin network. But it doesn't look normal. I can see in ddms only "Peer died" or "connection timed out": http://i58.fastpic.ru/big/2013/0922/90/d002e868d692611371894ba100c11490.jpg
What is the problem if the "seedAddrs" is online?

[1714161850]

1714161850

[1714161850]

1714161850

[1714161850]

1714161850

[Mike Hearn]

Why are you so keen on using the seed nodes?

I haven't updated them for a long time. Neither has anyone updated the ones in bitcoind. We really should, it's a good reminder to do so .... it hasn't been done though. So by now the IPs are all long since stale.

[Green Lantern]

I haven't updated them for a long time. Neither has anyone updated the ones in bitcoind.

You said that IPs hadn't been updated for a long time in bitcoinj and in bictoin client for windows. Right?
So can you tell why seedAddrs (https://code.google.com/p/bitcoinj/source/browse/core/src/main/java/com/google/bitcoin/discovery/SeedPeers.java line 87) in bitcoinj differ from pnSeed in bitcoin cpp client for windows (https://github.com/bitcoin/bitcoin/blob/0.8.1/src/net.cpp line 1226) though there (https://code.google.com/p/bitcoinj/source/browse/core/src/main/java/com/google/bitcoin/discovery/SeedPeers.java line 28) is such sentence: "The list comes from the Bitcoin C++ source code"? I can not see concurrences.

Why are you so keen on using the seed nodes?

So all of 200 IPs are stale? Neither are working?
Hm... It will break my suppositions about not correct work of "SeedPeers" class.  
Why am I keep using seed nodes? Because I am trying to modify bitcoinj and I need work throw "SeedPeers" class not throw "DnsDiscovery". I'll explain more in PM if you are interested.

[Mike Hearn]

Well, the comment was correct when I first wrote that code. Obviously the C++ code was refreshed at some point (many months ago) but I didn't copy the IPs into bitcoinj.

We should improve the process for refreshing the seed peers. I guess the DNS seeds have sufficient data to calculate which peers are the most stable.

[Green Lantern]

Well, the comment was correct when I first wrote that code. Obviously the C++ code was refreshed at some point (many months ago) but I didn't copy the IPs into bitcoinj.

We should improve the process for refreshing the seed peers. I guess the DNS seeds have sufficient data to calculate which peers are the most stable.

I want to translate part of pnSeed form C++ code to seedAddrs in java bitcoinj.
So what about little-endian and big-endian notation? Are seedAddrs (https://code.google.com/p/bitcoinj/source/browse/core/src/main/java/com/google/bitcoin/discovery/SeedPeers.java lin 87) using little-endian and pnSeed (https://github.com/bitcoin/bitcoin/blob/0.8.1/src/net.cpp line 1226) using big-endian notation whether it is other way round?
It is mean that I should write first element "0xe473042e" from pnSeed like "0x2e0473e4" in seedAddrs, doesn't it?

[Mike Hearn]

I filed this bug:

https://code.google.com/p/bitcoinj/issues/detail?id=457

The right way to do it is load the text file from that URL dynamically (I mean load it from a JAR resource, and we'll copy it into the source tree from time to time).

[Mike Hearn]

Yes, you have to reverse the integers.

[Green Lantern]

Yes, you have to reverse the integers.

So, my example is correct?

It is mean that I should write first element "0xe473042e" from pnSeed like "0x2e0473e4" in seedAddrs, doesn't it?

Thanks so much!

[Mike Hearn]

Well, if it doesn't work you got them the wrong way around

[Green Lantern]

Well, if it doesn't work you got them the wrong way around

I have reversed all integers in pnSeed array from https://github.com/bitcoin/bitcoin/blob/0.8.1/src/net.cpp (line 1226). It doesn't work!
Then I simply pasted part of pnSeed to seedAddrs. It works.
So, reverse isn't needed.

[Mike Hearn]

OK, funny, I seem to remember having to reverse them, but it was years ago so I obviously misremembered.