I think that's a great approach. Here is the patch Satoshi sent me back in 2010. All the code has changed since but it can be a source of inspiration. From Satoshi: *The simplified payment verification in the paper imagined you would receive transactions directly, as with sending to IP address which nobody uses, or a node would index all transactions by public key and you could download them like downloading mail from a mail server. Instead, I think client-only nodes should receive full blocks so they can scan them for their own transactions. They don't need to store them or index them. For the initial download, they only need to download headers, since there couldn't be any payments before the first time the program was run (a header download command was added in 0.3.18). From then on, they download full blocks (but only store the headers). Code for client-only mode is mostly implemented. There's a feature branch on github with it, also I'm attaching the patch to this message. Here's some more about it: "Here's my client-mode implementation so far. Client-only mode only records block headers and doesn't use the tx index. It can't generate, but it can still send and receive transactions. It's not fully finished for use by end-users, but it doesn't matter because it's a complete no-op if fClient is not enabled. At this point it's mainly documentation showing the cut-lines for client-only re-implementers. With fClient=true, I've only tested the header-only initial download. A little background. CBlockIndex contains all the information of the block header, so to operate with headers only, I just maintain the CBlockIndex structure as usual. The nFile/nBlockPos are null, since the full block is not recorded on disk. The code to gracefully switch between client-mode on/off without deleting blk*.dat in between is not implemented yet. It would mostly be a matter of having non-client LoadBlockIndex ignore block index entries with null block pos. That would make it re-download those as full blocks. Switching back to client-mode is no problem, it doesn't mind if the full blocks are there. If the initial block download becomes too long, we'll want client mode as an option so new users can get running quickly. With graceful switch-off of client mode, they can later turn off client mode and have it download the full blocks if they want to start generating. They should rather just use a getwork miner to join a pool instead. Client-only re-implementations would not need to implement EvalScript at all, or at most just implement the five ops used by the standard transaction templates." * diff -u old\db.cpp new\db.cpp --- old\db.cpp Sat Dec 18 18:35:59 2010 +++ new\db.cpp Sun Dec 19 20:53:59 2010 @@ -464,29 +464,32 @@ ReadBestInvalidWork(bnBestInvalidWork); // Verify blocks in the best chain - CBlockIndex* pindexFork = NULL; - for (CBlockIndex* pindex = pindexBest; pindex && pindex->pprev; pindex = pindex->pprev) + if (!fClient) { - if (pindex->nHeight < nBestHeight-2500 && !mapArgs.count("-checkblocks")) - break; - CBlock block; - if (!block.ReadFromDisk(pindex)) - return error("LoadBlockIndex() : block.ReadFromDisk failed"); - if (!block.CheckBlock()) + CBlockIndex* pindexFork = NULL; + for (CBlockIndex* pindex = pindexBest; pindex && pindex->pprev; pindex = pindex->pprev) { - printf("LoadBlockIndex() : *** found bad block at %d, hash=%s\n", pindex->nHeight, pindex->GetBlockHash().ToString().c_str()); - pindexFork = pindex->pprev; + if (pindex->nHeight < nBestHeight-2500 && !mapArgs.count("-checkblocks")) + break; + CBlock block; + if (!block.ReadFromDisk(pindex)) + return error("LoadBlockIndex() : block.ReadFromDisk failed"); + if (!block.CheckBlock()) + { + printf("LoadBlockIndex() : *** found bad block at %d, hash=%s\n", pindex->nHeight, pindex->GetBlockHash().ToString().c_str()); + pindexFork = pindex->pprev; + } + } + if (pindexFork) + { + // Reorg back to the fork + printf("LoadBlockIndex() : *** moving best chain pointer back to block %d\n", pindexFork->nHeight); + CBlock block; + if (!block.ReadFromDisk(pindexFork)) + return error("LoadBlockIndex() : block.ReadFromDisk failed"); + CTxDB txdb; + block.SetBestChain(txdb, pindexFork); } - } - if (pindexFork) - { - // Reorg back to the fork - printf("LoadBlockIndex() : *** moving best chain pointer back to block %d\n", pindexFork->nHeight); - CBlock block; - if (!block.ReadFromDisk(pindexFork)) - return error("LoadBlockIndex() : block.ReadFromDisk failed"); - CTxDB txdb; - block.SetBestChain(txdb, pindexFork); } return true; diff -u old\main.cpp new\main.cpp --- old\main.cpp Sat Dec 18 18:35:59 2010 +++ new\main.cpp Sun Dec 19 20:53:59 2010 @@ -637,6 +637,9 @@ if (!IsStandard()) return error("AcceptToMemoryPool() : nonstandard transaction type"); + if (fClient) + return true; + // Do we already have it? uint256 hash = GetHash(); CRITICAL_BLOCK(cs_mapTransactions) @@ -1308,23 +1311,26 @@ if (!CheckBlock()) return false; - //// issue here: it doesn't know the version - unsigned int nTxPos = pindex->nBlockPos + ::GetSerializeSize(CBlock(), SER_DISK) - 1 + GetSizeOfCompactSize(vtx.size()); - - map mapUnused; - int64 nFees = 0; - foreach(CTransaction& tx, vtx) + if (!fClient) { - CDiskTxPos posThisTx(pindex->nFile, pindex->nBlockPos, nTxPos); - nTxPos += ::GetSerializeSize(tx, SER_DISK); + //// issue here: it doesn't know the version + unsigned int nTxPos = pindex->nBlockPos + ::GetSerializeSize(CBlock(), SER_DISK) - 1 + GetSizeOfCompactSize(vtx.size( )); + + map mapUnused; + int64 nFees = 0; + foreach(CTransaction& tx, vtx) + { + CDiskTxPos posThisTx(pindex->nFile, pindex->nBlockPos, nTxPos); + nTxPos += ::GetSerializeSize(tx, SER_DISK); - if (!tx.ConnectInputs(txdb, mapUnused, posThisTx, pindex, nFees, true, false)) + if (!tx.ConnectInputs(txdb, mapUnused, posThisTx, pindex, nFees, true, false)) + return false; + } + + if (vtx[0].GetValueOut() > GetBlockValue(pindex->nHeight, nFees)) return false; } - if (vtx[0].GetValueOut() > GetBlockValue(pindex->nHeight, nFees)) - return false; - // Update block index on disk without changing it in memory. // The memory index structure will be changed after the db commits. if (pindex->pprev) @@ -1378,7 +1384,7 @@ foreach(CBlockIndex* pindex, vDisconnect) { CBlock block; - if (!block.ReadFromDisk(pindex)) + if (!block.ReadFromDisk(pindex, !fClient)) return error("Reorganize() : ReadFromDisk for disconnect failed"); if (!block.DisconnectBlock(txdb, pindex)) return error("Reorganize() : DisconnectBlock failed"); @@ -1395,7 +1401,7 @@ { CBlockIndex* pindex = vConnect[i]; CBlock block; - if (!block.ReadFromDisk(pindex)) + if (!block.ReadFromDisk(pindex, !fClient)) return error("Reorganize() : ReadFromDisk for connect failed"); if (!block.ConnectBlock(txdb, pindex)) { @@ -1526,7 +1532,7 @@ txdb.Close(); - if (pindexNew == pindexBest) + if (!fClient && pindexNew == pindexBest) { // Notify UI to display prev block's coinbase if it was ours static uint256 hashPrevBestCoinBase; @@ -1547,10 +1553,6 @@ // These are checks that are independent of context // that can be verified before saving an orphan block. - // Size limits - if (vtx.empty() || vtx.size() > MAX_BLOCK_SIZE || ::GetSerializeSize(*this, SER_NETWORK) > MAX_BLOCK_SIZE) - return error("CheckBlock() : size limits failed"); - // Check proof of work matches claimed amount if (!CheckProofOfWork(GetHash(), nBits)) return error("CheckBlock() : proof of work failed"); @@ -1559,6 +1561,13 @@ if (GetBlockTime() > GetAdjustedTime() + 2 * 60 * 60) return error("CheckBlock() : block timestamp too far in the future"); + if (fClient && vtx.empty()) + return true; + + // Size limits + if (vtx.empty() || vtx.size() > MAX_BLOCK_SIZE || ::GetSerializeSize(*this, SER_NETWORK) > MAX_BLOCK_SIZE) + return error("CheckBlock() : size limits failed"); + // First transaction must be coinbase, the rest must not be if (vtx.empty() || !vtx[0].IsCoinBase()) return error("CheckBlock() : first tx is not coinbase"); @@ -1623,13 +1632,14 @@ return error("AcceptBlock() : out of disk space"); unsigned int nFile = -1; unsigned int nBlockPos = 0; - if (!WriteToDisk(nFile, nBlockPos)) - return error("AcceptBlock() : WriteToDisk failed"); + if (!fClient) + if (!WriteToDisk(nFile, nBlockPos)) + return error("AcceptBlock() : WriteToDisk failed"); if (!AddToBlockIndex(nFile, nBlockPos)) return error("AcceptBlock() : AddToBlockIndex failed"); // Relay inventory, but don't relay old inventory during initial block download - if (hashBestChain == hash) + if (!fClient && hashBestChain == hash) CRITICAL_BLOCK(cs_vNodes) foreach(CNode* pnode, vNodes) if (nBestHeight > (pnode->nStartingHeight != -1 ? pnode->nStartingHeight - 2000 : 55000)) @@ -2405,6 +2415,8 @@ { if (fShutdown) return true; + if (fClient && inv.type == MSG_TX) + continue; pfrom->AddInventoryKnown(inv); bool fAlreadyHave = AlreadyHave(txdb, inv); @@ -2441,6 +2453,9 @@ if (inv.type == MSG_BLOCK) { + if (fClient) + return true; + // Send block from disk map::iterator mi = mapBlockIndex.find(inv.hash); if (mi != mapBlockIndex.end()) @@ -2486,6 +2501,8 @@ else if (strCommand == "getblocks") { + if (fClient) + return true; CBlockLocator locator; uint256 hashStop; vRecv >> locator >> hashStop; @@ -2556,6 +2573,8 @@ else if (strCommand == "tx") { + if (fClient) + return true; vector vWorkQueue; CDataStream vMsg(vRecv); CTransaction tx; @@ -2620,6 +2639,33 @@ if (ProcessBlock(pfrom, &block)) mapAlreadyAskedFor.erase(inv); + } + + + else if (strCommand == "headers") + { + if (!fClient) + return true; + vector vHeaders; + vRecv >> vHeaders; + + uint256 hashBestBefore = hashBestChain; + foreach(CBlock& block, vHeaders) + { + block.vtx.clear(); + + printf("received header %s\n", block.GetHash().ToString(). substr(0,20).c_str()); + + CInv inv(MSG_BLOCK, block.GetHash()); + pfrom->AddInventoryKnown(inv); + + if (ProcessBlock(pfrom, &block)) + mapAlreadyAskedFor.erase(inv); + } + + // Request next batch + if (hashBestChain != hashBestBefore) + pfrom->PushGetBlocks(pindexBest, uint256(0)); } On Tue, Jul 16, 2013 at 12:17 PM, Wendell wrote: > I for one would be interested in taking a look at that. > > In San Jose I was asking around about the possibility of hiring someone to > complete such a patch. Pieter Wuille introduced me to Eric Lombrozo, who > expressed interest, but has since gotten quite busy. So we haven't arrived > at a detailed estimate of what it would involve. > > Maybe it would be better to start a completely new thread for this topic, > but I would very much be interested in an open dissection of what adding > SPV support to bitcoind would take. I am willing to fund or (ideally) > co-fund this endeavor, if I can ever get my head around it. I'm super > interested in all of these possibilities (including micro-stripped-VMs and > transpilation), but would simply like to encourage the proliferation of > _options_ whenever possible. > > -wendell > > grabhive.com | twitter.com/grabhive > > On Jul 16, 2013, at 11:51 AM, Mike Hearn wrote: > > > If you wanted to implement SPV mode in bitcoind, Gavin or I could send > you Satoshi's old patch although of course it is no longer usable. It would > indicate the basic cut lines though. > >