The length of vectors, maps, sets, etc are serialized using
Write/ReadCompactSize -- which, unfortunately, do not use a
unique encoding.
So deserializing and then re-serializing a transaction (for example)
can give you different bits than you started with. That doesn't
cause any problems that we are aware of, but it is exactly the type
of subtle mismatch that can lead to exploits.
With this pull, reading a non-canonical CompactSize throws an
exception, which means nodes will ignore 'tx' or 'block' or
other messages that are not properly encoded.
Please check my logic... but this change is safe with respect to
causing a network split. Old clients that receive
non-canonically-encoded transactions or blocks deserialize
them into CTransaction/CBlock structures in memory, and then
re-serialize them before relaying them to peers.
And please check my logic with respect to causing a blockchain
split: there are no CompactSize fields in the block header, so
the block hash is always canonical. The merkle root in the block
header is computed on a vector<CTransaction>, so
any non-canonical encoding of the transactions in 'tx' or 'block'
messages is erased as they are read into memory by old clients,
and does not affect the block hash. And, as noted above, old
clients re-serialize (with canonical encoding) 'tx' and 'block'
messages before relaying to peers.
Orphan transactions were stored as a CDataStream pointer;
this changes the mapOrphanTransactions data structures to
store orphans as a CTransaction.
This also fixes CVE-2013-4627 by always re-serializing
transactions before relaying them.
Compiling on my OSX 10.6 build machine, I get:
Undefined symbols:
"boost::chrono::steady_clock::now()", referenced from:
boost::cv_status boost::condition_variable::wait_for<long long, boost::ratio<1ll, 1000000000ll> >(boost::unique_lock<boost::mutex>&, boost::chrono::duration<long long, boost::ratio<1ll, 1000000000ll> > const&)in bitcoinrpc.o
Linking against the boost_chrono fixes the issue.
Windows builds already link against boost_chrono; Linux doesn't, but compiles (on pull-tester / gitian, at least).
New method in bitcoinrpc: RunLater, that uses a map of deadline
timers to run a function later.
Behavior of walletpassphrase is changed; before, calling
walletpassphrase again before the lock timeout passed
would result in: Error: Wallet is already unlocked.
You would have to call lockwallet before walletpassphrase.
Now: the last walletpassphrase with correct password
wins, and overrides any previous timeout.
Fixes issue# 1961 which was caused by spawning too many threads.
Test plan:
Start with encrypted wallet, password 'foo'
NOTE:
python -c 'import time; print("%d"%time.time())'
... will tell you current unix timestamp.
Try:
walletpassphrase foo 600
getinfo
EXPECT: unlocked_until is about 10 minutes in the future
walletpassphrase foo 1
sleep 2
sendtoaddress mun74Bvba3B1PF2YkrF4NsgcJwHXXh12LF 11
EXPECT: Error: Please enter the wallet passphrase with walletpassphrase first.
walletpassphrase foo 600
walletpassphrase foo 0
getinfo
EXPECT: wallet is locked (unlocked_until is 0)
walletpassphrase foo 10
walletpassphrase foo 600
getinfo
EXPECT: wallet is unlocked until 10 minutes in future
walletpassphrase foo 60
walletpassphrase bar 600
EXPECT: Error, incorrect passphrase
getinfo
EXPECT: wallet still scheduled to lock 60 seconds from first (successful) walletpassphrase
So we stop getting pull requests (like #2604) fixing problems with disabled Script opcodes.
A hard fork would be required to re-enable these, and if we ever did that we'd require extensive review and testing.
When debugging another issue, I found a hang-during-startup race condition due to
LoadWallet calling SetMinVersion (via LoadCryptedKey).
Writing to the file that you're in the process of reading is a bad idea.
Bitcoin-Qt could core dump if application initialization failed in certain ways.
I double-fixed this:
1) qt/bitcoin.cpp now shuts down core threads cleanly if AppInit2 returns false
2) init.cpp now exits before StartNode() if strErrors is set (no reason to StartNode if we're just going to exit immediately anyway).
Tested by triggering all of the various ways AppInit2 can fail, either by passing bogus command-line arguments or just recompiling tweaked code to simulate failure.
This is a partial fix for #2480
Two reasons for this change:
1. Need to always use boost::thread's sleep, even on Windows, so the
sleeps can be interrupted (prior code used Windows' built-in Sleep).
2. I always forgot what units the old Sleep took.
Create a boost::thread_group object at the qt/bitcoind main-loop level
that will hold pointers to all the main-loop threads.
This will replace the vnThreadsRunning[] array.
For testing, ported the BitcoinMiner threads to use its
own boost::thread_group.
Gavin Andresen