27f76e2 blockchain_import: Build string for db type list (warptangent)
1aa8a9d blockchain_import: Add mode argument representing multiple DB flags (warptangent)
cffc411 blockchain_import: Support BerkeleyDB (warptangent)
19c1aaa blockchain_import: Add database type argument (warptangent)
0fedce0 blockchain_import: Make LMDB-specific names general (warptangent)
When keys are contiguous and monotonically increasing, this gets
denser page utilization (doesn't leave padding in page splits).
Can't be used for keys that are inserted in random order (e.g. hashes)
In total this only saves around 1.5% of space compared to original
DB code. The previous patch accounted for 0.8% savings on its own;
the blocks tables just aren't that big.
These modes match those optionally provided as part of the daemon's
--db-type argument.
Argument after the # is interpreted as a composite mode if there's only
one (no comma separated arguments).
Sample usage:
blockchain_import --database lmdb#fastest
blockchain_import --database berkeley#fastest
Multiple specific DB flags are still supported, e.g.
blockchain_import --database lmdb#nosync,nordahead
blockchain_import --database berkeley#txn_nosync
1995923 BlockchainLMDB: Deal with DB exceptions at block level with particularity (warptangent)
c16cc20 BlockchainLMDB: Add sanity check for inconsistent state (warptangent)
9118d0a BlockchainLMDB: Call destructor on allocated txn if setup fails (warptangent)
f5581c3 BlockchainLMDB: Replace remaining txn pointer NULLs with nullptr (warptangent)
c7e6b77 crypto: only check MONERO_USE_SOFTWARE_AES once (moneromooo-monero)
74aef21 crypto: use software AES based on the MONERO_USE_SOFTWARE_AES env var (moneromooo-monero)
Add another DB error exception type to distinguish failed txn setup from
general use of txn.
This keeps the error handling flow the same as before the block-level
txn setup changes that moved control up a layer to BlockchainDB.
Remove LINK_SEARCH_START_STATIC and LINK_SEARCH_END_STATIC. This is more
appropriate when the compiler flag -static is used.
This had been causing CMake to omit the linker flags necesssary to
distinguish static and dynamic library linking. CMake had assumed static
linking for the target, causing it to omit explicit static link flags.
That is problematic without the -static compile flag being set.
With a library located in system directories, like libboost_date_time,
the full static path (.a), though found correctly by CMake, was treated
by the linker as a dynamic library. This is because
target_link_libraries() transforms the full path to -l<libname> if it's
in a system directory. Without -static or explicit linker flags, the
dynamic library (.so) is linked.
Removing the above two properties removes the assumption of static. So
-Wl;-Bstatic is inserted where needed. This causes -l<libname> to
properly refer to the static library instead of dynamic.
Setting to no or 0 also works. If set, any other value enables it.
Useful for running with valgrind in cases where it fails at
properly implementing AES-NI.
We also replace the --fakechain option with an optional structure
containing details about configuration for the core/blockchain,
for test purposes. This seems more future friendly.
7658ac0 blockchain: revert handle_get_objects adding block id on tx not found (moneromooo-monero)
3a0f4d8 berkeleydb: fix delete/free mismatch (moneromooo-monero)
1642be2 minor bugfixes and refactoring (Thomas Winget)
098dcf2 unit_tests: fix mnemonics unit test testing invalid seeds (moneromooo-monero)
If it is, it points to reuse of a tx key, which isn't meant to happen.
If it does, a key image collision means that only one of those
outputs is spendable, so the wallet selects the larger amount,
unless that output was spent already.
This causes a discrepancy betewen reported received inputs and
payment total.
Since tx keys are 256 bits, this should never happen except if
done on purpose, or if a sender uses a bad PRNG.
Locking just one db turns out to not have been a good idea, since
the pool and p2p state fdles have to be used anyway.
Also ensure the directory exists before tring to lock.
- Blockchain should store if it's running on testnet or not
- moved loading compiled-in block hashes to its own function for clarity
- on handle_get_objects, should now correctly return false if a block's
transactions are missing
- replace instances of BOOST_FOREACH with C++11 for loops in Blockchain.
7fc6fa3 wallet: forbid dust altogether in output selection where appropriate (moneromooo-monero)
5e1a739 blockchain: log number of outputs available for a new tx (moneromooo-monero)
bcac101 daemon: fix a few issues reported by valgrind (moneromooo-monero)
a7e8174 tx_pool: fix serialization of new relayed data (moneromooo-monero)
601ad76 hardfork: fix mixup in indexing variable in get_voting_info (moneromooo-monero)
444e22f blockchain: remove unused timer (moneromooo-monero)
7edfdd8 blockchain: fix m_sync_counter uninitialized variable use (moneromooo-monero)
d97582c epee: use generate_random_bytes for new random uuids (moneromooo-monero)
17c7c9c epee: remove dodgy random code that nobody uses (moneromooo-monero)
A boost lock is used to determine whether more than one process
wants to access the database. The boost file_lock doesn't seem
to like locking directories, so we use an arbitrary file in it.
This allows to still run two daemons if they have different
database directories (ie, LMDB/BDB, different data directories).
In particular, ensure we check the status of RPC response structures,
as some functions will return success, but with a BUSY status, when
the daemon is not yet synced, and the response will not filled.
94b98fb tx_pool: do not accept txes not in a block if they timed out before (moneromooo-monero)
eadbdf3 tx_pool: fix use of invalidated iterator (moneromooo-monero)
3b1d7e0 Fix V1/V2 use of hard fork related parameters (moneromooo-monero)
This is intended to avoid cases where a timed out tx will be
re-relayed by another peer for which it has not timed out yet,
which would cause the tx to stay in the network's pool for a
long time (until all peers time it out before another one
tries to relay it again).
This improves blockchain reorganization time by allowing one of the more
expensive DB lookups when popping a block to not have to seek through a
long dup list in the "output_amounts" db. This is most noticeable for
HDDs.
See ffcf6bdb95
Data should be removed in the reverse order it was added.
This matches the order of removal in
blockchain_storage::pop_transaction_from_global_index.
See f11def012f
