get_ordered_blocks() support #20
Conversation
It is now possible to get the block ordered by their height, to do so the unordered block that come early are stored untill they are needed then deleted. from block 0 to 150000 the maximum number of blocks in memory was around 800.
Block height is now filled by the get_ordered_blocks() Blockchain method.
|
It would seem that this method doesn't actually guarantee that each block will be a subsequent block from the last one. |
|
I've fixed this bug and updated the code to use Python3 in a PR to @LoudMurmur's branch: https://github.com/LoudMurmur/python-bitcoin-blockchain-parser/pull/1. @LoudMurmur you can review and merge, or @alecalve if you would like I can make a PR with these fixes to this repo directly. @alecalve, it would seem that there are three solutions to this
I've tested #17 and it seems to corrupt the |
|
I had the corruption issue indeed when trying it. I think using a more modern leveldb library would fix it, since it would also handle locking, which this PR doesn't. |
|
I was in email correspondence with the author of that fork and he has a hunch that that is correct. |
|
Fixed by #28 |
4 participants
Add support to get block ordered by their height.
If a block from the blockchain is too early (blockchain gives us block 75 while we are only at height 55 for example), then this block is stored in memory and at each iteration the code will check if the next block, according to height, is not already available in the "early_block_pool' before trying to make a disk access.
if the next block is available in the pool the disk acess won't be done, the block is yielded and removed from the early block pool.