You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
There are cases when you are interested only in a small fraction of the block and paying the cost of allocating all the block objects could impact performance. In other words, the cases where bitcoin_slices make sense.
What could help would be a BlockExtra version with Vec<u8> instead of a Block so that the downstream user could avoid instantiating a Block and use a bitcoin_slices::Visitor instead.
We have the byte size of the Block in BlockExtra, however in the binary format it appears after the block data, thus we cannot "skip" block parsing that is variable lenght.
Binary format version 1 could instead write the size before the Block so that if the user require the faster version could simply allocate a single Vec<u8> for the block .
steps:
release a 1.x version where version greater than 1 errors
make the PR, change the binary format and change crate version to 2.x, this version could decode both version 0 and 1, and encode the 2 versions according to a parameter, use version 0 by default so that version upgrade is opt-in
The text was updated successfully, but these errors were encountered:
At the moment
BlockExtra
contains abitcoin::Block
blocks_iterator/src/block_extra.rs
Line 17 in 73812b9
There are cases when you are interested only in a small fraction of the block and paying the cost of allocating all the block objects could impact performance. In other words, the cases where bitcoin_slices make sense.
What could help would be a
BlockExtra
version withVec<u8>
instead of aBlock
so that the downstream user could avoid instantiating aBlock
and use abitcoin_slices::Visitor
instead.We have the byte size of the
Block
inBlockExtra
, however in the binary format it appears after the block data, thus we cannot "skip" block parsing that is variable lenght.Binary format version 1 could instead write the size before the Block so that if the user require the faster version could simply allocate a single
Vec<u8>
for the block .steps:
The text was updated successfully, but these errors were encountered: