Change Details

This seems like it shouldn't have a big impact, but bridging to and from C++ has a measurable impact. Most of the compression time is spent in ToRustVec (44ms) and FromRustSlice (19ms), and just 45ms in ScriptCompression. While chronikreindexing blocks 22000 to 221000, the compression cost went from 162ms to 41ms (which is even below C++'s impl). Without this change, compression cost of chronikreindexing all blocks up to 300000 is 80 seconds, so after this change, this could be just 20s. For a reindex up to block 800000, this could shave away 10 minutes of indexing time. As a nice side benefit, this removes having to thread `fn_compress_script` through the code, which makes the code more readable and maintainable.

TBDhis seems like it shouldn't have a big impact, but bridging to and from C++ has a measurable impact. Most of the compression time is spent in ToRustVec (44ms) and FromRustSlice (19ms), and just 45ms in ScriptCompression. While chronikreindexing blocks 22000 to 221000, the compression cost went from 162ms to 41ms (which is even below C++'s impl). Without this change, compression cost of chronikreindexing all blocks up to 300000 is 80 seconds, so after this change, this could be just 20s. For a reindex up to block 800000, this could shave away 10 minutes of indexing time. As a nice side benefit, this removes having to thread `fn_compress_script` through the code, which makes the code more readable and maintainable.