I try run lumosql benchmarks, but get some error like this: ```sh ./tool/get-lumo-sources /opt/osx/lumosql/sources test 3.33.0 Cannot load VCS(git): Can't locate Git.pm in @INC (you may need to install the Git module) (@INC contains: /usr/local/bin/../lib /Library/Perl/5.18/darwin-thread-multi-2level /Library/Perl/5.18 /Network/Library/Perl/5.18/darwin-thread-multi-2level /Network/Library/Perl/5.18 /Library/Perl/Updates/5.18.4/darwin-thread-multi-2level /Library/Perl/Updates/5.18.4 /System/Library/Perl/5.18/darwin-thread-multi-2level /System/Library/Perl/5.18 /System/Library/Perl/Extras/5.18/darwin-thread-multi-2level /System/Library/Perl/Extras/5.18 .) at /Library/Perl/5.18/NotFork/VCS/Git.pm line 13, <$fh> line 10. BEGIN failed--compilation aborted at /Library/Perl/5.18/NotFork/VCS/Git.pm line 13, <$fh> line 10. Compilation failed in require at (eval 6) line 2, <$fh> line 10. make: *** [build/3.33.0] Error 1 ``` I already run this: ```sh perl -MCPAN -e shell install Git::Repository ``` > No information about the logical location of this error, eg what row(s) it affects. The application knows nothing about how rows map to pages. All the application knows is that SQLITE_IOERR_DATA was returned during a read operation. That's a lot better than silent corruption, but also not as helpful as it could be. This is very true. I try get the page information for each transition, to know if they are conflict, to use with the Raft pipelining optimistic(still a work at early stage). > Brittle implementation due to requirements of the file format. The "bytes of reserved space on each page" value at offset 20 in the SQLite database header must be exactly 8. What if we want a better or different checksum? I consider checksum as a reference implementation, I try build a VFS 64 byte reserved for encryption.