This is just a toy to search flumedb backwards looking for a message and output its offset. I decided to build this after seeing Cel implementation in C in %5Pmvr9fijzjX/ouPf3pq+RfkvH5KX7RqYqN8RMltCDc=.sha256.
I've used his C implementation as a reference and included it on this repo so that I don't lose it.
The Racket version is much slower for obvious reasons since it is interpreted and not compiled like the C version but also because the Racket version parses the json of every message which the C version doesn't.
Cel's C version:
soapdog@FafiGo$ time ./flumefind ./log.offset %5Pmvr9fijzjX/ouPf3pq+RfkvH5KX7RqYqN8RMltCDc=.sha256 676860406 real 0m0.256s user 0m0.047s sys 0m0.141s
SoapDog's Racket version:
soapdog@FafiGo$ time racket ./flumefind.rkt --id %5Pmvr9fijzjX/ouPf3pq+RfkvH5KX7RqYqN8RMltCDc=.sha256 ./log.offset 676860406 real 0m15.977s user 0m11.422s sys 0m4.547s
Implementing this has been a great experience. It is the first time I work with files or bytes in Racket. I am confident now that I can write a parser that moves all the content from flume into SQLite in Racket now.
Thanks for everyone that supported me in %LTiHz/i2WHnNv9T2bwdETlxscwDcCtvTQ4gIuzFQud0=.sha256