the corner office : tech blog

Dusted off my old 'metalib' project. I'm reading and checksumming the contents of directory trees, and then importing the details into a db. Notes:

• The file reading was being done with python (os.walk and friends) and the uploading with java - after all manner of python encoding headaches, I reverted to java for file reading too. It wasn't a bad choice - it's given me a chance to play around with some Java 7 nio additions, all pretty cool.

• I got worried about hurting my MySQL server (hidden away on a VirtualBox VM), and set up a local H2 server. What an impressive bit of software, from the code to the website and documentation. I could do everything in MySQL mode, so my schemas worked out the box, and zippy as can be. My only gripe is that the DB is saved as a single big file, which is less useful for seeing table sizes. That, I think, I can live with.

• my machine can generate md5sums (using Guava) at a rate of about 45 MB/s. CPU usage hardly lifts, which tells me that CPUs have come a very long way, and hard drives haven't, since mucking around with hdparm more than a decade ago. Hrm.