hamsterdb: the current status
I decided to postpone the port to big-endian architectures, because I lack a decent machine. I have access to a PowerPC system at work, but it’s old, awfully slow and the hardware has some issues. If anybody wants to use hamsterdb on big-endian machines, I can tell you that the in-memory-databases work flawlessly, but at least one test fails with file-based databases.
Over the last weeks I did not make as much progress as I wanted, but I managed to create over 70 unittests, and refactored some parts of the code. On my TODO list there’s one more important item: rewrite the freelist and replace it with a bitmap. This will not only be more efficient in regards to performance, but it should also keep the database files smaller, because file fragmentation is avoided.
I do not know when the next release will be available, but I hope that I’m done in about four to six weeks.