JDPFu.com 2025

Ok, so back to the reason for this entry. The current version of swish++ indexes my entire site in under 30 seconds. The prior version took 10 minutes or so. A major speed improvement to say the least. I’m doing local file indexing, not going over HTTP.

According to the Feature List, they have

• Lightning-fast searching
• Use the same mmap(2) technique used for indexing and used again for searching.
• The generated index file is written to disk such that it can be mmap’ed back into memory and
• binary searched immediately, with no parsing of the data, also in O(log n) time.

Fantastic!

I had to change some of my search command options – forcing the results separator to be a TAB character ‘\t’. At first, I searched for a config file setting – no joy. Then I checked command line options and didn’t see any. Into the code. It is C++ … soon, I was modifying the ResultSeparator.h file and replacing the default " " into “\t”. Recompile, install. Update the cgi script that is the front-end to the program so the split() uses a tab. Done. Reindex the entire site and test. Working. Done. 20 minutes of effort.

F/LOSS Rocks. Just try to do that with closed source software. You can’t. If you are trying to run a company, how much is that worth to you? Do you really want to be held hostage by proprietary, closed source software? I’ve seen trivial changes to code estimated at over $100K by software vendors AND I’ve seen large companies pay it. Seems like blackmail to me. Then add in the 15%/yr maintenance charges to jeep updates coming. Nice work if you can get it.