RAD II: Back to Ruby

[RAD stands for Ruby Ape Diaries, of which this is part II.] As of today the Ape is on Ruby, not JRuby; read on for the sad story and the Ape’s address.

Today I was going to test against my first APP end-point that claimed to offer proper TLS authentication, and I hit a brick wall, because JRuby doesn’t have that yet, and won’t for a while.

So I moaned and started switching back to Ruby. First, I chopped out the Jing validation (which I could, and may, revive with an external call). Also, the Ape now rests shakily on REXML’s parsing prowess.

Amazingly, the whole process took less than an hour. I’d put together a glue class called JREXML to stand in for Ruby’s REXML, and had got the API subtly wrong in a couple of places. Once that was working, it was gratifying that all the XPaths worked the same in Javaland and REXML.

Anyhow, the startup-time problem is gone, and I guess I can tell the world where the Ape lives: www.tbray.org/ape. It’s still a CGI and I may have to shut it down if abuse happens. If anyone wants to volunteer an APP end-point for the world to turn the Ape loose on, the world will be able to check out the Ape’s output.

Speaking of which, the past day or so I’ve focused on capturing the client/server dialogue and including it in the Ape’s report, and I think it’ll be a massively helpful debugging aid.

I’ll still report on the JRuby lessons, and when it has TLS I’ll switch back, because there are some other Java libraries that I’m thinking of.