LJ hacking; may not make sense even to most geeks

And this, of course, is the problem with hacking a system in ways not anticipated by the architects: the new format of voicepost entries breaks my custom S2 layer. I had my data-munging hacks as well isolated as I could, and I backed off of them in a way that should have restored the most vanilla LJ style, but no luck. Creating a brand new child of the same parent fixes the problem, but loses my layout tweaks, colors, comment hack, auto-cutting long posts, and the admittedly scattershot quiz-blocker. I don't know whether it will be easier to debug the error in my current style or recode that stuff starting with the updated layer code. Either way it's kind of annoying; I appreciate the high degree of flexibility LJ allows me, but given that they let me do so much I wish they wouldn't toss in those few major hurdles that make S2 development such a pain.

Hmm. I could just add voiceposts to my set of widgets to be blocked... but it's probably better ultimately to do a clean rework.