[14:08] <chris2> abentley: hi. i
[14:08] <chris2> i'm curious if the code to http://code.aaronbentley.com/weavediff is still online somewhere?
[14:11] <abentley> chris2: It's been a very long time.  I'm sure whether I implemented it.
[14:12] <chris2> there is a link on your homepage, but broken :)
[14:12] <chris2> but if you dont find it, not problem
[14:12] <chris2> i wonder if its possible to store a new revision without extracting the last one first?
[14:16] <abentley> Sorry, I can't find easily.
[14:16] <chris2> ok, no problem
[14:16] <chris2> i'll try to do it myself :)
[14:21] <abentley> chris2: You don't have to extract the last version first.  As part of extracting the current version, you'll need to fill in holes in the current version with lines from the previous version.  As part of reconstructing those segments of the previous version, you'll need to fill in holes with an even earlier version, and so on until you've found all the lines that appear in the current version.
[14:23] <abentley> chris2: Best case, all lines in the current version are new, and you can stop right away.  Worst case, there's one line in the current version that was introduced in the very first version...
[14:34] <chris2> that's what i thought, but i think there are counterexamples where you need a real lcs?
[14:35] <chris2> if you try to insert changed lines into the first possible position, the weave can become bigger than it needs no?
[14:38] <chris2> need to contemplate about that :)