Michael

@Nitemice Nice debugging there! :) Change 1 is very interesting, and I'd like to learn more about this—is this Windows specific, is there something else going on? I wonder if...

@Nitemice Digging some more, I've noticed the pattern of the keys here seem to be: `PREFIX {SPACER} SUFFIX` where PREFIX can be something like: - `!pl#add#` - `!pl#changes#` - `!cit#cit#`...

@Nitemice FYI, pushed a new version based on our current understanding.

> I think I've found a simpler solution that shouldn't slow down the runtime much if at all. What if we modify the `bytestring_less_or_equal` function so that space and group-separator...

Thank you, @Nitemice, this is very helpful. I now figured out the spacer symbol. It encodes the bytelength of the upcoming data. The same pattern seems to be used for...

Ok, I think I figured out the comparator approximately too now. It's essentially alphanumeric, but ignores the varint {length} sections. Pushed new version. Leaving in slow_search as a backup in...

@Nitemice Can you confirm that this works on your end?

> I wonder if the header in base 2 should be something like `01234567890123456789....`? so bit offset instead of byte offset? Seeing the bit offset address could be nice, especially...

> Would it look strange to show binary in 8 bit groups somehow? 🤔 I checked HexFiend. It has an option called "Byte Grouping".  Similarly, it also has "line...

Some options. I like D most.