On Thu, 14 Aug 2014 05:57:17 +0000, Carl Sturtivant wrote: > This is the idea I mean. > http://citeseer.ist.psu.edu/viewdoc/download? doi=10.1.1.14.9450&rep=rep1&type=pdf > Here's a C++ implementation supported I think by gcc. > http://www.sgi.com/tech/stl/Rope.html > > Is there a D implementation of a rope out there somewhere? > > How is unicode handled? Just by using dchar?
I was working on a D Rope implementation years ago (now lost to time), when I came across this article and ditched it: http://scienceblogs.com/goodmath/2009/02/18/gap-buffers-or-why-bother- with-1/ The short version is that while Ropes are interesting theoretically, they have a bad implementation complexity compared to much simpler structures which give more than adequate results for most tasks.
