Igor van den Hoven

I've got it disabled since it makes `#config mouse info` obnoxious to use. `#showme \e[?1003h\` and `#showme \e[?1003l\` work though as an alternative to #config. ENTER/LEAVE might be possible, I'll...

Thanks, where would I find the documentation for those?

Thanks, much appreciated!

I know it took a bit, but at last I've added a precompiler check for 64-bit long doubles, and I've documented the ambiguity of long doubles in quadsort.h and fluxsort.h.

@mlochbaum I noticed you had an interest in Rust, and I came across this Rust port recently. https://github.com/google/crumsort-rs Might be interesting to bench to see how it performs. The main...

I don't quite get the Rust hype myself, probably because I love writing unsafe code. Makes sense that parallel only does well on larger distributions. It's used by https://github.com/google/forma, so...

Thanks, tintin is generally not compiled with cpp though, so shouldn't it be as following? ``` pcre-config --cflags pcre-config --libs ./configure CFLAGS=-I/usr/local/include LDFLAGS=-L/usr/local/lib make CFLAGS=-I/usr/local/include LDFLAGS=-L/usr/local/lib make install ```

I'm quite sure Timsort, at least originally, performs a copy back. The performance loss is minimal. I've updated the readme to "Most textbook mergesort examples merge two blocks to swap...

I don't think modulating by 256 is enough for quadsort to detect it as non-random data, or otherwise benefit. Modulating by 16 should trigger galloping mode and some skips. I...

I'm aware, the performance you are reporting for mod 256 is quite unusual for a merge. Gains from simple branched merges are pretty limited generally. Either you're doing something quite...