wren romano

> I could also add a change that allows to add new VCS. Then you could do something like this: > > ```viml > let g:signify_vcs_cmds = { 'sd': ......

@treeowl if we do accept this benchmark (whether in this repo or in a separate one), it may help for comparing the current big-endian patricia trie based `IntMap` with an...

If using new classes in lieu of `Enum`, another option to consider is the classes from [prelude-safeenum](http://hackage.haskell.org/package/prelude-safeenum-0.1.1.2/docs/Prelude-SafeEnum.html). (I'm not saying they're necessarily the right idea here; just another option to...

Note that `HashMap` does have different performance tradeoffs compared to `IntMap` just due to the internal representation of the tree. Which is why I'm thinking of adding a non-hashed version...

@zaxtax how did you make the ticky boxes?

@cartazio I believe it's articulated in the paper where pseq was first introduced. AFAICR: seq ensures that the left argument is forced some time before the whole seq expression returns...

Fwiw, this can only (potentially) affect functions that delete several values (and not in a delete-a-whole-subtrie way). If only a single value is deleted, then while there may be iterated...

I was about to say that `text-trie` isn't mine, but then I saw the ping

Note to self: `runRW#` appears to have been added in base-4.10.0.0 / ghc-8.2.1, and the kinding of the result type was generalized in base-4.13.0.0 / ghc-8.8.1

Sorry for the long delay. I'll review this in more detail shortly