Samuel Gélineau

I have added [a branch](https://github.com/Simspace/resource-pool-benchmarks/tree/pool-vs-ioref-vs-mvar-vs-tvar) which fixes the bug, but also switches the focus of the benchmark from the number of stripes to the "doing tons of retries in stm"...

Personally I would prefer if the error message was `expecting a, 'b', or 'd'`, because the user has three ways to try to fix the problem: they can fix the...

How about keeping track of both the error message at the latest-consumed position and at the rightmost-observed position? Here is a long explanation of why I think it's a good...

I'm not in a hurry, take your time!

Hmm, I shared what I have so far (the version with exactly 2 panes per tab) so that you could redirect my efforts in the right direction if needed, but...

Thanks! I don't promise to be a very active co-maintainer, as I already maintain too many projects, but precisely for that reason, I appreciate how important it is to have...

> Do you know of a terminal emulator on Linux that has this same functionality? I found one, [Konsole](https://konsole.kde.org). > Of course, the easiest way to do what you want...

> I imagine GTK provides some sort of widget to allow you to arbitrarily split a layout in half, but I'm not sure what would be the best widget to...

> There are probably also a few places in the code that make an assumption that there is a one-to-one mapping between Notebook tabs and Terminals My attempts so far...

Any particular reason why `assertInvariantTMState` is returning a list of errors instead of using a Validation monad, or even just Writer or ExceptT?