I’m about to release Pijul 0.10, but before I can do so, we need to decide what to do about liner/rustyline/… which apparently have been stalled for about a year.
After evaluating a number of alternatives, which involved writing bindings to C libraries, I think the most sane option seems to drop support for this at the moment, and wait until someone provides a clap-quality version of this, with full support for linux, windows and different variants of bsd.
I am in favor of this solution; the other possibility is to keep this dependency for now, release anyway, then decide for pijul-0.11 what to do. We are still unstable, so having an issue with one system target temporary does not bother me too much (but of course, it is not my target of choice, so it is easy for me to say)
Since pijul-0.8.3, we can write our patch name and description from an external editor, so we can say we are dropping the support because there is a usable alternative (yet, I never took the time to properly document how it works, and I can fix that if you wish).
Ok, I’ve actually learned about VT100, and I’ve implemented this, since this is the usual way I type my patch names. The crate is called line, is super small, works on windows and unix. I managed to avoid blinking lines, which is better than all alternatives I’ve tested (with the exception of readline).