• 11652975 repl: add readline-persistent-history option

Ok, I've just pushed a commit that adds a --readline-persistent-history option that you can set to true or false. Maybe before I go to bed I'll look at what it would take to add expeditor support. I don't think it will be very hard.

• 1f1d63f0 add first draft of a repl with expeditor support

I also made a branch with an experimental repl-expeditor file. So far when it works it works better than with readline (eg. with some highlighting, reasonable auto-indentation, etc), but it has some weird issues. For example, I'm not confident it is any better at not borking the racket-prefs.rkt file.

One thing I was wondering about while looking at the code is why save-readline-history! exists as separate from the save-history in ‘pread.rkt’ (based on what winds up in the file .racket/racket-prefs.rktd, that seems to be what is used to save the history for the racket REPL, in which the issue with ffi-wrapper:history_get doesn’t seem to occur).

I wrote a different function to save it with a different key, so that rash-repl and regular racket repl history don't get mixed up. I think I've had issues both ways. I'm not sure what the core problem is.

I wrote a different function to save it with a different key, so that rash-repl and regular racket repl history don't get mixed up.

Ah right, I thought I was just missing something about how the storage location for the history is parameterized in the library, but it really is just hardcoded then.

Thanks for adding the option, seems to work fine now for me with that disabled!

In reply to @yasuakikudo:matrix.org
I am really thinking we should combine technology and democratic community building. They can be positioned to support each other!

In reply to @nojyuum:matrix.org
I dig bash and racket. bash is a huge help for doing runbooks, but it is very limited in all sorts of ways, it is showing its age, in a way. and the rash approach seems well-positioned for longevity. plus, it seems super intuitive to me, compared to other super-bashes or bash replacements.

In reply to @ormaaj:matrix.org
i need to look into designing a decent mechanism for interactive use. I write most nontrivial scripting code with rash, scsh-process, or ksh93, but do the interactive entry with bash or xrepl, which is a headache.

In reply to @yasuakikudo:matrix.org
And to bootstrap this, interpreters and shells are the first step?

In reply to @willghatch:matrix.org
In other news, I'm going to give you all unsolicited insight into my personal life: I just had a son the other day! Being a dad is cool.

In reply to @willghatch:matrix.org
I agree that technology and democratic community building can be good together, and each probably needs the other to be done well. But I can't say that I know much about how to go about building them together or what the first steps are. I think free software is important but not alone sufficient. While I have a lot of thoughts about software, politics, their intersection, and what's needed for both to be good, I think many of those thoughts (especially when they venture into politics) are... well let's say I'm uncertain and doubtful of even my own opinions with regard to politics. Anyway, long story short, I don't know :)