@kanwei Yes, there’s an issue about that, thanks.

@genekim I’’d be interested to know what you find works - scratch files can be weird, I have an outstanding issue to test them properly. I like manutter51's idea, I think.

@cfleming Scratchfiles work, with surprisingly little effort — was able to make a new Scratchfile, go into it, and do some forms — (+ 1 1), 1. No problems there. What was a little surprising was that it kept me in the same namespace as before, which was great. I thought I would have to make a new (ns xxx) or something... (That's actually a pleasant surprise. So easy!)

@cfleming @manutter51 — to be specific, what appeals to me about the "personal" directory is that I can work on different machines. Already, there have been times when I went home w/o checking everything into the repo, and couldn't work on something. The idea that there would be another bunch of work I could do that I might forget to check in somewhere seems... dangerous. 🙂

@genekim re: your initial comment: I often work inside of a comment block when I have scratch work.

I did the scratch file thing for a while, but I didn’t find it worth the effort.

I do use a dev dir where people can be much more loosie goosie.

But comment forms are super for exploring things in-place.

(Perhaps not directly answering your question, but thought you mind find my “there and back again” story useful 🙂)

I didn’t even know cmd-/ did form comments 😛

I use slurp/barf — I don’t think there’s a special command

bear in mind: #_ and comment are _slightly_ different. comment returns nil, #_ ignores at the reader.

this has gotten me before

@potetm — what a game changer! Working in a comment form is magical! I'm loving it. Thanks!!!

my pleasure 🙂

@genekim No, the IntelliJ “Comment with Block Comment” command should do that, but doesn’t (yet)

Also, one of the nice things about using comment is that you can use “Send top form to REPL” from within a comment block and it will send the top-level form within the comment block, not the comment itself.

So if you’re here:

(comment
  (+ 1 2|))

^^ that’s how I use it

it worked so naturally, I didn’t even notice it’s sending not-quite-the-top form

is there a way to automatically reorder alphabetically ns in :require ?

@cfleming @potetm Yes, I noticed that right away. For a couple of marvenlous hours last night, I worked within a source file comment block, sending forms to the REPL, and then copying code outside the comment block when done. Describing what was so different and awesome defies easy explanation, but I felt so much more productive vs. not working within a comment block. I'll try to write something up in the next week or two — it's not an understatement to say that it's one of my biggest aha moments of Clojure. Thank you!!!

PS: I think it has to do with the way it encourages documentation and the evolution of the code, as well as how it keeps codes around that actually calls the function — as opposed to burying the function inside of another function, inside a let block. Makes it much easier to change and test it later.

I have some videos of how I tend to work here: https://potetm.com/videos.html

I made those because I totes agree w/ @genekim, and I wanted to show other people what it’s like.

And I don’t really have words for it, so I hope you find some 🙂

I've created a REPL configuration and I'm trying to set -e "(require 'dev) (in-ns 'dev)" as its Parameters. However, when I start it, it turns into -e \"(require 'dev) (in-ns 'dev)\" and leads to Exception in thread "main" clojure.lang.LispReader$ReaderException: java.lang.RuntimeException: EOF while reading string Is there any way to prevent that seemingly unnecessary escaping of the double quotes?

@cfleming since a month or so everytime that I am in the debugger and try to realize a lazy sequence (with a watch expression), the debugger gets stuck in “Collecting data …“. If I afterwards remove the watch, and click the play button, the repl then gets stuck as well. Is there some kind of hidden setting that I screwed or is it a bug ?

@carocad I don’t think there’s a setting for that, it sounds like a bug. Do you have some repro steps, and I can try to reproduce it?

mmmm besides what I mentioned above … nothing really. It happens all the time when trying to get info on lazy sequences

if there is a way to give you a log I am happy to help

But you have a watch on the sequence itself too?

yes. For example, lets say that I have a function that returns a lazy seq like (map str something-else). Then in the debugger I add a watch + containing (vec my-lazy-result). I can see the watch in the debugger but it is stuck collecting data

Ok, thanks - I’ll try to reproduce that.