Fork me on GitHub
#calva
<
2019-06-17
>
pez07:06:27

This with that the repl window has problems with large data structures is pretty bad. I am not sure where it bogs down, but I suspect it is in the lexing and building up of the model.

slipset07:06:20

@pez/@lspector Cider has had problems with “large” values in the repl window as well.

slipset07:06:06

Don’t remember if it’s a problem pasting large values, but it has deffo been a problem when you have large values returned from nrepl.

slipset07:06:22

AFAIK there is some stuff in recent nrepls which handles ellisions.

slipset07:06:18

Oh, there can be some stuff in future nrepls which helps with this https://github.com/nrepl/nrepl/pull/106

pez08:06:41

Thanks, @slipset! I added an issue for this now, and included that nrepl PR as a clue to solving this: https://github.com/BetterThanTomorrow/calva/issues/228

👍 4
pez08:06:33

This actually makes me doubt the decision to go with the Webview, writing a new Clojure editor... It is such a big thing to take on. We should investigate how far we could take things using a regular editor window instead. It would bring other problems, but so much of the fundamentals of Clojure editing is already solved by VS Code and Calva... Hmmm... Anyway, it is not something I can or should jump to right now. Just something to keep considering.

pez11:06:39

@lspector , this is on another Mac i sometimes use. Funny that Apple seem so undecided on what it should be called.

👍 4
pez11:06:21

Also, evaluate to repl window now uses ctrl+alt modifier instead of just alt, so you shouldn't run into troubles with that dead key any more.

Felix Linker15:06:09

what's up peeps!

Felix Linker15:06:09

I have a question regarding the REPL window. Nested changes only seem to be imported when I do (use 'user :reload-all) whereas I'm used to do (use 'user :reload) when not in the REPL window. Is this a bug or am I missing something?

Felix Linker15:06:16

And are there any best practices for debugging? Like showing the stack trace? I can only see a part of that in the task-console-window that launched the REPL server but all the important frames aren't shown but hidden in the ... 39 more ... line at the end of the trace. In the REPL window *e is bound to the last exception that occured in that window and not the one that is shown in the task window.

pez15:06:24

What is the task window in this context? The “Calva says” output?

Felix Linker15:06:03

No, the window that launches when I jack-in

pez15:06:47

There are stack traces there? From something failing at startup?

grierson15:06:38

Is there a way to 'evalCurrentFormInREPLWindow' without the cursor jumping into the REPL?

Felix Linker15:06:44

From a server. But I just recognized that those might get logged to that window because of logger settings. Maybe, I'm able to redirect them somehow. But it'd still be nice to know if there are some kind of best practices for debugging with calva

pez16:06:41

@grierson , not at the moment. Can you tell a bit about why you would want that?

grierson16:06:01

@pez The output window works like that but doesn't have any colours or pretty print like the Calva repl.

grierson16:06:05

@pez also I like to eval each expr at a time. With the cursor jumping between windows I have to keep switching between the two.

pez16:06:29

Thanks. Can you file as an issue?

pez16:06:57

@linkerfelix if you do find some log settings that work better for you, please create a wiki page for debugging and write some tips there.

Felix Linker17:06:24

@pez I will have a look 🙂 Do you have some comments on the :reload vs. :reload-all subject? ☺️

pez18:06:44

I am not quite following that question, @linkerfelix. Have you tried the Calva refresh commands?

pez18:06:01

(I suspect they behave the same, but anyway.)