I cannot render any notebook on that branch exec-state

When doing clerk/show! I get always:

*e
;; => #error {
 :cause "Cannot invoke \"clojure.lang.IObj.withMeta(clojure.lang.IPersistentMap)\" because \"x\" is null"
 :via
 [{:type java.lang.NullPointerException
   :message "Cannot invoke \"clojure.lang.IObj.withMeta(clojure.lang.IPersistentMap)\" because \"x\" is null"
   :at [clojure.core$with_meta__5485 invokeStatic "core.clj" 220]}]
 :trace
 [[clojure.core$with_meta__5485 invokeStatic "core.clj" 220]
  [clojure.core$vary_meta invokeStatic "core.clj" 677]
  [clojure.core$vary_meta doInvoke "core.clj" 677]
  [clojure.lang.RestFn applyTo "RestFn.java" 142]
  [clojure.lang.Atom swap "Atom.java" 79]
  [clojure.core$swap_BANG_ invokeStatic "core.clj" 2372]
  [clojure.core$swap_BANG_ doInvoke "core.clj" 2362]
  [clojure.lang.RestFn invoke "RestFn.java" 497]
  [nextjournal.clerk.webserver$set_status_BANG_ invokeStatic "webserver.clj" 192]
  [nextjournal.clerk.webserver$set_status_BANG_ invoke "webserver.clj" 191]
  [nextjournal.clerk$show_BANG_ invokeStatic "clerk.clj" 37]
  [nextjournal.clerk$show_BANG_ invoke "clerk.clj" 22]
  [example$eval60102 invokeStatic "NO_SOURCE_FILE" 9]

@U7CAHM72M should be fixed with https://github.com/nextjournal/clerk/pull/429/commits/b2cb2144a89ab0e6c69a25ca5cc2d38f0c5672d9, thanks for trying it!

1. i think this is looking great! 2. I haven't come across any weird behavior / potential bugs 3. really appreciate knowing that "something is actually happening right now". When running slow computations, i've often been wondering if something is actually happening, or if i've done something wrong and made something crash. Clerk feels snappier with this somehow. (or perhaps you've made clerk faster in other ways!) 4. also really appreciate that the execution progress indicator isn't taking up tons of screen estate. I think I'd appreciate some visual signifier that "the evaluation has come this far in space" as well. I'm looking at the form -- if I can easily see that "clerk is evaluating this form right now", I don't have to break my attention and read a var name / read a form aiming my attention back to the form. (i tried {:sha "b2cb2144a89ab0e6c69a25ca5cc2d38f0c5672d9"})

@U3X7174KS thanks for the feedback, great to hear it feels snappier with this! Indicating the executing cell in the doc will come in a follow-up, currently we only refresh the doc after the evaluation completed so the cell might not be there, yet. But yes, agree that that’ll be good to have, just want to make sure it doesn’t become slower when we do this.

@U5H74UNSF i did stumble over something weird now, still on {:sha "b2cb2144a89ab0e6c69a25ca5cc2d38f0c5672d9"}: I expected to see :abc, not #viewer-eval (keyword ":abc").

clerk makes it really easy to reproduce things 😁

this looks weird too, expecting :abc, not ::abc. Edit: my fault -- (keyword ":abc") is indeed ::keyword!

Am I doing something wrong ? For me it does not work at all:

Failed to load module script: Expected a JavaScript module script but the server responded with a MIME type of "". Strict MIME type checking is enforced for module scripts per HTML spec.
(index):420 Uncaught ReferenceError: nextjournal is not defined
    at (index):420:515
(anonymous) @ (index):420

I just run it from branch exec-state via bb dev :browse\? true and then clerk/show! the exec_status.cljnotebook

Firefox is more clear on error message:

Loading module from "" was blocked because of a disallowed MIME type ("").

> I just run it from branch exec-state via bb dev :browse\? true and then clerk/show! the exec_status.cljnotebook I'm not running Clerk from the Clerk repo -- I'm using io.github.nextjournal/clerk {:sha "b2cb2144a89ab0e6c69a25ca5cc2d38f0c5672d9"} from my https://github.com/teodorlu/lab/blob/a4336c561426947e273359bb9fdb2ba924a496ae/deps.edn#L5-L12. I find that easier.

doesn’t support it, yet. Issue welcome, think we can fix this shortly.

https://github.com/nextjournal/clerk/issues/430 …thanks! 😀

@UFJ5R302F thank you!

I suppose the nice thing to do would be to turn that into an example in clerk.garden

what about https://scicloj.github.io/tablecloth/index.html#Printing ?

Is native printing not enough?

other than me completely forgetting about all the wonderful things you've built? No, not really 😄

I'd like to be able to file off the table name (or make it prettier) as this is for delivery to clients

mostly I was after stupid tc/markdown tricks

look ma, I'm famous https://github.clerk.garden/otfrom/md-tc-clerk-example/commit/91f0306de4f95f7530610ef5d797b8dfd90da4cf/

@U1EP3BZ3Q should p/dataset->str be wrapped in a println by default?

(defn print-dataset
  "Prints dataset into console. For options see
  tech.v3.dataset.print/dataset-data->str
"
  ([ds] (println (p/dataset->str ds)))
  ([ds options] (println (p/dataset->str ds options))))

@U1EP3BZ3Q now updated to use dataset->str https://github.clerk.garden/otfrom/md-tc-clerk-example/commit/5a658b87eefb3548ca3e61384d4fbddeb7779cff/

which has nicer formatting except for the dataset name and shape (which is good for me but not for clients), but I could take that off by chopping off the first bit of the string

Actually it's produced by TMD. Yeah, remove two first lines by hand, there is no option to skip it.

done: https://github.clerk.garden/otfrom/md-tc-clerk-example/commit/3659b98c0347bc172ec4b2368122e792b0958e96/

Looks great! 👏

@U1EP3BZ3Q thx 🙂

@U1EP3BZ3Q is this a bug in tc or is there a reason for the side effect and returning nil here? https://clojurians.slack.com/archives/C035GRLJEP8/p1678380031309319?thread_ts=1678376426.783319&cid=C035GRLJEP8

It's a bug for sure. PR?

hmm... doesn't seem to be like that in latest github. Now I'm confused

@U1EP3BZ3Q I'll dig, I think I can actually fix this one. 🙂

found it

https://github.com/scicloj/tablecloth/pull/98 💥

@U0525KG62 I was too quick (heading home and it was quick answer without much thinking 😕, my bad, sorry for that ). The idea behind print-dataset is actual printing 🙂 I though it might return a dataset itself, not a string representation. To get a string you can call tc/dataset->str which does exactly what you need - a formatted string.

I think we can enhance dataset->str to have an option of stripping first to line (ie. dataset name and empty line). What do you think?

that all sounds good. I'm easy either way on the stripping off of the dataset name as it is easy enough to do

Hi @U0525KG62 if markdown is intended for prose, maybe. But you can provide your own style. The source of this notebook shows how to customise markdown rendering https://snapshots.nextjournal.com/clerk/build/2356ac2e3a2f1063b65e84ef8019595d7bda2557/index.html#/notebooks/viewers/custom_markdown.clj

clerk/table could be more suited for the job of transforming data and strings are rendered sans by default https://snapshots.nextjournal.com/clerk/build/2356ac2e3a2f1063b65e84ef8019595d7bda2557/index.html#/notebooks/viewers/table.clj

Where can I find out more about add-child-viewers that looks like it might solve some alignment and formatting problems for me

> looks like it might solve some alignment and formatting problems You might have a look at Clerk default markdown viewers https://github.com/nextjournal/clerk/blob/2356ac2e3a2f1063b65e84ef8019595d7bda2557/src/nextjournal/clerk/viewer.cljc#L662-L731 and override them (by name) in the style of https://github.com/nextjournal/clerk/blob/2356ac2e3a2f1063b65e84ef8019595d7bda2557/notebooks/viewers/custom_markdown.clj#L18-L25

> add-child-viewers is mentioned in the book:tm: in the context of tables, but it’s the same principle https://book.clerk.vision/#transform