after checking ob-clojure.el I determined that

#+begin_src clojure :results  pp
(pp/print-table [{:foo 10} {:foo  11}])
#+end_src

Mind pasting in the output you got? I've been occasionally frustrated when org tries to interpret everything as a table myself.

#+begin_src clojure :results  pp
(pp/print-table [{:foo 10} {:foo  11}])
#+end_src

#+RESULTS:
: 
: | :foo |
: |------|
: |   10 |
: |   11 |
: 
: 

I'm not sure what's up with : Also there is 1 extra blank line because it says "pp nil" once because of the hack

#+begin_src clojure :results  pp table
(pp/print-table [{:foo 10} {:foo  11}])
#+end_src

#+RESULTS:
| :foo |
|------|
|   10 |
|   11 |

Ohhh, nice! Gonna steal that one 😊😁

Do your clojure blocks run against a persistent process (REPL)? I'm curious if it would be possible to execute stuff like that with babashka.

proof of concept:

#+begin_src elisp
(defun org-babel-execute:bb (s params)
  "Execute `S` with babashka."
  (shell-command-to-string (format "bb -e '%s'" s)))

(add-to-list 'org-src-lang-modes '("bb" . clojure))
#+end_src

#+begin_src bb
  (+ 1 2 3)
#+end_src

#+RESULTS:
: 6

extending ob-clojure is maybe the right way though

Oh, cool! 😎

You can toggle-debug-on-error and check what’s the problematic line

let me see...

Debugger entered--Lisp error: (wrong-type-argument stringp (require . package))
  string-match("\\(\\`\\|/\\)text-mode\\(\\.so\\|\\.dylib\\|\\.elc\\|\\.el\\)?\\..." (require . package))
  load-history-filename-element("\\(\\`\\|/\\)text-mode\\(\\.so\\|\\.dylib\\|\\.elc\\|\\.el\\)?\\...")
  eval-after-load("text-mode" #f(compiled-function () #<bytecode 0x1fc49373f5612>))
  byte-code("\300\301\302\"\210\300\303\304\"\210\300\305\306\"\210\300\307\310\"\210\300\311\312\"\210\300\313\314\"\210\300\315\316\"\210\300\317\320\"\210\300\321\322\"\210\300\323\324\"\210..." [eval-after-load clojure-mode #f(compiled-function () #<bytecode 0x1fc4934087f72>) css-mode #f(compiled-function () #<bytecode 0x1fc4934084cd2>) erlang-mode #f(compiled-function () #<bytecode 0x1fc49333ef98e>) feature-mode #f(compiled-function () #<bytecode 0x1fc4933291926>) sgml-mode #f(compiled-function () #<bytecode 0x1fc4933291872>) rhtml-mode #f(compiled-function () #<bytecode 0x1fc4933291872>) nxhtml-mode #f(compiled-function () #<bytecode 0x1fc4933291872>) web-mode #f(compiled-function () #<bytecode 0x1fc4933290c0a>) js #f(compiled-function () #<bytecode 0x1fc4933290dd6>) js2-mode #f(compiled-function () #<bytecode 0x1fc4933290dd6>) #f(compiled-function () #<bytecode 0x1fc4933290ba2>) js3-mode #f(compiled-function () #<bytecode 0x1fc4933290dd6>) latex #f(compiled-function () #<bytecode 0x1fc4933290b6e>) nxml-mode #f(compiled-function () #<bytecode 0x1fc4933290a3a>) octave-mod #f(compiled-function () #<bytecode 0x1fc49332909d2>) octave #f(compiled-function () #<bytecode 0x1fc49332909d2>) python #f(compiled-function () #<bytecode -0xbcec82b8f601688>) python-mode #f(compiled-function () #<bytecode 0x1fc49332938b6>) ruby-mode #f(compiled-function () #<bytecode 0x1fc4934084a6a>) org #f(compiled-function () #<bytecode 0x1fc4933290102>) cc-mode #f(compiled-function () #<bytecode 0x1fc4934087cde>) "text-mode" #f(compiled-function () #<bytecode 0x1fc49373f5612>) cperl-mode #f(compiled-function () #<bytecode 0x1fc493408786e>) sml-mode #f(compiled-function () #<bytecode 0x1fc49373f56de>) enh-ruby-mode #f(compiled-function () #<bytecode 0x1fc49333eca12>) ...] 3)
  autoload-do-load((autoload "expand-region" "Increase selected region by semantic units.\n\nWith ..." t nil) er/expand-region)
  command-execute(er/expand-region)

It seems that require expected a string (as usual), but it got other thing. Maybe try to bisect your config. For example, try using emacs -q and loading only expand-region.

The way I read it the problem is with string-match, not require

Ah, yeah, maybe. It’s a bit hard to read 😅

bisecting the init.el is a good idea

hm, I don't even know how to require the package

ah I figured it out

M-x package-initialize

and sure enough it works with only expand-region loaded

so now I have to really bisect the init.el I guess?

I’d recommend that, yeap

ok so this is strange with emacs -q it works without -q but with an init.el that only contains (package-initialize) I get the error message

better yet: emacs -q: works emacs with empty init.el: doesn't work

shouldn't those two be identical?

yeah. strange. not sure what’s going on. maybe wipe out your elpa folder and reinstall all packages (you can backup the current emacs.d first, just in case)

Ok so after much digging I found that the offending line is this https://github.com/magnars/expand-region.el/blob/master/expand-region.el#L188

all other calls to eval-after-load take a symbol, but this one takes "text-mode" as a string

from what I can tell, that odd call (introduced in this pr https://github.com/magnars/expand-region.el/pull/213) seems to break with Emacs 28

Created a PR to fix the problem upstream https://github.com/magnars/expand-region.el/pull/275

Thanks for the help @U976F1AR2

cool, I’m glad you found the issue

I confess I found it surprising that more users didn’t report it (I suppose), though – AFAIK expand-region is widely used :thinking_face:

Well, expand-region doesn't accept bug reports anymore, so people don't have a venue where they can report issues

(But I may also be wrong about this, as I know very little about elisp)

> Well, expand-region doesn’t accept bug reports anymore, so people don’t have a venue where they can report issues Oh, got it.

Makes sense, yeah

I think it was supposed to be there, isn’t it?

yup you answered in the other thread