eval-after-load: Symbol’s value as variable is void: org-brain-poly-hostmode

[DSMasterson]

I've added the following to my (non-Evil) Gnu Emacs 27.1 .emacs, but get the above error when I hit "C-c b" -- what's wrong?

(use-package org-brain
:after org polymode
:bind ("C-c b" . org-brain-visualize-mode)
:init
(setq org-brain-path org-directory)
;; For Evil users
(with-eval-after-load 'evil
(evil-set-initial-state 'org-brain-visualize-mode 'emacs))
:config
(add-hook 'before-save-hook #'org-brain-ensure-ids-in-buffer)
(push '("b" "Brain" plain (function org-brain-goto-end)
"* %i%?" :empty-lines 1)
org-capture-templates)
(setq org-brain-visualize-default-choices 'all)
(setq org-brain-title-max-length 12)
(setq org-brain-include-file-entries nil
org-brain-file-entries-use-title nil)
)

(use-package polymode
:config
(add-hook 'org-brain-visualize-mode-hook #'org-brain-polymode)
)

[Kungsgeten]

Hi!

It looks like there's something problematic with the way polymode is loaded. Have you installed the package (assuming you're using use-package-always-ensure)? You could try this:

(use-package polymode :defer t)

and then in the :config part of org-brain:

(add-hook 'org-brain-visualize-mode-hook #'org-brain-polymode)

That works for me. Perhaps I need to update the README.

[nanjigen]

I too am now getting this issue, was there a recent update?
Specifically:

org-brain-replace-links-with-visible-parts: Wrong type argument: stringp, nil
File mode specification error: (void-variable org-brain-poly-hostmode)

The first of those errors is in relation to me trying to visualize an entry from helm-brain.

Edit: Hmmm, it seems that there may be a deeper issue, as now .org files open as Fundamental instead of in org-mode...

P.S. I am a Doom user

[DSMasterson]

I'm going to have to look at this more. There's a lot to the README that I have to figure out. In the meantime:

1. I'm not sure org-brain-path is working right. I set it to a directory and org-brain adds "/brain" to it,
2. You have a README.org already. Why not, as part of your packaging/release process, org-publish that into a .texi file that could then be run through 'makeinfo' to produce an org-brain.info file. Then create a simple 'dir' file in the package and the package install process should take care of adding your info file to Emacs Info.

[Kungsgeten]

@nanjigen No, I don't think there's a recent update causing this. Seems to be something with your config if org files aren't opened in org-mode.

@DSMasterson The only function that changes org-brain-path is org-brain-switch-brain. The function org-brain-maybe-switch-brain could trigger that function. If you've changed org-brain-path after org-brain is loaded then the variable org-brain-data-file might have the wrong value, triggering org-brain-switch-brain when you enter the directory of the data file. Regarding the texi thing I have never tried that (and don't know how), but it seems like a good idea.

[DSMasterson]

Curiousity question: I see the LICENSE.org, README.org, and org-brain.el files in Github -- how do you package this for upload to MELPA? I thought there would be some sort of Makefile for that process.

[Kungsgeten]

Curiousity question: I see the LICENSE.org, README.org, and org-brain.el files in Github -- how do you package this for upload to MELPA? I thought there would be some sort of Makefile for that process.

I don't know how MELPA works, but when you upload to MELPA you only include a very small file containing information about the repo. Most information in the package manager is taken from the package description header in the main .el file. I don't think it cares about LICENSE and README, but I might be wrong. I've never written av Makefile, so I'm sure that is not required :) My "packaging/release process" is simply to push a new version to the Github repo. I think tagged releases are pushed onto MELPA Stable, but I'm not 100% here (I seldom tag releases).

[DSMasterson]

Interesting, Thanks.

[nanjigen]

@nanjigen No, I don't think there's a recent update causing this. Seems to be something with your config if org files aren't opened in org-mode.

I keep circling around this issue, first I will get this

File mode specification error: (void-variable org-brain-poly-hostmode)
org-brain-headline-at: Wrong type argument: stringp, nil

Then, if I go to the polymode section in org-brain.el and manual evaluate it, I get the following:

File mode specification error: (invalid-function define-hostmode)
org-brain-headline-at: Wrong type argument: stringp, nil

Then, moving around a visualise buffer causes a stream of display errors:

Error during redisplay: (poly-lock-function 77) signaled (void-variable matcher)
Error during redisplay: (poly-lock-function 83) signaled (void-variable matcher)
Error during redisplay: (poly-lock-function 85) signaled (void-variable matcher)

[jacereda]

I'm also facing this under NixOS and Nixpkgs/Darwin (without :ensure t).

It's weird, but I tried installing org-brain.el from a local checkout (as opposed to the version available in Melpa) and the error disappeared.

I deduced the version in melpa was lagging behind, but after diffing the checked out file and the Melpa version, the only difference is a ;; Package-Version: 20210108.1512 line. Well, that and the fact that the local checkout is being installed straight as $out/share/emacs/site-lisp/org-brain.el instead of the location chosen by the Nix packaging.

I must be lacking some sleep or something, because I can't think of any possible explanation.

[poulpoulsen]

Hello,
unfortunately here is the same error.
I'm on Org-mode 9.4, actual MELPA version of org-brain, Emacs 28, Fedora 34.
I load it with (require 'org-brain),
(require 'polymode)
No other configuration for polymode.
Is it necessary?

Any ideas how to solve it?
Without polymode org-brain works for me.

Regards
Poul

[swapneils]

I'm having a similar issue (on Windows); .org files open in org-mode (as they should) but org-brain gives this error in the use-package declaration and when trying to use any of its commands.

Based on the discussions in #320 I tried putting a no-compile setting:
;; --no-byte-compile: t; --
at the top of org-brain.el.

This fixed the initial problem and allows me to open org-brain-visualize, but causes a different issue when trying to move between nodes in org-brain (by using RET or by mouse-1).
Removing the compiled .elc files in the org-brain folder and restarting Emacs lets org-brain work for some time. The first time I tried it the org-brain.elc file (and the problem) reappeared for some reason, but I removed it again and it hasn't caused any problems today.

P.S.: Here's the error I'm getting after putting the above "no-byte-compile" warning on org-brain.el and then trying to move between entries in org-brain-visualize. Adding the warning to both org-brain.el and org-brain-autoloads.el didn't get rid of this either, so I'm not sure why it's popping up:
Lisp error: (void-variable entry)
(org-brain-visualize entry)
(lambda (_x) (org-brain-visualize entry))(#<marker (moves after insertion) at 83 in org-brain>)
button-activate(#<marker (moves after insertion) at 83 in org-brain> nil)
push-button(83)
funcall-interactively(push-button 83)
call-interactively(push-button nil nil)
command-execute(push-button)

I must be lacking some sleep or something, because I can't think of any possible explanation.

It seems the issue has to do with the byte-compiler's treatment of the code, though I don't know enough about it to find an exact solution.

[nanjigen]

Adding a (require 'polymode) to the top of org-brain.el fixes this for me.

[RecentlyRezzed]

I've had this problem too, when I did use straight.el without use-package. But it works for me in combination with use-package. I presume the minimal working example is:

;;; BEGIN bootstrap straight.el
(defvar bootstrap-version)
(let ((bootstrap-file
       (expand-file-name "straight/repos/straight.el/bootstrap.el" user-emacs-directory))
      (bootstrap-version 5))
  (unless (file-exists-p bootstrap-file)
    (with-current-buffer
        (url-retrieve-synchronously
         "https://raw.githubusercontent.com/raxod502/straight.el/develop/install.el"
         'silent 'inhibit-cookies)
      (goto-char (point-max))
      (eval-print-last-sexp)))
  (load bootstrap-file nil 'nomessage))
;;; END bootstrap straight.el

(straight-use-package 'use-package)

;; Configure use-package to use straight.el by default
(use-package straight
  :custom (straight-use-package-by-default t))

;; Allows you to edit entries directly from org-brain-visualize
(use-package polymode
  :defer t)

(use-package org-brain
  :config (add-hook 'org-brain-visualize-mode-hook #'org-brain-polymode))

[Kungsgeten]

I'm not sure if this would help (since I'm not having these issues) but perhaps someone could try:

In org-brain.el there's a line (with-eval-after-load "polymode". Try using "polymode-core" instead of "polymode".