Re: [BUG] org-element--cache: (org-agenda) Cached element is incorrect [9.6 (release_9.6-90-gf49ee9 @ /home/stefan/.emacs.d/contrib/org-mode/lisp/)]

[Stefan]

Hi,

On Thu, Dec 15, 2022, at 13:16, Ihor Radchenko wrote:
> 1. Check the value of `org-element--cache-change-tic' right after
>    opening an Org file 

first its 1, 

(when I start fresh and add 1 byte to the buffer the value is 13955)

> and the values of `after-change-functions' and
>    `after-change-functions' in Org buffer after you see the warning.

after starting the =t= agenda (which results in the warnings):

the value of org-element--cache-change-tic is still 1,

the value of after-change-functions is
`(org-element--cache-after-change jit-lock-after-change 
flyspell-after-change-function org-fold-core--fix-folded-region t 
ws-butler-after-change)`

(I disabled `ws-butler` because that sounded suspicious, but that did not 
resolve the warnings)


> 2. If you can, try to load Emacs with minimal Org-related settings.
>    Then, try to load agenda and let me know if you see the warning.

I did some manual bisecting of my config, and I noticed that the warnings in 
"Media.org" go away when I remove `(org-startup-align-all-tables t)` from my 
config.  

So I did some more digging, reduced 
my config to just contain `org-startup-align-all-tables` set to t, 
my agendas limited to Media.org, 
and removed most headings from Media.org.

With a minimal Media.org I get a "Org parser error",
see here for sample and backtrace:
https://gist.github.com/stefan2904/5a2b33d42a947f36402b548d9b186b67

(here the parser error turns into an incorrect error when I change the length 
of the URL ...)


With a very minimal Media.org I get a "Cached element is incorrect",
apparently because the table is not aligned?
see here for sample file and backtrace:
https://gist.github.com/stefan2904/688afdc2ebfeece8e0fa19a24d9ca321

Both errors go away when I set `(org-startup-align-all-tables nil)`.

>    It looks to me like you have something unusual in your Org setup or
>    some third-party package interfering.


Best,
 Stefan