Re: saving registry is fixed, but gnorb still can't gnorb-org-view

[hobbes@poukram.net]

Eric Abrahamsen <eric@ericabrahamsen.net> writes:

>> In your case, there's a link that has no message ID part. You could take
>> a look at the Org tree in question and see which it is, and if you know
>> how that link might have been created (or mangled).
>
> Ooh, I might be wrong -- I got bitten by Org's changing link-related
> regular expressions, and fixed some code for the new regexps, and
> apparently wasn't careful enough with the fix. I'm assuming in your
> subtree there's a "plain" link to a Gnus message, ie one that isn't
> wrapped in brackets. Would you eval the function below and see if it
> fixes the problem?
>
> (defun gnorb-scan-links (bound &rest types)
>   (unless (= (point) bound)
>     (let ((alist (mapcar #'list (copy-sequence types))))
>       (while (re-search-forward org-link-any-re bound t)
> 	(pcase-let* ((`(,type ,link) (split-string
> 				      (or
> 				       (match-string-no-properties 2)
> 				       (match-string-no-properties 0))
> 				      ":"))
> 		     (sym (intern-soft type)))
> 	  (when (and sym (memq sym types))
> 	    (push link (alist-get sym alist)))))
>       alist)))

So, after some more experimenting: it's indeed linked to the gnus links
in the org tree.

But all links have brackets, and all have a message id, and I can follow
them. Evaluating the defun didn't help.

If I remove the links, C-c v works like it should, so it's clearly
linked to those links...

Thanks,
-- 
Rémi