RE: [O] FW: [RFC] Link-type for attachments, more attach options

[Gustav Wikström]

Hi!

org-attach-store-link-p with option t is supposed to store a link to the 
original location (i.e. the location the file was/is in before it was attached 
to the node. That was the default before I started working with attachments I 
believe. Haven't ever used that feature myself but the patch you provide would 
change the functionality which I don't think is correct. It would also not 
match the documentation any longer.

See the documentation for the customization parameter:

#+begin_src emacs-lisp
  (defcustom org-attach-store-link-p nil
    "Non-nil means store a link to a file when attaching it."
    :group 'org-attach
    :version "24.1"
    :type '(choice
            (const :tag "Don't store link" nil)
            (const :tag "Link to origin location" t)
            (const :tag "Link to the attach-dir location" attached)))
#+end_src

Regards
Gustav

> -----Original Message-----
> From: stardiviner <address@hidden>
> Sent: den 18 januari 2020 15:56
> To: Gustav Wikström <address@hidden>
> Cc: address@hidden; address@hidden
> Subject: Re: [O] FW: [RFC] Link-type for attachments, more attach options
> 
> 
> stardiviner <address@hidden> writes:
> 
> I finally figured out why the link always failed. Because it use wrong
> variable which is old filename path. I attached a patch.
> 
> > Gustav Wikström <address@hidden> writes:
> >
> >> Hi,
> >>
> >>> -----Original Message-----
> >>> From: stardiviner <address@hidden>
> >>> Sent: den 15 januari 2020 07:21
> >>> To: Gustav Wikström <address@hidden>
> >>> Cc: address@hidden; address@hidden
> >>> Subject: Re: [O] FW: [RFC] Link-type for attachments, more attach
> >>> options
> >>>
> >>> [...]
> >>>
> >>> >> I found when I set option ~(setq org-attach-store-link-p t)~.
> >>> >> Then attach a file, store file link with =[C-c C-l]=. The stored
> >>> >> link. I open this link got error "No such file: ....". I tested
> >>> >> this with minimal Emacs config. confirmed this problem.
> >>> >>
> >>> >
> >>> > I cannot reproduce this. In my try with a minimal Emacs (emacs -q)
> >>> > and
> >>> with only that single customization it works for me. I'm testing it
> >>> in linux. A wild guess.. Could it be that you used the move
> >>> operation instead of the copy operation when attaching the file?
> >>> >
> >>> > Regards
> >>> > Gustav
> >>>
> >>> Did you reproduce this issue with =emacs -q= ? That is a built-in
> >>> Org Mode version which does not contains the latest version =org-
> attach.el=.
> >>>
> >>> Here is my minimal Emacs config:
> >>>
> >>> [...]
> >>>
> >>> ;;==================================================================
> >>> ======
> >>> ======
> >>> ;;; Here is org-attach.el customization
> >>>
> >>> (require 'org-attach)
> >>>
> >>> ;; store link auto with `org-store-link' using `file:' link type or
> >>> `attachment:' link type.
> >>> (setq org-attach-store-link-p 'attached) (setq
> >>> org-attach-dir-relative t) (setq org-attach-preferred-new-method
> >>> 'ask) #+end_src
> >>>
> >>> #+begin_src sh :eval no
> >>> emacs -q -l '~/.config/emacs/minimal-init.el'
> >>> #+end_src
> >>
> >
> >> Hmm, in the first mail you said that you set org-attach-store-link-p
> >> to t, but in your config it says 'attached.
> >
> > Sorry about this.
> >
> >> I've tried with a minimal config as well (using emacs -q because I
> >> build the newest org mode version into the emacs
> >> folder) and can only reproduce your issue when using the attached
> >> option for org-attach-store-link-p and then inserting that link with
> >> C-c C-l /in another heading/. Pasting the link in another heading is
> >> expected to break since the attachment link is context dependent (i.e.
> requires an attachment folder).
> >> Makes sense? If I'm still misunderstanding your use-case, would you
> >> care to describe the steps to reproduce it more in detail?
> >
> > After updated commit, don't know why, but all links worked again. I'm
> > not good at expressing thanks, but you got all my thanks on this. :)
> >
> >> Regards
> >> Gustav
> >>
> 
> --
> [ stardiviner ]
>        I try to make every word tell the meaning what I want to express.
> 
>        Blog: https://stardiviner.github.io/
>        IRC(freenode): stardiviner, Matrix: stardiviner
>        GPG: F09F650D7D674819892591401B5DF1C95AE89AC3
>