emacs-orgmode@gnu.org archives
 help / color / mirror / code / Atom feed
From: Ihor Radchenko <yantar92@gmail.com>
To: John Herrlin <jherrlin@gmail.com>
Cc: emacs-orgmode@gnu.org
Subject: Re: Bug: [patch] Fix org-babel-result-to-file never expanding links when babel is evaluated in indirect buffer [9.3.7 (release_9.3.7-728-g1efc4e @ /home/yantar92/.emacs.d/straight/build/org/)]
Date: Mon, 03 Aug 2020 06:50:34 +0800	[thread overview]
Message-ID: <87zh7cy85h.fsf@localhost> (raw)
In-Reply-To: <877duglulm.fsf@gmail.com>

> I don't have a real world scenario, just the one I made up below. I used
> that to play around with the function.

I see. FYI: I do change default-directory during src block execution in
my config. The following code makes src blocks execute inside attachment
folder by default:

(define-advice org-babel-execute-src-block (:filter-args (&optional args) set-detault-dir-to-org-attach-path)
  "Set working directory to the current entry's attach directory."
  (if (eq major-mode 'org-mode)
      (let* ((directory (file-name-as-directory (org-attach-dir 'create-if-none)))
	     (arg (car args))
             (info (cadr args))
             (params (org-babel-merge-params (nth 2 info) (caddr args)))
             (dir-param (alist-get :dir params)))
        (unless (and dir-param (or (equal (f-full default-directory) (f-full dir-param))
				   (f-absolute-p dir-param)))
          (setf (alist-get :dir params)
		(if dir-param
                    (f-join directory (alist-get :dir params))
                  directory)))
	(list arg info params))
    (list arg info params)))

That's actually how I found the reported bug.

Best,
Ihor

John Herrlin <jherrlin@gmail.com> writes:

> Ihor Radchenko <yantar92@gmail.com> writes:
>
>>> It's intended to change directory and dont clone base buffer values. I
>>> tried to accomplish "If the `default-directory' is different from the
>>> containing file's directory then expand relative links."
>>
>> Not sure if I understand your use-case. Can you explain why you need to
>> achieve such behaviour? There might be an easier way to get what you
>> want.
>
> I don't have a real world scenario, just the one I made up below. I used
> that to play around with the function.
>
> Best regards
> John
>
>>
>> Best,
>> Ihor
>>
>>
>> John Herrlin <jherrlin@gmail.com> writes:
>>
>>> It's intended to change directory and dont clone base buffer values. I
>>> tried to accomplish "If the `default-directory' is different from the
>>> containing file's directory then expand relative links." As I
>>> understands it that's when it inserts a relative link, otherwise it just
>>> `result'. I was looking from the indirect buffers point of view and from
>>> there the relative link is off in this case. But from the base buffers
>>> point of view everything works fine. I guess that it was my use case
>>> that is strange.
>>>
>>> Looks good!
>>>
>>> Best regards
>>> John
>>>
>>>>> In my case the result is a relative link, but if I follow the link I end
>>>>> up in the wrong place.
>>>>
>>>> Do you mean that the generated script.sh ends up in your home dir?
>>>> It is expected since default-directory in your buffer is "~/":
>>>>
>>>>>   (dired "~/")
>>>>>   (make-indirect-buffer "tmp.org" "tmp-indirect")
>>>>
>>>> Note that you called make-indirect-buffer without third argument.
>>>>
>>>> According to manual:
>>>>
>>>>> (make-indirect-buffer BASE-BUFFER NAME &optional CLONE)
>>>>> CLONE nil means the indirect buffer's state is reset to default values.
>>>>
>>>>>     Org first tries to generate the filename from the value of the
>>>>>     ‘file’ header argument and the directory specified using the
>>>>>     ‘output-dir’ header arguments.  If ‘output-dir’ is not specified,
>>>>>     Org assumes it is the current directory.
>>>>
>>>> No surprise you got your file in the default-directory (~/).
>>>>
>>>> If you use (make-indirect-buffer "tmp.org" "tmp-indirect" t) instead,
>>>> the script.sh link will be relative, as expected.
>>>>
>>>> Let me know if I miss something.
>>>>
>>>> Best,
>>>> Ihor
>>>>
>>>>
>>>>> Hey Ihor,
>>>>>
>>>>> Could not get the patch to work.
>>>>>
>>>>> I used this use case.
>>>>>
>>>>> (with-current-buffer
>>>>>   (find-file "/tmp/tmp.org")
>>>>>   (insert "* A\n** b\n** c\n\n   some text")
>>>>>   (dired "~/")
>>>>>   (make-indirect-buffer "tmp.org" "tmp-indirect")
>>>>>   (switch-to-buffer "tmp-indirect")
>>>>>   (goto-char (point-max))
>>>>>   (funcall 'org-mode)
>>>>>   (insert (concat "\n\n| default-directory | " default-directory " |\n"
>>>>>                   "| org-link-file-path-type | " (symbol-name org-link-file-path-type) " |\n"
>>>>>                   "| (expand-file-name default-directory) | " (expand-file-name default-directory) " |\n"
>>>>>                   "| (buffer-file-name (buffer-base-buffer)) | " (buffer-file-name (buffer-base-buffer)) " |\n"))
>>>>>   (org-table-align)
>>>>>   (insert "\n\n#+BEGIN_SRC shell :results file :file script.sh
>>>>>        echo \"#!/bin/bash\"
>>>>>        echo \"echo Hey\"\n#+END_SRC")
>>>>>   (org-babel-execute-src-block))
>>>>>
>>>>> In my case the result is a relative link, but if I follow the link I end
>>>>> up in the wrong place.
>>>>>
>>>>> src_emacs-lisp{emacs-version} {{{results(=26.3=)}}}
>>>>> src_emacs-lisp{org-version} {{{results(=9.3.7=)}}}
>>>>>
>>>>> Best regards
>>>>> John
>>>>>
>>>>>
>>>>> Ihor Radchenko <yantar92@gmail.com> writes:
>>>>>
>>>>>> Remember to cover the basics, that is, what you expected to happen and
>>>>>> what in fact did happen.  You don't know how to make a good report?  See
>>>>>>
>>>>>>      https://orgmode.org/manual/Feedback.html#Feedback
>>>>>>
>>>>>> Your bug report will be posted to the Org mailing list.
>>>>>> ------------------------------------------------------------------------
>>>>>>
>>>>>> I noticed that org-babel-result-to-file never expands file: links in
>>>>>> src block results when ran from indirect buffer and default-directory
>>>>>> is not same with base buffer file's directory. This is against the
>>>>>> docstring stating that
>>>>>>
>>>>>>> If the `default-directory' is different from the containing
>>>>>>> file's directory then expand relative links.
>>>>>>
>>>>>> The misbehavior happens because buffer-file-name variable is always
>>>>>> nil in indirect buffers. We need to call (file-name-directory
>>>>>> (buffer-file-name (buffer-base-buffer))) instead.
>>>>>>
>>>>>> The patch is attached.
>>>>>>
>>>>>> Best,
>>>>>> Ihor
>>>>>>
>>>>>> diff --git a/lisp/ob-core.el b/lisp/ob-core.el
>>>>>> index f86282f0f..d37eef5f9 100644
>>>>>> --- a/lisp/ob-core.el
>>>>>> +++ b/lisp/ob-core.el
>>>>>> @@ -2561,16 +2561,16 @@ If the `default-directory' is different from the containing
>>>>>>  file's directory then expand relative links."
>>>>>>    (when (stringp result)
>>>>>>      (let ((same-directory?
>>>>>> -	   (and buffer-file-name
>>>>>> +	   (and (buffer-file-name (buffer-base-buffer))
>>>>>>  		(not (string= (expand-file-name default-directory)
>>>>>> -			      (expand-file-name
>>>>>> -			       (file-name-directory buffer-file-name)))))))
>>>>>> +			    (expand-file-name
>>>>>> +			     (file-name-directory (buffer-file-name (buffer-base-buffer)))))))))
>>>>>>        (format "[[file:%s]%s]"
>>>>>> -	      (if (and default-directory buffer-file-name same-directory?)
>>>>>> +	      (if (and default-directory (buffer-file-name (buffer-base-buffer)) same-directory?)
>>>>>>  		  (if (eq org-link-file-path-type 'adaptive)
>>>>>>  		      (file-relative-name
>>>>>>  		       (expand-file-name result default-directory)
>>>>>> -		       (file-name-directory (buffer-file-name)))
>>>>>> +		       (file-name-directory (buffer-file-name (buffer-base-buffer))))
>>>>>>  		    (expand-file-name result default-directory))
>>>>>>  		result)
>>>>>>  	      (if description (concat "[" description "]") "")))))
>>>>>>
>>>>>> Emacs  : GNU Emacs 28.0.50 (build 4, x86_64-pc-linux-gnu, GTK+ Version 3.24.20, cairo version 1.16.0)
>>>>>>  of 2020-07-27
>>>>>> Package: Org mode version 9.3.7 (release_9.3.7-728-g1efc4e @ /home/yantar92/.emacs.d/straight/build/org/)

-- 
Ihor Radchenko,
PhD,
Center for Advancing Materials Performance from the Nanoscale (CAMP-nano)
State Key Laboratory for Mechanical Behavior of Materials, Xi'an Jiaotong University, Xi'an, China
Email: yantar92@gmail.com, ihor_radchenko@alumni.sutd.edu.sg


  reply	other threads:[~2020-08-02 22:51 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-07-31  8:49 Bug: [patch] Fix org-babel-result-to-file never expanding links when babel is evaluated in indirect buffer [9.3.7 (release_9.3.7-728-g1efc4e @ /home/yantar92/.emacs.d/straight/build/org/)] Ihor Radchenko
2020-07-31 17:24 ` John Herrlin
2020-07-31 23:32   ` Ihor Radchenko
2020-08-01 16:33     ` John Herrlin
2020-08-02  4:36       ` Ihor Radchenko
2020-08-02 19:24         ` John Herrlin
2020-08-02 22:50           ` Ihor Radchenko [this message]
2020-09-05 14:42 ` Bastien
2020-09-05 23:05   ` Ihor Radchenko

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

  List information: https://www.orgmode.org/

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=87zh7cy85h.fsf@localhost \
    --to=yantar92@gmail.com \
    --cc=emacs-orgmode@gnu.org \
    --cc=jherrlin@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
Code repositories for project(s) associated with this public inbox

	https://git.savannah.gnu.org/cgit/emacs/org-mode.git

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).