Eric Schulte <eric.schu...@gmx.com> writes:
> Andreas Leha <andreas.l...@med.uni-goettingen.de> writes:
>
>> Hi all,
>>
>> this small patch adds :filelinkdescr to the list of possible header
>> arguments for babel source blocks.
>> Its intented functionality is to allow source blocks to return links to
>> files with description.
>>
>> Two small examples:
>>
>> ,----[ example1 ]
>> | #+begin_src R :results graphics :file test.png :filelinkdescr here
>> | plot(1:10)
>> | #+end_src
>> |
>> | #+results:
>> | [[file:test.png][here]]
>> `----
>>
>> ,----[ example2 ]
>> | #+begin_src R :results graphics :file test2.png :filelinkdescr
>> | plot(1:10)
>> | #+end_src
>> |
>> | #+results:
>> | [[file:test2.png][test2.png]]
>> `----
>>
>> If used with value ("here" in the example1) that value will become the
>> description. If used without value (as in example2), the description is
>> taken from either the :file header argument or the results of the code block
>> evaluation (in that order).
>>
>> Have not done much testing yet, though.
>>
>> Regards,
>> Andreas
>>
>
> Hi Andreas,
>
> I would like to apply this patch, however I have some questions about
> the use of the :file argument and the result itself to fill in the
> description if the :filelinkdescr header argument is left blank. Also,
> I would suggest that :file-desc is a better name for this header
> argument.
Hi Eric,
thanks for considering this patch. Feel free to change the term to
anything more appropriate.
>
> In my opinion using the :file argument as a description changes (or at
> least adds to) the meaning of the :file header argument. Is there a
> significant benefit to this behavior? If not I'd prefer that an empty
> :file-desc header argument have no effect.
My original use case was indeed making a code block to
produce [[file.pdf][file.pdf]] instead of [[file.pdf]] (see
http://permalink.gmane.org/gmane.emacs.orgmode/53941). So, if you
remove the possibility to specify an empty :file-desc, for me that would
mean to give the file name twice, as in
: #+begin_src R :results graphics :file test2.png :file-desc test2.png
which would be avoidable with my proposed behaviour.
I could live with that, though I'd still prefer not to have to type more
than necessary...
Regards,
Andreas
>
> Thanks for this patch, which I think adds nice functionality and which I
> would like to apply -- although I'll likely want to make the changes
> above after application.
>
> Cheers,
>
>>
>>
>> From 698570cb9938ee1f4dd1472cb13b32d08d108cbb Mon Sep 17 00:00:00 2001
>> From: Andreas Leha <andr...@lehas.net>
>> Date: Tue, 27 Mar 2012 10:33:32 +0200
>> Subject: [PATCH] new source block header argument :filelinkdescr
>>
>> ---
>> lisp/ob.el | 10 +++++++++-
>> 1 files changed, 9 insertions(+), 1 deletions(-)
>>
>> diff --git a/lisp/ob.el b/lisp/ob.el
>> index d27a394..0efef3e 100644
>> --- a/lisp/ob.el
>> +++ b/lisp/ob.el
>> @@ -400,6 +400,7 @@ then run `org-babel-pop-to-session'."
>> (eval . ((never query)))
>> (exports . ((code results both none)))
>> (file . :any)
>> + (filelinkdescr . :any)
>> (hlines . ((no yes)))
>> (mkdirp . ((yes no)))
>> (no-expand)
>> @@ -1833,7 +1834,14 @@ code ---- the results are extracted in the syntax of
>> the source
>> (progn
>> (setq result (org-babel-clean-text-properties result))
>> (when (member "file" result-params)
>> - (setq result (org-babel-result-to-file result))))
>> + (progn (when (assoc :filelinkdescr (nth 2 info))
>> + (let ((filelinkdescr (or (cdr (assoc :filelinkdescr (nth 2
>> info)))
>> + (cdr (assoc :file (nth 2 info)))
>> + result)))
>> + (when filelinkdescr
>> + (setq result
>> + (list result filelinkdescr)))))
>> + (setq result (org-babel-result-to-file result)))))
>> (unless (listp result) (setq result (format "%S" result))))
>> (if (and result-params (member "silent" result-params))
>> (progn
