Thank you. Good to know there is an active spam detector.
New users can’t post more than 2 links. Makes sense.
Here is a translation table to bypass that limit for this info:
git.nixo.xyz == h t t p s : / / g i t . n i x o . x y z / n i x o / o b - j u l i a
gjkerns == h t t p s : / / g i t h u b . c o m / g j k e r n s / o b - j u l i a / b l o b / m a s t e r / o b - j u l i a - d o c . o r g # r e s u l t s - v a l u e
For what it’s worth, I took quite a bit of time trying to install the
“ob-julia” inside git.nixo.xyz but ran into a conflict whereby there
are two ob-julia.el’s in the system if you do that, and that is how I
discovered that gjkerns was already installed into my setup!
The one provided by the org-plus-contrib package is loaded first, and
the ob-julia.el from inside git.nixo.xyz never got found in the load
path.
The one coming in by way of org-plus-contrib looks to be the same
“old” package from looking at the header in its ob-julia.el file, at
.emacs.d/elpa/org-plus-contrib-20200914/ob-julia.el, since it contains
this at the top:
;; See https://github.com/gjkerns/ob-julia/blob/master/ob-julia-doc.org
;; for detailed instructions on how to get started. The git repository
;; contains more documentation: git://github.com/gjkerns/ob-julia.git
It could be argued that the one coming from org-plus-contrib should be
removed, and replaced with the one from git.nixo.xyz, but I cannot
tell if the latter is better since I cannot give up on using the
org-plus-contrib package since it contains things I rely upon already
in my setup that I cannot disturb just to experiment with Julia+Org
integration.
Testing the gjkerns version of ob-julia.el shows that using :results output
works:
#+NAME: with-results-output
#+BEGIN_SRC julia :results output
rand(2,3)
#+END_SRC
#+RESULTS: with-results-output
: "somefile.org"
: 2×3 Array{Float64,2}:
: 0.471239 0.297008 0.663586
: 0.0963117 0.585604 0.642755
but using :results value
this gives no output:
#+NAME: with-results-value
#+BEGIN_SRC julia :results value
rand(2,3)
#+END_SRC
#+RESULTS: with-results-value
So, as found before, the one at gjkerns is broken indeed, but this is
proof of it.