ob-core missing?

[fasiha]

When I try to use jmax, via emacs -q -l init.el, Emacs complains on the minibuffer that "Cannot open load file: ob-core" and indeed this file isn't in the repository. It appears to be something that ought to be autogenerated by running make autoloads [1] while in the org-mode directory, but nonetheless this file is never generated by make autoloads.

Running make also reports many errors with the same message.

Have you recently tried jmax on a fresh setup? Or is there just a step I'm missing?

[1] http://orgmode.org/manual/Installation.html#Installation

[jkitchin]

true enough. I thought I took the org-mode dir complely out, but I see that is not the case. I confirmed your error.

I put my org-mode and magit back into the repository. When I clone it and run emacs -q -l init.el, then it loads with no errors. hopefully it does for you too.

[fasiha]

After installing aspell for Windows [1] and customizing the hostname check at the top of jmax.el, it is working great.

I hope you don't mind me using this channel to ask an additional question (I am a failure at email). In your pycse.org book, there's a code block under "\ Defining functions in Python" which creates a matplotlib figure and then saves it as a png. In the org file, a link to the png file is embedded under a #+RESULTS: banner.

My question is, did org-mode auto-generate this reference to the png file by somehow knowing what savefig was calling the file, or did you manually add the link to the figure? I recognize that org-mode would have to be really smart to have done the former, but that's how the Sage and IPython browser-based notebooks work. But I don't think org-mode is doing this automatically because when I delete the results and the link to the image and evaluate the code block, the new results section is empty.

[1] http://aspell.net/win32/

[fasiha]

One small issue that might or might not be important: when I run emacs -q -l init.el, M-x org-mode, and check the org-mode version number, emacs says, Org-mode version N/A-fixup (N/A-fixup !!check installation!! @. This persists even after I run make in the org-mode subdirectory.

[jkitchin]

I had to write the [[./images/figure.png]] link. org-mode does not do that automatically. I agree, it would be great for that to automatically happen, but sometimes I save more than one figure in a code block. I don't know how you would do that automatically.

The hostname check is just there because I sync this directory on dropbox, and use it on many different windows/Linux computers.

[jkitchin]

Ok, I figured out what the org-version business was. the org-mode directory is under git vc from the org-repository, and it has a very large ignore file that made a lot of files not get committed to jmax. some of those did the version business. I think I added those, and now the version works the way it should I think. you will have to pull the new files in.

[fasiha]

About org-mode: in my usual .emacs.d directory, which is also a git repo, I include org-mode's repo as a git submodule, that's worked pretty well for me.

About the image link: I tried adding print "\b\b\b[[./images/figure.png]]" to the end of the code block, but of course it didn't fool emacs into thinking I had explicitly typed in those. There is I'm sure a good way to have the code block export a list of file names to the org file, I'll continue looking around.

[jkitchin]

It is possible to do this:

+BEGIN_SRC python :results output org raw

print '[[./images/figure.png]]' print '\n\n[[./images/figure.png]]'

+END_SRC

+RESULTS:

[[./images/figure.png]]

[[./images/figure.png]]

It does not always replace the output on multiple runs though.