Still haven't tracked down the exact steps to reproduce this problem, but
it happened again recently -- and starting a new emacs from the command
line solved it, i.e., my usual daemon-run emacs seems to be "crudding up"
somehow.

On Sat, Nov 4, 2017 at 1:05 PM, Lawrence Bottorff <borgauf@gmail.com> wrote:

> Trying the babel gnuplot the other day, I had trouble getting the results
> to C-c C-c to actually materialize in the place I'd given. Taking the
> example from the org-mode babel gnuplot page, (changing the location):
>
> #+begin_src gnuplot :exports both :file omnigraphics/omnig10.png
> reset
>
> set title "Putting it All Together"
>
> set xlabel "X"
> set xrange [-15:15]
> set xtics -15,2,15
>
>
> set ylabel "Y"
> set yrange [-15:15]
> set ytics -15,2,15
>
> f(x) = sqrt(x - 1) + 1
> g(x) = x**3
> h(x) = sqrt(-x - 1) + 1
>
> plot f(x) w lp lw 1, g(x) w lp lw 1, h(x) w l lw 1
> #+end_src
>
> The system was not creating the png in the given folder. Finally it
> occurred to me to look in another place where I'd previously been creating
> org-mode gnuplot pngs -- and there it was! I could do C-c C-c again and
> again and see new versions of the png being created each time in
> yesterday's folder, i.e., not the location given with :file. So the
> :file omnigraphics/omnig10.png was being blown off and org-mode internal
> had remembered where I'd previously created pngs and kept creating them
> there. I've tried to recreate this scenario with just the code snippet
> above -- fresh (non-daemon) emacs processes, changing the :file location
> each time -- and I've reproduced the problem sometimes. I've combed through
> my personal init els and it's definitely nothing I'm telling it. Strange. .
> . .
>
> LB
>