We shouldn't always build our documentation afresh. Indeed, if you run `./cleanall` and `./make` then OpenCOR will build fine. However, if you rerun `./make` straight after that, you will be told that `[OpenCOR]/ext/doc/user` (or `[OpenCOR]/ext/doc/developer`, depending on which one gets built first) doesn't appear to contain a CMakeLists.txt file. So, instead, we have to update our `./clean` script to always clean our `[OpenCOR]/ext/doc` folder.
Simulation Experiment view: make sure that curves are plotted with all of their original settings when reloading a SED-ML file (#2176). This issue was that we were tracking the original settings using a graph's address as a key. However, when reloading a SED-ML file, that address changes since a graph gets removed and then readded.
Reverting part of commit 13a0ef1. Indeed, if everything works fine (with our copy of Sphinx) then we shouldn't have to clean our `ext/doc` folder. So, the fact that we are having some problems with our copy of Sphinx means that it must be related to our packaging of Python and its packages?