protect pre_update_intracellular from changes made in post_update_intracellular

[drbergman]

In models with both pre_update_intracellular and post_update_intracellular functions, it is currently possible for some cells to complete their post-update before other cells begin their pre-update. If the post-update modifies values in those cells that others use in the pre-update function (for example, surface-bound ligand expression), the pre-update function will use the incorrect values to update. This PR addresses this by performing the intracellular update as follows:

1. Loop over all cells
   a. Determine which are ready for intracellular updates
   b. For these, do their pre_update_intracellular (if defined)
2. If no updates found, return from update_all_cells_intracellular
3. Loop over all cells ready for update and call their update function
4. Loop over all cells ready for update and call their post_intracellular_update (if defined)

Commit comments

• first do all pre updates, then all updates, then all post updates
• this ensures that any changes manifested in updates/post-updates are not immediately visible to the pre-updates of other cells
• with the more verbose logic controlling intracellular updates, make a separate function to handle it
  • in that function, encode directly dependence on diffusion_dt rather than the argument diffusion_dt_
  • initialzed used to determine when to call intracellular rather than passed in
  • if after first looping over all cells, none were found to require intracellular updates, then return early