Moving GAAS to use ExtData fundamentally breaks GAAS until new features can be implemented in ExtData

Just making a meta issue here for reference. As the issues exposed in GAAS with the latest model are fundamentally an issue with GAAS.

GAAS was modified to use ExtData. This was a serious mistake that should not have been done without a new feature request for ExtData first and the change was apparently not tested.

The issue is that before, if turned on, GAAS would literally decide whether to perform whatever it is it does based on whether the file it was looking for existed.
Like if you gave the GAAS grid comp a file template. Suppose the template was like this: gaasfile_%y4%m2%d2_%h2%n2z.nc4 if it did not find an existing file after evaluating that, it would just do nothing.

GAAS was modified to use ExtData. This is a problem in that ExtData is just trying to fill a file. It it doesn't find a file it just fails. In the DAS case, you start at say 21z, give ExtData a template gaasfile_%y4%m2%d2_%h2%n2z.nc4 like this, there is no 21z file, you don't have a file until 0z. It just fails which is exactly what it was supposed to do which is what happened when tested in the GEOSadas.

I am modifying ExtData to allow for missing files in a sequence and mark the field with something to indicate not to use if the file is missing which ExtData.
In the future we will need to add a test that simulates the way the GAAS runs in the model when driven by the GEOSadas when not actually running in the GEOSadas framework.