I've seen numerous inconsistencies between elements and export modules when it comes to exporting "giveIPValue"

Primarily 3 sources of errors:

1. Not exporting the same numbering of components the same way

2. Not exporting the same type of tensor (S3E, S3 or G tensors)

3. Exporting in either local or global c.s.

As well as some other serious problems, like dimension errors. For examle, the ShellForceTensor  and ShellMomentumTensor should be force/length, not actually a force.  Some elements multiplied with the thickness (which wouldn't be the correct dimension to multiply with) to obtain a force.

Most of it probably stems from copying from different elements over and over, and I suspect very little of this is actually tested.
I fixed up (i think i did) all of them, except for number 3, global v.s. local c.s.

I added an option "giveGlobalIPValue" some time ago, which works for any vector or tensor. I think this is the best option, and the element should just output it's local variables when "giveIPValue" is called. This simplifies the shell/plate elements as well.

But, for output, I'm not sure which is the best way to go. Sometimes users would like local c.s., sometimes global. Global is of course always the safe option, as it will always be sensible answers even when nodal averaging and other recovery models are used.
Either way, we should do something, and keep it consistent