Bdy arrays are filled out of their bounds if bdy data is taken at initial step

[jchanut]

Context

bdydta module with initial state as boundary data

Analysis

if nn_dyn2d_dta==0, 2d boundary data arrays are allocated over rim points only (and no matter the use of frs for barotropic variables which is an other error). However these are filled over the whole open boundary set of points later on which causes an out of bound error.

Fix

Restrict boundary data initialization to rim points only in bdy_dta as a quick remedy.

[smasson]

We also saw this bug with Gaston last Friday...

Ssh is always defined and use only on the rim so it is easy, however u2d et v2d can be defined over the whole boundary if ln_full_vel = T (see definition of iszdim in bdy_dta_init). So we were thinking to define the loop upper limit with a size to make sure we always have to proper loop size...

I don't really see the need of the variable ilen1...

[smasson]

In 12547:

Error: Failed to load processor CommitTicketReference

No macro or processor named 'CommitTicketReference' found

[smasson]

fixed in [12547]