v0.2.1

v0.2.1 (2019-09-10)

Diff since v0.2.0

Merged pull requests:

• Add features, improve docs, and fixes (#26) (gaelforget)
• Improve documentation + related simplifications & changes (#25) (gaelforget)

v0.2.0

v0.2.0 (2019-08-23)

Diff since v0.1.10

Merged pull requests:

• set MeshArray=gcmarray (new default instead of gcmfaces) (#24) (gaelforget)
• Documentation updates and improvements (v1) (#23) (gaelforget)

v0.1.10

v0.1.10 (2019-08-21)

Diff since v0.1.9

Merged pull requests:

• fix remaining issues with gcmarray (demo3, view, etc) (#22) (gaelforget)
• Revisit custom types & add potential replacement for gcmfaces (gcmarray) (#21) (gaelforget)

v0.1.9

v0.1.9 (2019-08-15)

Diff since v0.1.8

Merged pull requests:

• introduce read / write functions and use in place of read_bin & convert2gcmfaces (#20) (gaelforget)
• improve transport example (#19) (gaelforget)

v0.1.8

v0.1.8 (2019-08-04)

Diff since v0.1.7

Merged pull requests:

• update and improve docs (#18) (gaelforget)
• Overhaul of the handling of grid specifications and variables (#17) (gaelforget)

v0.1.7

v0.1.7 (2019-07-31)

Diff since v0.1.6

Closed issues:

• can't install on juliabox (#7)

Merged pull requests:

• streamline and cleanup documentation (#16) (gaelforget)
• add example of meridional transport computations & plots (#15) (gaelforget)
• Add findtiles to MeshArrays (#14) (lmilechin)
• Remove unnecessary Documenter dependency. (#13) (fredrikekre)

Add documenter

v0.1.6

Update Project.toml

Revisit type definitions, constructors, etc. and indexing

Merge pull request #8 from gaelforget/AbstractArrayAsSupertype

Abstract array as supertype

Rewrite exchange methods and streamline operations to improve performance.

• gcmfaces_exch.jl: rewrite exchange functions in a much more performant
  fashion based on views and subfunctions; exch_cs_target, exch_cs_sources,
  and exch_cs_viewfunctions do most of the work for both cs and llc cases.
• gcmfaces_cal.jl: revisit methods to improve performance of e.g. smooth
  mask: revised behavior and handling of input arguments
  gradient: now 3 methods; dispatch -> different treatments of 1/DXC, 1/DYC
  mask: revised behavior and handling of input arguments
  convergence: remove mask calls, use view, avoid end
  smooth: revised handling of masks, 1/DXC etc factors, precomputed
  KuxFac/KvyFac/dtFac before main nbt loop, and add nFaces-loops
  within main nbt loop to improve performance (time and memory)
• gcmfaces_type.jl: streamline arithmetic operation methods
• MeshArrays.jl: add convergence to export list

Abstract array as supertype + add DOI and doc in README.md

Merge pull request #5 from gaelforget/AbstractArrayAsSupertype

Abstract array as supertype