Follow-up from "[MeL/IO] Xdmf Writer without xdmf library"
The following discussions from !3703 (merged) should be addressed:
-
@endJunction started a discussion: It might be wrong to use resize here instead of reserve. The following push-back will extend the grids vector further...
-
@endJunction started a discussion: - The
move
seems inappropriate here withgeometry
being a const-ref in the arguments... - Is a copy of geometry needed?
- The
-
@endJunction started a discussion: Should be
{:s}
. fmt will throw otherwise when about to print the error message. -
@endJunction started a discussion: Indentation...
-
@endJunction started a discussion: Most of the empty lines can be removed... clang-format.
-
@endJunction started a discussion: Something using std::distances would be nicer. Don't have a suggestion yet.... Anyway pointer arithmetic looks troubling, assumption of linear container...
-
@endJunction started a discussion: pass int and double by value.
-
@endJunction started a discussion: filename is stored by value, so passing the string by value and moving in the initializer list could be beneficial.
-
@endJunction started a discussion: Wouldn't static constexpr array work instead of function?
-
@endJunction started a discussion: function names
meshItemTypeString()
... -
@endJunction started a discussion: else
after return is not needed...Because the line 47 was not executed by any test, an OGS_FATAL might be more appropriate....
-
@endJunction started a discussion: resize here looks wrong. Should be reserve, I'd think.
-
@endJunction started a discussion: const
-
@endJunction started a discussion: array might be better here, either std::array or c-array.