Open herwinw opened 3 years ago
About the first point: jteps(jt, needle, haystack)
implements x e. y
which is "member in".
So in english, for sparse array jtisempty
checks if zero is in the shape of the array (as the comment also says in j).
Generally the functions often have commented the corresponding j phrase, if not t.c
is good place to check for primitives. Here for example jtrazein
is the monadic case, and jteps
is dyadic case of e.
. https://github.com/codereport/jsource/blob/de91efce021f9e184113527d01a1e3bb373b3079/jsrc/t.c#L816 (and j nuvoc to see what the primitive does)
That makes more sense than the epsilon
how I read it
A branch I had around. It's pretty much a direct copy from the C function, with a few remarks:
jteps
, which confuses me a lot. I'm not even sure why it does this.AN
to something more descriptive, similar to theis_sparse
function. However, this macro is used for assignment as well (seemake_array
), so I just left the C macro here. I think the consistency here is more important.