collate vs. multiple-document-collated-copies

[George Liu]

Looks like
"lp -d xxx -n 2 -o collate=true" and
"lp -d xxx -n 2 -o multiple-document-handling=multiple-document-collated-copies" are handled differently.


for -o collate=true, it looks like CUPS searches PPD to see whether printer supports hardware Collate, if so, merely set "*RBINumCopies=2".

For multiple-document-collated-copies, CUPS just assume printer doesn't support hardware Collate, use software to emulate collate effect and set "*RBINumCopies=1" (As a result, the size of datastream is multiplied)


Unless there's any underlying reason, I believe CUPS should use a consistent behavior to handle collate. I prefer the way "-o collate" is handled. (meaning CUPS checks whether printer supports hardware collate)