Open hackerb9 opened 3 years ago
This was more noticeable before I increased the INCR block size, but sometimes xclip's pastes (acting as a requestor) felt slower than if I used xsel.
Transfers over a network seemed to exacerbate the problem.
Basic speed tests can be done with Pipe View (pv): host1$ dd if=/dev/zero bs=1M count=16 | tr '\0' a | xclip host2$ xsel | pv >/dev/null host2$ xclip | pv >/dev/null
Note that while requesting a selection is slow (paste), having xclip own a selection (copy) is 100x faster than xsel for xfrs.
This was more noticeable before I increased the INCR block size, but sometimes xclip's pastes (acting as a requestor) felt slower than if I used xsel.
Transfers over a network seemed to exacerbate the problem.
Basic speed tests can be done with Pipe View (pv): host1$ dd if=/dev/zero bs=1M count=16 | tr '\0' a | xclip host2$ xsel | pv >/dev/null host2$ xclip | pv >/dev/null
Note that while requesting a selection is slow (paste), having xclip own a selection (copy) is 100x faster than xsel for xfrs.