Nice! Reviewed-by: Jamey Sharp <ja...@minilop.net>
I have a few observations to make. I don't know that changing any of these things would be an *improvement* but I'll note them anyway. :-) The xcb_flush call isn't necessary: each call to get a reply ensures that the corresponding request has already been flushed. In the current implementation, the first call to xcb_query_extension_reply here will cause all the requests to be flushed. Variable-length arrays are a C99 feature, as I recall. Is that OK for xdpyinfo? I'd probably just use malloc, or perhaps queue up a fixed number of requests at a time? The number of recv syscalls may be artificially high on Xlib's side right now, due to a known bug where we usually call recv at least twice per call to _XReply. Ignoring recv syscalls that returned EAGAIN might be more informative. Measuring number of packets on a TCP-based X socket might be even better? Jamey _______________________________________________ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: http://lists.x.org/mailman/listinfo/xorg-devel
