on Wed Sep 05 2012, Lars Ingebrigtsen wrote: > Lars Magne Ingebrigtsen writes: > >>> What I found was that the nntp-send-mode-reader function would have to >>> process an enormous number of lines every time I hit `G' in the >>> *Group* buffer, which was quite slow. >> >> Uhm... `nntp-send-mode-reader' doesn't process anything. >> >> Could you ELP-instrument nntp and see what takes so long? > > Any progress here? I just reproduced this behavior and ELP'd it with mkdir /tmp/larsi mv /tmp/larsi/ HOME=/tmp/larsi emacs `M-x el-get-install RET nognus RET' `M-x RET gnus RET' Naturally, I can't give you the .authinfo file that goes with this directory, but I'm sure you can massage things appropriately. el-get takes a *long* time to initialize the first time you start emacs with this init file; I'm sure you have a faster way to get the latest Gnus into your load-path, but this is reliable. It produced the following ELP results --8<---------------cut here---------------start------------->8--- nntp-request-list 3 12.511693000 4.1705643333 nntp-with-open-group-function 3 12.511677 4.170559 nntp-send-command-and-decode 3 12.511605 4.170535 nntp-accept-process-output 1984 12.323542999 0.0062114632 nntp-open-server 3 3.848937 1.2829789999 nntp-open-connection 3 3.848764 1.2829213333 nntp-send-mode-reader 3 0.361392 0.120464 nntp-decode-text 11 0.356008 0.0323643636 nntp-send-authinfo 4 0.05992 0.01498 nntp-handle-authinfo 1 0.057592 0.057592 nntp-find-connection-buffer 1984 0.0036499999 1.839...e-06 nntp-read-server-type 3 7.9e-05 2.633...e-05 nntp-make-process-buffer 3 3.8e-05 1.266...e-05 nntp-possibly-change-group 3 2.499...e-05 8.333...e-06 nntp-erase-buffer 17 2e-05 1.176...e-06 nntp-server-opened 9 1.499...e-05 1.666...e-06 --8<---------------cut here---------------end--------------->8---