Denys Duchier writes: >> One idea occurred to me -- perhaps what's unique about this group >> might be the distribution of articles in the group? Do you have any >> ticked/dormant articles in the group? In that case, altering >> `nntp-nov-gap' to, say, 100000 might help. altering nntp-nov-gap had no effect >> But the problem is that something in your connection chain is echoing >> when it shouldn't, really. > > Maybe it's not in my connection, but at the server's end. This server > was just taken down for filling its log file (and hence its partition) > with some sort of tracing info (I don't know the details yet). So > maybe it was ill configured (in a very strange and non-uniform way > :-). I'll know more this evening or tomorrow. The server was fixed and restarted and I still have the same problems. I have now noticed that I actually encounter the problem on other newsgroups, but there it has a different effect: if I am not mistaken correctly, it just gives me warnings about "invalid data on line so-and-so" (looking in the " *nntpd*" buffer reveals stray XOVER commands). However, in comp.lang.functional, it just won't go past that XOVER command and blows up loudly (why does it blow up in that group and not the others, I just don't get). For example here is an extreme case, where there are 55 stray XOVER lines, and yet all I get are messages flashing in the echo area informing me of "invalid data" on the corresponding line: