proff@suburbia.net writes: > A failure in this initial group setting code is treated by nntpcache as a > general server failure (I think the rmgroup case is so rare as to ignore) I disagree. The recent occurance was the comp.lang.java reorganisation, where groups were removed. This sort of remove one general group, and replace with several specific groups is common within Usenet. The problem seemed to be that a reference to a group remained in nntpcached's merged list of the active files long after the group was removed. Should it not be the case when merging the active lists, that a scan should be done to see that if wibble.group was not returned by the server that is deemed to hold wibble.*, then it should be dropped from the list? -- `O O' | Nick.Holloway@alfie.demon.co.uk http://www.alfie.demon.co.uk/ // ^ \\ | Nick.Holloway@parallax.co.uk http://www.parallax.co.uk/~alfie/
