On Saturday 23 October 2004 19:55, Richard Atterer wrote: > On Sun, Oct 24, 2004 at 02:44:26AM +1000, Robert Parker wrote: > > It's basically pointless. > > No, IMHO it would be a very good idea indeed! ;-7 > > That's because *I* get all the hate mail from people who think that jigdo > is buggy, fundamentally broken etc etc... ;-|
Richard, I understand the situation you have been undeservedly involved. Jigdo is a brilliant method (and implementation, along with the JTE), but its working schema is not fully understood by a great deal of newcomers (newcomers are too fast and impatient to read the docs first). It is hard fot some of them to see the phenomenal idea behind this method, and I think there must be some hints suggested by the jigdo program itself in case like these. Since it is very hard to stipulate (as well as to complete) for all .jidgo/.template producers: 'hey you producers, keep your last_week_purged_packages directory or the like' , then please let's make the jigdo-lite program to print some explanations that it is not a program malfunction, but a producer's one, and this might be completed by: *loop mouting the incomplete iso and searching for newer .jigdo/.template pair *using rsync to complete that version of the iso, if available on any the server(s) *fallbacking to the snapshot server as the last resort. This is why the snapshot server was created for, to search and download for packages that have been drop from the current official debian archives. Can we make something like a servey of how many people use jigdo and try to predict what could be the impact on the snapshot server. (some more hints might be suggested by the others, and prioritisized as appropriate) Yes I remember about your idea to implement a 'phone call system', but IMHO it will add another level of sofistication on the top of the jigdo schema and it wont solve the problem of downloading the 'dropped from the archives' files. -- pub 4096R/0E4BD0AB 2003-03-18 <keyserver.bu.edu ; pgp.mit.edu> fingerprint 1AE7 7C66 0A26 5BFF DF22 5D55 1C57 0C89 0E4B D0AB -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
