Richard Atterer wrote: >On Wed, May 29, 2002 at 01:14:23AM +1200, Philip Charles wrote: > >>The extra H4 files came from alpha.gnu.org/../debian-libio, future >>sets will be using alpha.gnu.org/../debian. This illustrates the >>problem of producing kits like jigdo for the Hurd, the Hurd is a >>rapidly moving target. >> > >It's true that this makes using jigdo difficult, but as you can see >from Attila's sid CDs and DVDs, it's possible. > >For files that change very frequently, grep them out of the filenames >passed to jigdo-file so they end up in the template. > >For other files (packages on the CD that become outdated and disappear >from the main mirror), one possibility is to set up an md5sum-based >fallback directory; for a "MD5Sum=http://foo/"; entry in the jigdo >file, jigdo will first look in the main mirror for a file, then for at > That's what Attila does in debain-superseded, AFAIK. Quite big, 9000 packages.
So we need to know how much packages changed since creation of the image, if the changing file group is quite the same, and how "fast we are moving". If we can guaranty that certain packages do not change in a fixed period of time, making stably usable jigdo files much more possible. Otherwise we would need the fallback directory. But that would limit mirroring of the packages and we would loose on big advantage, beeing independant of a certain server to download / distributing network load. Some statistics for the hurd-H4-CD1 image: 2069 files, 1888 packages, 381 could not be fetched from the Deibian mirrors (of which 3 where duplicates. Can this even happen?) 14 Packages* files, 4 Release files, As I understand the system, that last two simply mean, the archive is not hte same as on CD. 91 packages can be fetched from alpha.gnu.org, 154 packages could not ber found on alpha either. I expect them to be superseded. ~120 remain as non-package files. These are: docs, tools, TRANS.TBLs, disks, etc. The remaining, Release and Packages* files should go into the template as they will never be mirrored anyway. When including the 154 not found packages with 0.3 Mb/package, I estimate template to be about 50-70 Mb. This would be a template that worked at least until yesterday ;-) (At this point I'd like to thank my new friends grep, less and sed and of courde the pipe :) Patrick -- Engineers motto: cheap, good, fast: choose any two Patrick Strasser <pstrasser at bigfoot dot de> Student of Telematik, Techn. University Graz, Austria -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
