On Tue, Dec 09, 2003 at 09:34:40PM +0000, Colin Watson wrote: > On Tue, Dec 09, 2003 at 09:07:00PM +0100, Javier Fernández-Sanguino Peña wrote: > > In order to verify the integrity of the CVS archive, and restore the www > > CVS (and the generation of the website) [1] , it would be necessary to > > check the CVS _history_ (not just the last version) of all the files > > available in the CVS [2] > [...] > > [2] If not we will need to import just the latest branch after checking > > it's integrity with offsite copies (which are more readily available) thus > > losing all the www CVS history. > > Isn't that an overreaction? I don't think it would matter *that* much if > things not on HEAD were corrupted, particularly since webwml isn't a > repository where people are liable to check out old tags and try to run > them. In this particular case, I think it's clearly better to have > history that may be corrupt in places than no history at all. That's what I've been asked to do on IRC. I understand that stuff on HEAD should be checked first (joy is doing it and I'll try to help too). But the source code which does run when compiling the source might be eligible for a more intensive history revision than, say, chinese translation of content. In summary: my opinion is the same as yours, but we've been asked to review the CVS history too. In any case, debian-admin might want to pop-up and say "we have backups!" (hopefully they will) Regards Javi
Attachment:
signature.asc
Description: Digital signature