On Sat, 17 Aug 2002 12:15:18 +0200 Oskar Sandberg <[EMAIL PROTECTED]> wrote:
> > As long as bug reports around here stay just one step above "MOMMMEEE, > IT DOESN'T WORK - FIX IT NOW!!!!" It's a pretty difficult job. > > Is it always the same pages? Do they never work? What are they doing > that is different then the ones that do work? Is it tied to a special > metadata function? Does it effect the CLI clients as well? Why should > I waste my time trying to track these things down when nobody else > even bothers to check what is actually happening? > I don't happen to know anything about the current problem, nor do I have a node that I could get working in short order, but... do you think that, for debugging purposes, either or both of the following could be done? * Having the snapshot generator tag CVS with the date of the snapshot, right before it makes it, and/or * Having the developers tag CVS with "build_NNN" whenever they decide to release a new build ? That should help mightily with debugging (cvs [r]diff -r build_493 -r build_494), and also with time-travel (copy tree; cvs update -r build_493 ) And for those who may not remember this particular annoyance: always run 'cvs tag' _after_ committing, as it knows nothing of your local copy (silly silly silly). Anyway, my $0.02. Thanks -- hobbs
msg03612/pgp00000.pgp
Description: PGP signature
