On 05/06/07, Jesse Keating <jkeating@xxxxxxxxxx> wrote:
On Tuesday 05 June 2007 07:04:17 Axel Thimm wrote: > Why not? Care to detail this? It's a simple timing issue. Given you have a build chain 4 packages deep. First rebuild will rebuild the first package, and all other 3 packages against the old build of the first package. Second rebuild will rebuild the first package again, and all other 3 packages against the first rebuild of the first package. That still leaves packages 3 and 4 as not being rebuilt against the resultant rebuilt of package 2 against rebuild of package 1. You'd have to either rebuild 4 times, or insert delays into the rebuild so that 1 lands in buildroot before 2, then 2 lands in buildroot before 3, so on and so forth.
Right. That's the problem that needs solving, not a reason for not doing full rebuilds.
This is just one example where automated rebuild, while it does some good, doesn't really fix all the problems you think it would. It just hides them further under the rug under the assumption "But we did a full rebuild, everything should just build fine now..." when in reality your full rebuild didn't accomplish that, it just gave you a warm and (false) fuzzy feeling.
Isn't the warm fuzzy feeling warranted if the above issue is solved reproducibly ? J. -- Fedora-maintainers mailing list Fedora-maintainers@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-maintainers -- Fedora-maintainers-readonly mailing list Fedora-maintainers-readonly@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-maintainers-readonly
