Hi Richard, That makes perfect sense. I'll take another look and see if I can find what else fsck might be doing to increase delta.
Thanks! - Levi On 8/6/26 06:39, Richard Purdie wrote: > On Tue, 2026-08-04 at 15:45 -0600, Levi Shafter via lists.openembedded.org > wrote: >> Hi Ross, >> >> I'd like to bump this patch and ask if you've had a chance to review my >> last message. Considering additional delta beyond differences in >> timestamp information is observed between builds, do you think a toggle >> that allows for a choice between problems represents an improvement >> until the path to full reproducibility with the fsck command becomes clear? > > It is really easy just to put a conditional code path in for every > issue someone encounters. The trouble is that if we do that every time > someone encounters an issue you end up with an unmaintainable and > untestable maze of "if" conditions. > > It is easy to argue this is just one more conditional code path so > people can avoid an error but they do stack up and I'm still very > reluctant to add conditional codepaths unless we really have to and > there will be a clear number of users for both paths. > > In this case, it will create an obsure variable only a small number of > people will set and lead to fragile code which is easily broken. > > I still think we need to get to the bottom of the issues with fsck and > the only way we'll do that is if there is some natural pressure to do > so. I'd therefore still be reluctant on the original patch. > > Cheers, > > Richard >
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#242969): https://lists.openembedded.org/g/openembedded-core/message/242969 Mute This Topic: https://lists.openembedded.org/mt/117020092/21656 Group Owner: [email protected] Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [[email protected]] -=-=-=-=-=-=-=-=-=-=-=-
