On 25 Oct 2024, at 14:49, Richard Purdie <[email protected]> 
wrote:
> I've sent out a change for do_package which stops the messing with the
> timestamps. It breaks reproducibility of the do_package sstate but that
> is probably the right compromise. See what you think...
> 
> Does that mean we don't need to this patch (since this feels like a bit
> of a hack)?

Oh it’s very much a hack.

I’m actually surprised the hash equiv logic tweak was that simple, I should 
have looked at that instead of assuming eldritch code!

I agree that interim do_package sstate that is essentially an optimisation is 
fine to be non-reproducible, because the final will be clamped by SDE anyway.  
Drop this patch, and merge yours.

Ross
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#206443): 
https://lists.openembedded.org/g/openembedded-core/message/206443
Mute This Topic: https://lists.openembedded.org/mt/109194063/21656
Group Owner: [email protected]
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub 
[[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-

Reply via email to