What if we use a different file ? This would allow us to use a different semantic an adjusted plan for a smoother transition.

In D10082#157428, @joerg.sonnenberger wrote:

In D10082#157189, @marmoute wrote:

What happens in the stream clone case ? Or during the transition from older to newer client ? streaming content from an older clone will not contains a fully up to date cache, and existing repository neither. How are we dealing with this ?
You mention the need to update the cache after stripping, but I don't see such change in the code. am I missing something ?

The cache will fix up itself when it detects missing entries.

(I am stopping my review here since I am the actual author of those)

why ?

(maybe explain why in your description, the behavior is already "encoded" in the patch)

Do you have any performance number for that update ? for example the time we need to do that full some known large repositories, and the impact on a clone ?

This is all true.

+1 (the set binding was not available at the time of writing, but they are not need to keep doing this now.)

In D9765#156614, @pulkit wrote:

Ping on this, let's get this ready and pushed for upcoming release.

What happens in the stream clone case ? Or during the transition from older to newer client ? streaming content from an older clone will not contains a fully up to date cache, and existing repository neither. How are we dealing with this ?

In D10272#157092, @pulkit wrote:

I am +1 on this. I remember there was a discussion on IRC around it but not sure what the end result of that was. @marmoute does having experimental.rebaseskipobsolete serves any purpose?

In D10297#157076, @pulkit wrote:

@marmoute we should add pytest-vcr to heptapod.