Skip to content

Conversation

@jonboulle
Copy link
Contributor

@jonboulle jonboulle commented Mar 14, 2017

As noted in
#608 (comment) , we shouldn't be mandating that implementers use something like
tar-split to facilitate reproducible DiffIDs, but rather recommend it.

Signed-off-by: Jonathan Boulle [email protected]

As noted in
opencontainers#608 (comment)
- we shouldn't be mandating that implementers use something like
tar-split to facilitate reproducible DiffIDs, but rather recommend it.

Signed-off-by: Jonathan Boulle <[email protected]>
@jonboulle jonboulle mentioned this pull request Mar 14, 2017
@vbatts
Copy link
Member

vbatts commented Mar 14, 2017

fair, though it doesn't say they must use tar-split, it is just an example right?

and what's this even saying is that implied by having a tar archive that matches checksum, so how that is achieved is up to the implementer, right?

@jonboulle
Copy link
Contributor Author

@vbatts see my comment at #608 (comment) , this is probably just the ambiguity of natural language. Nevertheless seems cleaner if we can standardise on the RFC2119-style

@stevvooe
Copy link
Contributor

stevvooe commented Mar 14, 2017

LGTM

fair, though it doesn't say they must use tar-split, it is just an example right?

Agreed.

@vbatts We could probably relax this altogether. In fact, if the exact DiffID is not reproduced, everything will still work.

Approved with PullApprove

@vbatts
Copy link
Member

vbatts commented Mar 14, 2017

LGTM

Approved with PullApprove

@vbatts vbatts merged commit 8d87da9 into opencontainers:master Mar 14, 2017
@vbatts vbatts mentioned this pull request May 19, 2017
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants