Thursday July 28th
Video recording: https://www.youtube.com/watch?v=2CN1jB3OuOY
Chairs (10 minutes)
Introduction - Lou Berger presenting
Please see the note well.
bluesheets are handled automatically by meetecho.
discussion to be had around whether to reach out to the IEEE
Working remotely chairs are here to support virtual interims and side meetings.
proposal to not include in-disards-overflow
no objections in meeting
concerns with the subinterface vlan model changes?
Lou Berger - are we ready for publication?
will confirm on list for resolved comments
do we need to send a liason over the vlan terms
Rob Wilton - will hand ad responsibility for these drafts to warren
don’t think formal liason is required. they have been involved.
use the ieee-ietf coordination mailing list
also use the ieee yangsters review since they have done it.
Lou Berger - I can send those if you prefer
Warren Kumar - sounds good to me, prefer that the mail not come from me.
Lou Berger - I’ll send it
Lou Berger - believe that we previously agreed to hold the documents until all are ready to go.
Reshad Rahman - authors plan to get modver and semver to wg lc first and then focus on progressing the other documents
interpreting revision labels in semver
use of revision labels by submodules
review section 3.3 update
revision label in filename
lada - having a symbolic link is an issue due to the changing filename
these is even worse then date
Reshad Rahman - downside is without it you have to look at the contents of the file.
Rob Wilton - git is to not have the revision date or label in the git repo itself but publish it somehwere else
Kent Watsen - agree with rob
revision label uniquely identifies a module version.
Ladislav Lhotka - this also includes changes in whitespace?
NBC changes can break imports
Kent Watsen - the break in backwards compatibility may not be in the area the matters.
Rob Wilton - not using the revision label to make decision, rather the revision history is used.
Kent Watsen - python packages declare compatibility with current / past versions but not future versions.
Kent Watsen - it does sound confusing the purpose of this effort is to address this problem
Lou Berger - would be really helpful if you (rob / reshad) would send your example to the list.
impact of changing import statement
Reshad Rahman - consider rough consensus to be option b (not sure now)
Rob Wilton - is the example sufficiently clear
Kent Watsen - are you speaking as contributor?
Lou Berger - do send the email I thought I understood it until your last statement
should we import yang revision identifier?
insignificant whitespace changes
Ladislav Lhotka’s earlier question.
Reshad Rahman - should result in a new change e.g. editorial.
Charles Eckel - don’t userstand the usecase for whitespace but nothing else
Reshad Rahman - don’t have an answer for you
via jabber - Jan Lindblad
There is probably no reason to publish with only whitespace changes. The point is that there should be a version number for every different checksum
Ladislav Lhotka - we still have this alternative syntax which raises the case of two syntaxes of the same version.
Kent Watsen - the yin files have a different suffix
Joe Clarke - if I decide to publish something and all I change is whitespace I still want to note that.
Ladislav Lhotka - starting from a YIN module and using different tools for translating it to YANG, one can easily end up with modules differing in whitespace.
No comments - please send other comments/questions to the list.
Kent Watsen - will take questions to the list.
Lou Berger - we will need check IPR
Qin Wu - okay
Kent Watsen - similar work in NETCONF, need to ensure importance of work.
Qin Wu - Not a strong a dependency.