[netmod] YANG Versioning Weekly Call Minutes - 2020-12-01

["Sterne, Jason (Nokia - CA/Ottawa)" <[e-mail]>]

YANG Versioning Weekly Call Minutes - 2020-12-01

We talked about:
1) preparing for the virtual interim, and
2) the whitespace issue (item B of the virtual interim agenda)

Jason to work with the chairs to go ahead with Dec 14th.

We'll try to encourage/build some good participation for the interim over the coming weeks.

The interim will be the final discussion for these issues - after that we will make a decision and put the result into the final last-call drafts.

Need to progress on materials preparation before next weekly call. Target: get materials reviewed and out by Monday Dec 7th.

Some rough notes from the whitespace issue (B):

Should insignificant whitespace changes (i.e. outside quotes) to a YANG module be allowed without changing the revision?
(e.g. windows CR/LF)
- decorate in separate file
- scope of this issue is insignificant whitespace changes
- is this a checksum of a file, or of the significant YANG statements? Would have to define canonical YANG.

Treating whitespace changes as insignificant (i.e. as the same version) would make clear that two modules have the same meaning of they differ only in insignificant whitespace. It would also make things like Windows vs. Unix line endings an insignificant matter.
- have to define canonical output format (not just a tool):
   - CR/LF
   - order of sub-statements:  is that editorial change? Or is it "insignificant" ?
   - order of leafs within a container?  No - you can't change order of leafs (that matters)
- does 7950 actually define significant vs insignificant
- any "YIN" considerations ?

Treating whitespace changes as editorial changes (i.e. as a new version) would make it possible to use regular checksum tools to check the integrity of YANG modules and packages. It would also make clear which of two modules with whitespace differences is the newer version.

Rgds,
Jason