Re: [netmod] Joint WGLC on "semver" and "module-versioning" drafts

[Carsten Bormann <[e-mail]>]

On 2023-06-04, at 19:42, Jürgen Schönwälder <[e-mail]> wrote:
> 
>> 
>> I’m not sure I understand the current discussion, but wouldn't
>> 
>> curl -s https://www.rfc-editor.org/rfc/rfc9127.xml | xmlstarlet sel -T -t -v "//sourcecode[@name='[e-mail]']/text()”
>> 
>> be considered an authoritative source for that YANG file in that RFC?
>> 
> 
> There are many ways to extract YANG modules from RFCs and the results
> they produce are not necessarily byte-level identical.

Thanks.  This particular method only works for RFC 8650 onwards, but could be considered canonical for those.  As XML is neutral with respect to the bytes that constitute a line break, if byte level determinism is required, the canonical form of a line break (#xA, i.e. a bare LF) should be assumed.  There also is https://github.com/ietf-tools/xml2rfc/issues/986 (YANG files extracted this way typically start with an empty line).

> So far this was
> not considered necessary. Before people go and try to engineer a
> solution, it may be useful to understand why solving this problem is
> relevant or important.

Right.  I wasn’t about to engineer a solution, but more interested in finding out what people consider the authoritative solution today, and I now have a useful answer.

Grüße, Carsten