[top-posting, as a contributor]

I was discussing this issue with others during the Farewell reception.

The best idea that surfaced was:

add to Datatracker the ability to upload a ".tgz" file
"tgz" is a file format that extracts a directory hierarchy
require that the extracted directory contains a well-defined entry point. 
this could be a file called "build.sh" or possibly "Makefile"
the purpose of the entry point is to:
validate examples (dynamically downloading dependencies)
emit coverage statistics
generate tree diagrams
generate XML-file for `xml2rfc` (vis-a-vis Kramdown if needed).
add to Datatracker the ability to also download the ".tgz" format of the 
document (if available)


Of course, the ".tgz" could be 1-1 to a GitHub repo, but this solution stops 
short of mandating GitHub because:
Datatracker generally handles artifacts (the ".tgz" file)
Concern for incorporating 3rd-party infrastructure


Kent




> On Nov 17, 2025, at 11:01 AM, Rob Wilton (rwilton) 
> <[email protected]> wrote:
> 
> Hi Camilo,
> 
> I've added Mahesh/Med to the thread.  I think that this is a conversation 
> that would need to be had with IANA or the RFC Editor.  Arguably, I would 
> think that the RFC editor could/should be the long-term home for these 
> assets, but then that is not what is done with YANG Modules today, they are 
> instead hosted by IANA, so perhaps that is the more pragmatic choice.
> 
> I definitely agree about storing them on Github whilst the draft is being 
> developed.  Ideally, we would have some tooling that manages all of these.  
> This is similar what Med was doing with his YANG draft tooling project.  It 
> is also similar to what I am doing for my drafts (but with different tooling 
> that I started earlier and with a fast compile/resolve step), e.g., 
> https://github.com/rgwilton/draft-yp-observability/
> 
> Here, I use a markdown file for the draft that imports the YANG examples, 
> YANG code, tree diagrams etc, and a build file 
> (https://github.com/rgwilton/draft-yp-observability/blob/main/build.mill) 
> that stitches everything together, after compiling and validating the YANG 
> and instance data examples, so if I update a YANG module and then save the 
> file, then the build tool will automatically detect the change, revalidate 
> the examples and report any errors, and then produce the updated html for the 
> draft.  Tooling like this has some nice other properties since it will also 
> download and extract the YANG files from all the RFCs and drafts that I 
> depend on.  When I get a bit of time, I will probably try and refine this 
> further, e.g., defining named schema associated with the instance data, 
> automatically build instance data file versions of the examples, etc.
> 
> Kind regards,
> Rob
> 
> 
> From: Camilo Cardona <[email protected]>
> Date: Monday, 17 November 2025 at 15:04
> To: Rob Wilton (rwilton) <[email protected]>, [email protected] 
> <[email protected]>, Carsten Bormann <[email protected]>, Kent Watsen 
> <[email protected]>
> Cc: [email protected] <[email protected]>
> Subject: Re: [netmod] Re: Should example modules be allowed to use code tags 
> [Re: AD review of draft-ietf-netmod-intf-ext-yang-16]
> 
> Hi Rob,
>  
> How complicated is to find that “permanent” home? Github is, of course, ruled 
> by a private company and I am not sure the process should depend on it.. We 
> could always keep it on github until the release process where the files are 
> then stored in the permanent home.
>  
> Thanks,
> Camilo
>  
> From: "Rob Wilton (rwilton)" <[email protected]>
> Date: Monday, 17 November 2025 at 04:37
> To: "[email protected]" <[email protected]>, Carsten Bormann 
> <[email protected]>, Kent Watsen <[email protected]>
> Cc: "[email protected]" <[email protected]>, Camilo Cardona <[email protected]>
> Subject: Re: [netmod] Re: Should example modules be allowed to use code tags 
> [Re: AD review of draft-ietf-netmod-intf-ext-yang-16]
>  
> Hi Benoit,
>  
> My proposal (and what I mentioned in the WG) was slightly different:
>  
> I don't think that we should require complete examples in the RFC.  
> Specifically, I think that if we end up wrapping the examples with RFC 9195 
> instance data + YANG library schema information, and bearing in mind the 
> short line lengths allowed in RFCs, I think that we will find "extra noise" 
> and a lot of line wrapping in the examples which will make them hard to read 
> and understand, defeating the purpose of putting them in the document in the 
> first place.
>  
> I also don't think that we should putting the YANG modules into the RFCs, but 
> that is part of the separate Veloce discussion.
>  
> What I propose is:
>  
> Full compilable examples should be stored somewhere else, e.g., in GitHub, or 
> probably a more permanent home (e.g., IANA or RFC editor website).  These 
> should contain or reference any necessary extra metadata (e.g., the 
> associated schema, or the contents of datastores if validating notifications).
>  
> Examples in documents should just be there to aid the reader understand the 
> main structure of the module.  They should be allowed to be incomplete, or 
> leave out long noisy fields (e.g., eliding long keys/hashes in some of the 
> crypto related drafts).  However, the shorter examples in the draft should 
> include a reference/link to where the full validatable examples can be found.
>  
> Kind regards,
> Rob
>  
>  
> From: [email protected] <[email protected]>
> Date: Saturday, 15 November 2025 at 10:28
> To: Carsten Bormann <[email protected]>, Kent Watsen <[email protected]>
> Cc: [email protected] <[email protected]>, Rob Wilton (rwilton) 
> <[email protected]>, Camilo Cardona <[email protected]>
> Subject: Re: [netmod] Re: Should example modules be allowed to use code tags 
> [Re: AD review of draft-ietf-netmod-intf-ext-yang-16]
> 
> Dear all,
> 
> I am trying to summarize the conclusions of this email thread, in light of 
> the draft-cardona-claise-onion-yang-coverage 
> <https://datatracker.ietf.org/doc/draft-cardona-claise-onion-yang-coverage/> 
> presentation in the NETMOD meeting:
> - We want to have instance examples next to the IETF YANG module. Granted
> - We want to have those instance examples WITHIN the same document as the 
> YANG module. I guess so. 
>     This is ideal but is it always possible, because there might be different 
> contexts?
>     See Rob Wilton's comment in the IETF 124 NETMOD meeting minutes 
> <https://datatracker.ietf.org/doc/minutes-124-netmod-202511071430/>
> - I understand that the community would like a different tags for the 
> instances (next to CODE BEGINS/END)
>     And we have some in XML, great.
> - What I am not sure: do we also need the tags for the text draft/RFC (next 
> to XML)?
> 
> Where I would appreciate the help (of someone more clever that I): how to 
> practically add those tags, when you start from editing markdown draft.
> A (process) example would be ideal.
> 
> Thanks and regards, Benoit    
> 
> On Oct 13, 2025, at 16:30, Kent Watsen <[email protected]> 
> <mailto:[email protected]> wrote:
> I find that the ASCII-armor CODE BEGINS/CODE ENDS is an undesirable relic 
> from days before XML-based RFCs.  Now that RFCs are XML-native, better 
> constructs are possible.  I do not think that extracting from Text-formatted 
> RFCs is necessary.  Being able to extract from just XML is fine.  Therefore I 
> do NOT support adding support for code-tags for examples.
> RFCXML has markers= and related attributes name= etc. [1]; you never type 
> CODE BEGINS/ENDS.
> There is no need to ever apply heuristics to pull source from plaintext 
> renderings any more.
>  
> Try kramdown-rfc-extract-sourcecode on the XML if you need a tool.
> Try https://tzi.de/~cabo/i-d or /rfc for a prototype of the “RFC filesystem” 
> I’m proposing (not recently updated though).
>  
> Grüße, Carsten
>  
> [1]: https://authors.ietf.org/rfcxml-vocabulary#markers
>  
> _______________________________________________
> netmod mailing list -- [email protected] <mailto:[email protected]>
> To unsubscribe send an email to [email protected] 
> <mailto:[email protected]>
>  
> _______________________________________________
> netmod mailing list -- [email protected]
> To unsubscribe send an email to [email protected]

_______________________________________________
netmod mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to