Re: [tap] RFC Status?

[Leon Timmermans <[e-mail]>]

On Mon, Aug 11, 2014 at 8:42 PM, Andrew de Andrade <[e-mail]>
wrote:

> I'm working on some testing libraries for JavaScript based on TAP.
> Specifically, I'm trying to explore making it possible for library authors
> to not only check the results of their TAP tests, but also map-reduce the
> TAP results of the dependents of their library to see how changes to their
> library impact the correctness of their dependents.
>
> e.g. I am the author of library A. I publish version 0.1.0. Another author
> includes my library as a dependency in library B and in library C and uses
> TAP to test those libraries. Given this scenario, it would be nice if I
> could, when modifying my library, run not only my tests, but also run the
> tests of all my dependents built with my previous version and my current
> version and compare how my changes impacted the "correctness" of their
> programs.
>
> Beyond the benefits to a community, helping authors know when they can
> likely safely upgrade their deps, this type of runner would also help in an
> corporate environment with lots of code re-use between teams.
>
> While thinking about this, I decided to go out and figure out if there is
> such a thing as nested TAP in the protocol specification. While searching
> for this, I came across this group, which is a great, but it appears there
> has been no activity here in a while and it looks like all the wiki pages
> with prior information about the state of the TAP protocol becoming an IETF
> RFC have disappeared.
>
> What's the current state of TAP? Is there still interest in taking this to
> RFC status? I found some previous discussions on the list about nested TAP.
> Did those make it into the specification in any way? If so, where can I
> find examples of correct nested tap results?
>
> Furthermore, is there a standardized set of test fixtures for the current
> version that any implementation can be tested against for correctness and
> performance? Having worked a bit with JSON schema, I found that one of the
> most useful tools was a standardized set of tests to check if a particular
> implementation conforms to the protocol standard. (see:
> https://github.com/json-schema/JSON-Schema-Test-Suite )
>
>
I've been recently writing a TAP Harness, and for subtests I followed what
Perl has been doing for years (just like TAP::Stream): have indented
subtests follow by a non-indented summary line. None of this is described
in detail anywhere though :-/, nor is the interaction with the barely
defined YAMLish well-explored AFAIK.

Leon