PEP 684: Fix invalid format for Discussions-To header

[CAM-Gerlach]

In #2393 , the Discussions-To header of PEP 684 was updated to an invalid format, per the initial discussion in #2266 documentation in #2346 , initial processing implementation in #2351 , final implementation and full conform in #2361 , and as stated in PEP 1 (PEP-0001) and PEP 12 (PEP-0012) and approved by the Steering Council in python/steering-council#113 . This PR fixes that.

To note, I've already implemented and tested updates to the validators to catch regressions like this and hopefully avoid this incident in the future, so that the headers can be easily parsed programmatically (which allows us to do all sorts of nice things without having to touch existing PEPs or ask anything more of PEP authors), and PEP authors are given immediate and specific feedback in case of mistakes. They're just blocked on #2375 which is blocked on #2358 , which should hopefully all be merged very soon.

[ericsnowcurrently]

People should post to python-dev for discussion, not the archive that link points to. But I guess it isn't that important in the end. It just means Discussions-to isn't all that meaningful. :)

[CAM-Gerlach]

People should post to that thread on python-dev, not the mailing list archive. Hence the link text.

The linked mailing list archive thread contains a prominent Reply button at the bottom, which will send a reply to the canonical thread (rather than merely start an arbitrary new thread on the list, as would just a mailto: link to the address), while also allowing current and future readers to browse what has already been said about the PEP. In addition, there is a button on the top left for users to join the list in question, which is necessary so they can post and receive replies.

Next time, if you spot a problem with the current standard format and would like to propose an alternative, you are welcome to open an issue on this repo (though I suggest you review the issues and PRs I referenced, which contain extensive discussion, debate and eventual consensus among the PEP editor/infra team that resulted in the present status quo).

It just means Discussions-to isn't all that meaningful. :)

The link in the rendered output is of substantial importance, as it clearly and unambiguously directs users to where they can view, subscribe and reply to the current canonical discussion thread (without them having to know that they need to parse the Post-History field and presume they need to click the latest link).

However, you make a good point that with the changes in #2358 , including a Discussions-To header in the reST source is technically redundant in most cases with the latest link in the Post-History header. Since we can parse that now (given its consistent and validated format once the mentioned PRs are merged), once PEP 676 is live you can simply omit that field entirely, and the rendering infra can extract the link and display it with the existing processing in the Discussions-To field.