Re: [PATCH] git-repack doc: remove discussion of ancient caveat

["Ævar Arnfjörð Bjarmason" <avarab@gmail.com>]

On Mon, Jul 18 2022, Junio C Hamano wrote:

> Ævar Arnfjörð Bjarmason <avarab@gmail.com> writes:
>
>> Which I may have misread, but I understood as going beyond suggesting
>> that we cover #2 over (or in addition to) #1, and into speculation that
>> the change being suggested here was suspect because I hadn't carried out
>> a "solid study of history".
>
> OK, so there was a study of history, but the resulting commit did
> not interpret and reflect what's significant in the history
> correctly.  Sorry for mischaracterizing your mistake.
>
> Lets put it this way.  Here is a statement:
>
>     Since 1.6.0, people started to need to worry more about
>     compatibility with 1.4.4 and older.
>
> Now that statement, while it may be still correct, is irrelevant.
> Why?
>
> Even if there were tons of people who still use 1.6.0 (or 1.5.3 for
> that matter, which happens to be one of my favorite releases in the
> era), as long as nobody uses 1.4.4 or older, we can safely remove
> such a statement from our end-user facing documentation set.  Some
> archaeologists in us may care, but it is irrelevant to the general
> public, as long as 1.4.4 or older have died out.  "As continued use
> of 1.4.4 by people stopped being an issue long time ago, remove the
> warning about interoperability" is the only thing we need to say
> about this change.  We can add "that we needed to add in 1.6.0 era"
> at the end but that is already too verbose.
>
> Please do not be one of those folks we had to deal with in the past
> who for whatever reason cannot admit that they were wrong.

I won't, and I think on the subject of documentation & commit messages
you (as in me, in this case) have already lost the "argument" if you're
having to explain what was meant by the text, which clearly should be
self-contained enough to resolve any such ambiguities by itself.

I was just replying to clarify what I was trying to go for in the v1,
which in my mind isn't an argument *for* that version, but just an
explanation. I.e. sometimes the solution is to more clearly phrase what
I was going for, and sometimes it's to take another approach entirely.

In in case, I we may have crossed in E-Mails, here's a re-rolled v2:

    https://lore.kernel.org/git/patch-v2-1.1-98b6de56019-20220719T000847Z-avarab@gmail.com/

I tried to make the dates & versions involved clear, and also to work in
your preference for v2.0.0 (or possibly later) as a cut-off.

It may not at all be what you had in mind, or maybe you're willing to
queue it as-is, just let me know & I'll adjust it.

Thanks a lot for the review, and sorry about the misunderstanding.