Idea: Git as basis for future CTAN and TeX Live. (Discuss here or at tomorrow's TeX Hour)

[Jonathan Fine]

Hi Norbert

Thank you for your interest and for spotting the typo. Tomorrow's TeX Hour
will be Thursday 24 June, at 6:30 to 7:30pm UK time. As that is at 2:30am
Japan time, I don't expect you'll be attending. But there will be video
available afterwards.

You wrote:

> I guess one can try that out by git cloning my git-svn repo with
>         git clone --depth 1 --branch trunk ...
> that should only fetch one release of all the files.
>

My set of files was defined to avoid documentation files, and files that
aren't used by anyone, and binary executables. In other words, only files
that are input by tex and similar programs for typesetting purposes (and
which lie in TeX Live), and which are used by members of this list.

So your suggestion of fetching a single release would not be minimal for
the purpose I have in mind.

 You also wrote:

And we would need to push --force so that the history does not grow.
>

We don't know that yet. It all depends on how big 'all useful texlive input
files' is, over the years.

And there need not be any history, and so no need to use --force! An
'orphan' commit has no parents. We can create as many of these as we wish!
Not just the usual first commit to the main branch.

https://git-scm.com/docs/git-checkout#Documentation/git-checkout.txt---orphanltnewbranchgt

Please recall that I'm intending in this thread to use git as a distributed
peer-to-peer content addressable store. I'll use its version control
features only as strictly required for these purposes.

You have a special importance to TeX Live. To allow us to talk, I'm happy
to schedule a future TeX Hour meeting at a time that suits you (as long it
also suits me).

best regards

Jonathan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://tug.org/pipermail/tex-live/attachments/20210623/ff995aa5/attachment.html>