>>>>>  NP == Norbert Preining [2017-7-22]

NP> Combining what you two said, it seems that the easiest way forward for
NP> you (upstream) and not too disturbing for downstream to stop doing
NP> whatever "tarball releases" you are doing, and we consider the
NP> elpa releases as "the releases" which are properly tagged in git, and
NP> people can build from there, or get the tarballs from the elpa
NP> repository.

I've released Debian auctex 11.91-1 based on the AUCTeX release_11_91
Savannah tag.  As far as Debian packaging is concerned tarball releases
are no more necessary.  @Mosè: I'd suggest you to sign future release
tags with the same key you are using to sign the tarballs.

NP> In this case, both the current stable release tags, as well as the
NP> hopefully to be done elpa release tags can be used. I agree that
NP> this would be good.

>> > Apart from packaging, I think that AUCTeX users would benefit from
>> > having clearly marked stable releases.

NP> Well, I think this is what is happening by now, with release balls and
NP> tags in git. It is more about the additional releases to elpa and how
NP> they are reflected into stable releases, or whatever stable releases
NP> will be.

Yes, I agree.  ATM I do not understand if elpa releases should be
considered stable releases.  I tend to think of them as bleeding edge
releases, maybe suitable for Debian experimental, but unsuitable for the
unstable->testing->stable cycle.  WDYT?


auctex mailing list

Reply via email to