Re: Draft of Git Rev News edition 26

From: Christian Couder <christian.couder@gmail.com>
Date: Apr 19, 2017, 12:31 UTC
Message-ID: <CAP8UFD0MbRMwM-0AWc8_DgqE_67MC5J5qSfOdqj13TyRXcvRFA@mail.gmail.com>
In-Reply-To: <40c969f2-636e-f977-e385-d490f37f0ffa@grubix.eu>

Hi Michael,

> Hi Christian,
>
> Thanks for the ping on the draft.

Thanks for you input on this!

Show 6 quoted lines

> Re gpg: Maybe some valuable point of information is what Werner Koch
> himself said in that thread:
> "That [the command line is not a stable API to GnuPG] is not true.  The
> command line interface has been stable for the
> last 19 years.  We only removed a left over PGP-2 backward compatibility
> in 2.1 (-kvv).  I doubt that this has even been noticed."

Yeah, I could add the above, but there is already the following in the article (which is already quite long):

------------------ Bernhard then replied to each of the points Linus had raised. About "library versioning" his reply was:

> In my experience Werner (the lead GnuPG developers) is quite reasonable about
> keeping APIs stable (he often goes out of his way to keep even the command
> line version stable, maybe he shouldn't do that to the command line options
> so you are more motivated to go to this official API gpgme. >:) )

------------------

So I think Bernhard already knew and had already written that the command line API is basically stable thanks to Werner's efforts.

Show 5 quoted lines

> I think our conclusion was that on Git's side, there is no problem to
> solve (except, maybe, to use gpg2 by default when gpg is not installed)
> because the main problem is mixed installations of gpg1 and gpg2.1+, and
> we don't want to use a library instead of the command line API for the
> reasons mentioned by Linus and others.

I agree that one conclusion is that maybe we should use gpg2 by default when gpg is not installed or when both gpg and gpg2 are installed. But I think another important conclusion on the Git side is Peff's email, which basically says that gpg.program cannot be removed and "gpg.program = gpgme" could be added. But I prefer not to state these conclusions explicitly in the article as people might disagree :-)