Re: [Comm-en] RPM in ALT Linux (4.0.4 vs 4.13)

[Ivan Zakharyaschev <imz@altlinux.org>]

Hello!

On Sat, 3 Sep 2016, Neal Gompa wrote:

> I saw you guys listed as the most recent ones to change the rpm
> package in ALT Linux, and I was wondering if you guys had contemplated
> upgrading from rpm 4.0.4 to rpm 4.13?

glebfm@ and legion@ are busy now with this. 
https://lists.altlinux.org/pipermail/devel/2016-July/201603.html

They could give most details about this process.

The first thing to do on this way was to rebase many ALT's features[1] 
onto rpm(-install)-4.13. (Not yet features relevant for rpm-build.)

[1] https://www.altlinux.org/Rpm-4.13

> If so, why haven't you? What's holding you back from upgrading? I'd

Apart from the first important step (rebasing ALT's rpm-install) which has 
been done and is ready for testing, there are things would hold us back 
from putting the new version into ALT Sisyphus:

those packages which use librpm and/or access RPM's db will have to be 
adapted for the new version. (The first one, of course, is APT; then, 
there are some Perl bindings actively used by the tools for automatic 
package analysis, modification, submission; perhaps, some more, I don't 
know the list of things that hold this back well, but other involved 
people could tell you more.)

> like to see the ALT Linux rpm maintainer team be involved in upstream
> rpm.org development, as I'm sure your perspective would be valuable to
> ensure a vibrant ecosystem around rpm.

As said, there are a few ALT-specific nice, important and 
non-trivial features in RPM, which would always require maintaining a 
separate fork unless they are taken up by another RPM project, say, the 
rpm-4.13 project. Then the forces could be joined.

One of these features is the support for set-versions (the <= relation, 
which is used to constrain Requires/Provides, which would behave not like 
a linear order, but like inclusion of sets), developed by at@ in the past. 
Now, he has announced that he is developing an enhanced varaint of this 
feature and could tell the details about the current developments to those 
who are interested. 
(https://lists.altlinux.org/pipermail/devel/2016-July/201614.html : 
support for prototypes/signatures similar to C++ mangling, but for C).

at@ has pointed to his new work at https://github.com/svpv/rpmss -- 
https://lists.altlinux.org/pipermail/devel/2016-August/201701.html . At 
the same time, at@ shared his belief that if there is some code in ALT's 
RPM which was once written and works correctly since then, there will be 
no need to put efforts into maintaining it; and so, he sees no 
justification in the complex work of rebasing onto rpm-4.13 because this 
would not save us any future efforts in maintaining ALT's RPM compared to 
the current situation. (Zero efforts if the current code of RPM works 
correctly.)

glebfm@ -- Gleb Fontengauer-Malinovskiy

legion@ -- Alexey Gladkov

at@ -- Alexey Tourbin

community-en@lists -- a mailing list for the community around ALT for 
discussions in English

-- 
Best regards,
Ivan