Re: libassuan 3.0.1-1 pulled from testing after crashing pacman
On Mon, 2024-07-22 at 09:29 +0200, David Runge wrote:
Hi all, ...
have to use an installation media or static pacman to get back to 3.0.0-1. ...
David
Hey David - no worries these kind of things happen on occasion and it was only in testing, where users of which should mostly know how to recover from an occasional issue cropping up. It does bring up a question and am interested how you think about it. While pretty rare, there certainly can be occasions where an official static pacman could be pretty helpful to have. Its fairly non-trivial to build a static pacman since we don't have archive libs readily available for the 38 packages it depends on (at least that I can see). I wonder given the build tooling if : a) providing static pacman package would be a possibility? or b) provding 'archive' lib packages might be doable e.g. glibc-archive, openssl-archive, gpgme-archive ... Both seem painful: - (a) means either do (b) first or build archive versions of all 38 shared libs as part of pacman build. The latter approach is one taken by the AUR package. - (b) means changing the PKGBUILD of quite a lot of packages - and if we go this route then should it be done for other lib packages as well? Pro: While it's rarely needed it could be very useful. i.e. Protect against a Low frequency but high impact risk. Con: - A lot of work - is it worthwhile doing? - Can always boot install media for non-remote systems or use AUR package - its a lot of work for very small gain. thank you for all the work you and other devs/package managers do to keep Arch "-The-" stellar distro! -- Gene
While pretty rare, there certainly can be occasions where an official static pacman could be pretty helpful to have.
Its fairly non-trivial to build a static pacman since we don't have archive libs readily available for the 38 packages it depends on (at least that I can see). (…) Hello,
Morganamilo maintains a PKGBUILD in AUR⁽¹⁾ and the pinned comment links to the pre-build binary.⁽²⁾ Moranamilo is an official Arch PM⁽³⁾ and that binary is signed (.sig) with their key. One can verify using official Arch’s keyring: pacman-key --verify pacman-static.sig pacman-static Cheers, mpan ⁽¹⁾ https://aur.archlinux.org/packages/pacman-static ⁽²⁾ https://pkgbuild.com/~morganamilo/pacman-static/x86_64/bin/ ⁽³⁾ https://archlinux.org/people/package-maintainers/#morganamilo
On Mon, 22 Jul 2024 12:15:26 -0400 Genes Lists <lists@sapience.com> wrote:
It does bring up a question and am interested how you think about it. While pretty rare, there certainly can be occasions where an official static pacman could be pretty helpful to have.
See the note at https://aur.archlinux.org/packages/pacman-static It's from a PM and hosted on Arch infrastructure
On Mon, 2024-07-22 at 16:30 +0000, Doug Newgard wrote:
See the note at https://aur.archlinux.org/packages/pacman-static
It's from a PM and hosted on Arch infrastructure
Thank you (both) for pointing that out - indeed I missed that. 🙂 gene -- Gene
On Mon, 2024-07-22 at 12:53 -0400, Genes Lists wrote:
On Mon, 2024-07-22 at 16:30 +0000, Doug Newgard wrote:
See the note at https://aur.archlinux.org/packages/pacman-static
And thank you Morgan for providing package and pre-built binary. -- Gene
participants (3)
-
Doug Newgard
-
Genes Lists
-
mpan