Oops I messed up the quoting, for transparency here it is correctly:

On 11/22/25 6:26 AM, Joseph Brandon Kigen wrote:
> I disagree with this merge request. Here's the current situation:
>
> There are currently 6 AUR packages for Google Antigravity:
> - antigravity-bin (13 votes, most popular)
> - antigravity-bin-hardened (2 votes, specialized variant)
> - google-antigravity-bin (2 votes, *mine*)
> - antigravity-preview (0 votes)
> - antigravity-binary (0 votes)
> - antigravity (0 votes, *requester's package*)
>
> The `-bin` suffix is standard AUR practice for binary packages. The 
> requester's
> package name "antigravity" without suffix is actually the non-standard one.

According to the guidelines [1], the `-bin` suffix should be used when 
source code of the software is available, or if it's likely that source 
code will exist in the future. Google seeks a profit from Antigravity 
[2] so almost certainly we will never have source code for it. 
Therefore, there shouldn't be a `-bin` suffix.

> The requester's package (antigravity) has critical issues:
> - *Outdated* (v1.11.2 vs current v1.11.5)
> - *Broken binary symlink* (points to capital-A "Antigravity" which doesn't 
> exist)
> - *Not* maintained (last updated Nov 18, now Nov 22)
> - Zero votes, indicating no user adoption.

By the way, I am not the maintainer of `antigravity`. If they were to 
orphan it, I'd happily pick it up and fix it.

> google-antigravity-bin is:
> - Correctly structured and functional
> - Following proper naming conventions

The `google-` prefix does not follow naming conventions. Typically, 
package names keep the name of the software [3]. There's no need for a 
suffix of the name of the creator, "Google" can instead just be in the 
description.

> The most popular package (antigravity-bin with 13 votes) also uses the `-bin`
> suffix, confirming this is the accepted convention.

The convention is `-bin` when sources exist, no `-bin` when there are no 
sources. See Package Maintainer response on PRQ#77614 [4].

> If consolidation is desired, it should be around the most popular and 
> correctly- named packages
> (antigravity-bin or google-antigravity-bin),  not the broken, unmaintained
> "antigravity" package with zero adoption. The requester should orphan their
> package, not request others merge into it. The request should be rejected.

No, according to the guidelines [5], if a package is broken, changes 
should be submitted to the Maintainer.

> Given the fragmentation (6 packages doing more or less the same thing), I 
> propose:
> 1. Keep the two legitimate -bin packages:
> - antigravity-bin (most popular, 13 votes)
> - google-antigravity-bin (my package, proper Google branding)
>
> 2. Keep specialized variants:
> - antigravity-bin-hardened (serves specific security use case)
>
>     3. Orphan/remove broken/duplicate packages:
> - antigravity (broken, unmaintained)
> - antigravity-preview (outdated, unclear purpose[Suggests preview but fetches
> from the same source as per its PKGBUILD])
> - antigravity-binary (duplicate of -bin packages)
>
> I'm willing to orphan my package in favor of antigravity-bin IF the community
> prefers consolidation around that name. However, merging into a broken package
> makes no sense.

Since it is identical software, there should only be 1 package, 
`antigravity`. It doesn't matter if that package is currently broken, 
changes should be submitted to that maintainer, or it should be 
orphaned, and duplicate packages should not be created.

Best Regards, AlphaLynx [1] 
https://wiki.archlinux.org/title/Nonfree_applications_package_guidelines#Package_naming
 
[2] https://antigravity.google/pricing [3] 
https://wiki.archlinux.org/title/PKGBUILD#pkgname [4] 
https://lists.archlinux.org/archives/list/[email protected]/thread/IZY3AF6TT4Q64FG5INYNEJ2C5TN7YJQQ/#IZY3AF6TT4Q64FG5INYNEJ2C5TN7YJQQ
 
[5] 
https://wiki.archlinux.org/title/AUR_submission_guidelines#Rules_of_submission

Attachment: OpenPGP_0x100ED08AC2F74784.asc
Description: application/pgp-keys

Attachment: OpenPGP_signature.asc
Description: PGP signature

Reply via email to